 

. 7‘3

‘1},

;

‘ .. {H.112-

Ip. a t.
3%.:
111:3...

45!?

 

 

 

was

LIBRARY
> 3': 3 Michigan State
9 University

This is to certify that the
dissertation entitled

A MINIMALLY INVASIVE SYSTEM FOR THE EVALUATION
OF PROSTHETIC HEART VALVES

presented by

Shiu Chuen Chan

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Electrical and Computer
Engineering

 

 

Mic We

 

Major Professor‘s Signature
NM l; 049%
u I

Date

 

MSU is an affirmative-action, equal-opportunity employer

_.-.—.--—.—--—n—-—.--.—-o---—--—--—-u-n—.---—.-.--.-.-.--.--—--

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5I08 K:IProlecc&PrelelRC/DateDueAindd

 

A MINIMALLY INVASIVE SYSTEM FOR THE EVALUATION OF PROSTHETIC
HEART VALVES

By

Shiu Chuen Chan

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Electrical and Computer Engineering

2008

ABSTRACT

A MINIMALLY INVASIVE SYSTEM FOR THE EVALUATION OF PROSTHETIC
HEART VALVES

By

Shiu Chuen Chan

The Bjtirk-Shiley Convexo-Concave (BSCC) heart valve employs a pyrolytic carbon disc,
held in place by two struts, as an occluder to prevent retrograde ﬂow of blood in the heart.
The strut on the inlet side is integral to the suture ring while the strut on the outlet side is
welded to the suture ring. In some cases, for reasons yet unclear, the weld at one of the two
sites fractures after several years of use. This results in a single-leg separation (SLS) of the
strut. The fracture causes the load on the weld at the other end of the strut to increase,
thereby setting the stage for it to fail eventually. A complete fracture of the outlet strut
would cause the disc to escape resulting in mortality. This thesis describes a minimally
invasive electromagnetic approach, called the gradiometer-based method, for detecting SLS
failures. The feasibility of the method is investigated using three dimensional ﬁnite element
analysis techniques. A clinical scale gradiometer-based examination system is constructed

and in vitro studies of the system have shown 100% accuracy in detecting SLS cases.

To my wife, Suzanne, and our extended families.

iii

 

 

 

 

.14

 

ACKNOWLEDGEMENTS

I am grateful to the Bowling-Pﬁzer Settlement Funds Supervisory Panel for generously
supporting of this work. Their support has providedaccesses to invaluable resources and

opportunities that have helped made this project a success.

I would like to thank the members of my Guidance Committee: Dr. Christina Chan, Dr.
Pradeep Ramuhalli, Dr. Lalita Udpa, and my Major Professor, Dr. Satish Udpa for their
willingness and time spent serving on my committee, and the many intellectual discussions

on this and other subjects.

lam also deeply indebted to Drs. Lalita and Satish Udpa who had critically reviewed this

thesis despite their overwhelming schedules.

A project of this magnitude cannot be accomplished by one person alone. I am very thankful
for my colleagues and friends, past and present, in the “Heart Valve Project,” all of whom I
have the pleasure to work, brain-storm, and, at times, constructively disagree with. In
particular, I wish to thank Dr. Tian Yong for the many discussions on induction heating
modeling, Dr. Lalita Udpa and Dr. Zhiwei Zeng for the discussions on probability of
detection models, and Mr. Robert Clifford for the construction of the various mechanical

components of the system.

iv

A very special thank you goes to Dr. Mohachiro Oka of the Oita National College of
Technology (Oita, Japan), whose collaboration was instrumental in the development of the

ﬁrst gradiometer prototype that had since evolved into the current designs.

I would like to express my appreciation to all the world-class faculty, staff, members, and
visiting scholars of the Nondestructive Evaluation Laboratory at the Department of Electrical

and Computer Engineering at Michigan State University for all the opportunities and support

over the years.

A very special thanks goes to my mother, late father, brother and his family, as well as my

wife’s family for their unconditional support and care.

Last but not least, I wish to extend my deepest gratitude to my wife, Suzanne, who is an
Angel from Heaven sent by God. Her endless love, patience, care, and encouragement have

accompanied me over many challenges of this project as well as the completion of this thesis.

TABLE OF CONTENTS

LIST OF TABLES .................................................................................................................... x
LIST OF FIGURES ................................................... g ............................................................. xii
CHAPTER 1. INTRODUCTION .......................................................................................... 1
1.1 Organization of the Thesis ........................................................................................... 3
1.2 Beneﬁts of Work .......................................................................................................... 4
CHAPTER 2. THE GRADIOMETER-BASED METHOD .................................................. 5
2.1 Faraday’s Law and Maxwell’s Equations .................................................................... 5
2.2 Implementing the Gradiometer-Based Method ............................................................ 9
CHAPTER 3. BJORK-SHILEY CONVEXO-CONCAVE PROSTHETIC HEART VALVE
10
3.1 Outlet Strut Fracture of the BSCC Valve ................................................................... 11
3.1.1 Loading of the Outlet Strut .................................................................................. 13
3.2 Care for the BSCC Heart Valve Patients ................................................................... 16
3.2.1 BSCC Heart Valve Outlet Strut Fracture Mechanisms ....................................... 17
3.2.2 The Mechanics of Valve Closure ........................................................................ 18
3.2.3 Dynamics of the BSCC Valve ............................................................................. 22
3.2.4 Valve Closure Delay ........................................................................................... 25
3.3 Metallurgical Analyses of the Heart Valve and the Outlet Strut Weld Sites ............. 26
3.3.1 Risk Factor Analysis and Prophylactic Valve Replacement ............................... 28
3.4 Previous Approaches in BSCC Heart Valve SLS Detection ...................................... 31
3.4.1 Radiographic Detection Approach ...................................................................... 31
3.4.2 Echocardiography, Acoustics, Ultrasound and Electromagnetic Acoustic
Transduction Based Approaches ..................................................................................... 36
3.4.3 Electromagnetic Approaches ............................................................................... 45
3.4.4 Differential Eddy Current Approach ................................................................... 46
3.4.5 Catheter-Based Antenna Approach ..................................................................... 48
3.5 A Novel Approach for SLS Detection in BSCC Heart Valves .................................. 51
CHAPTER 4. MATHEMATICAL MODEL ....................................................................... 52
4.1 Induced Current in the Outlet Strut ............................................................................ 53
4.2 Magnetic Field Due to the Helmholtz Coil System ................................................... 56
4.3 Magnetic Field Due to the Outlet Strut Loop Current ............................................... 57
4.4 Induced EMF in the Gradiometer Coils ..................................................................... 61
4.5 Special Case: Gradiometer on Axis of Outlet Strut Loop ......................................... 63
4.6 Induced EMF in the Gradiometer Coils ..................................................................... 64
4.7 Outlet Strut Condition and Strut Loop Resistance ..................................................... 67
4.8 An Alternative Approach: Mutual Inductance Consideration .................................... 68
4.9 Factors Affecting the Gradiometer Signal Output ..................................................... 73
4.10 Spatial Dependency of the Gradiometer Output Signal ........................................... 74

vi

CHAPTER 5. NUMERICAL MODELING ......................................................................... 80

5.1 The Finite Element Model .......................................................................................... 80
5.2 Simulation Results ...................................................................................................... 85
5.2.1 The Excitation Coils ............................................................................................ 86
5.2.2 Field Perturbations Due to Outlet Strut Conditions ............................................ 88
CHAPTER 6. SYSTEM COMPONENTS ........................................................................... 98
6.1 The Excitation Subsystem .......................................................................................... 99
6.1.1 Signal Generator ................................................................................................ 100
6.1.2 Power Ampliﬁer ................................................................................................ 100
6.1.3 Impedance Matching Circuit ............................................................................. 101
6.1.4 Excitation Coils ................................................................................................. 102
6.1.5 Helmholtz Coil .................................................................................................. 103
6.2 Excitation Coils Mobile Custom Support System .................................................... 106
6.3 The Measurement Subsystem ................................................................................... 107
6.3.1 Lock-in Ampliﬁer ............................................................................................. 107
6.3.2 Data Acquisition System ................................................................................... 114
6.4 Custom Examination Table ...................................................................................... 115
CHAPTER 7. GRADIOMETER DESIGNS ...................................................................... 120
7.1 Gradiometer Output Signal ...................................................................................... 121
7.2 Physical Constraints ................................................................................................. 123
7.3 Delivery Approaches ................................................................................................ 124
7.4 Gradiometer Assemblies .......................................................................................... 126
7.4.1 The 9 F Gradiometer Assembly ........................................................................ 127
7.4.2 The 6 F Gradiometer Assembly ........................................................................ 132
7.5 Gradiometer Selection .............................................................................................. 137
7.5.1 Practical Considerations .................................................................................... 138
CHAPTER 8. SYSTEM CHARACTERIZATION AND TEST RESULTS ..................... 140
8.1 The Helmholtz Excitation Coil System .................................................................... 140
8.2 Lock-In Measurement of the Gradiometer Signal .................................................... 143
8.3 Clinical-Scale Prototype System .............................................................................. 143
8.4 Output Signal as a Function of Distance .................................................................. 145
8.4.1 Movement Model .............................................................................................. 148
8.5 Gradiometer Output as a Function of Excitation Frequency .................................... 152
8.5.1 In-Phase Signal, X ............................................................................................. 152
8.5.2 Quadrature, Y, and Amplitude, R, Signals ........................................................ 154
8.5.3 Output Signal Phase Angle, 0 ............................................................................ 156
8.5.4 Outlet Strut Condition Indicator and Excitation Frequency Speciﬁcation ....... 157
8.6 In Vitro Test System ................................................................................................. 157
8.6.1 In Vitro Test System Outputs and Classiﬁcations ............................................. 160
8.7 In Vitro Study Results .............................................................................................. 160
8.7.1 In Vitro Study: Trial A, 19 BSCC Heart Valves, 9 F Gradiometer .................. 166
8.7.2 In Vitro Study: Trial B, 19 BSCC Heart Valves, 9 F Gradiometer ................... 170
8.7.3 In Vitro Study: Trial C, 34 BSCC Heart Valves, 6 F Gradiometer ................... 173
8.7.4 Selection of the Computer Decision Threshold ................................................ 178

vii

 

8.7.5 Trial D: Threshold Veriﬁcation, 11 BSCC Heart Valves, 6 F Gradiometer.... 182

CHAPTER 9. PROBABILITY OF DETECTION MODEL ............................................. 186
9.1 Sources of Variability ............................................................................................... 186
9.2 POD Model .............................................................................................................. 187

9.2.1 Functional Model .............................................................................................. 192
9.3 Parametric Study ...................................................................................................... 192
9.3.1 POD With Respect to the Gradiometer Distance from the Outlet Strut ............ 193

CHAPTER 10. SIMULATION OF IN DUCTIVE HEATING EFFECTS ........................ 202
10.1 Model Formulation ................................................................................................. 203
10.2 Model Description .................................................................................................. 205
10.3 Simulation Results .................................................................................................. 209

10.3.1 Maximum Temperature Increase After Three Minutes of Exposure .............. 210
10.3.2 Maximum Temperature Increase After Fifteen Minutes of Exposure ............ 217
10.3.3 Special Case I: Excitation Without Blood Circulation ................................... 220
10.3.4 Special Case 11: Excitation without Convection and Conduction Cooling 225

CHAPTER 11. PATIENT SAFETY AND REGULATORY COMPLIANCE ................. 235
11.1 Health Effects of Radio Frequency EM Field Exposure ........................................ 235
11.2 Nature and Duration of EM Field Exposure in the Gradiometer-Based Examination
System ............................................................................................................................... 237
11.3 Speciﬁc Absorption Rate ....................................................................................... 238
11.4 Speciﬁc Absorption Rate Calculation .................................................................... 240

11.4.1 SAR Inference from Derived Electric Field Strength ..................................... 240
11.4.2 Region-Speciﬁc SAR Calculations ................................................................. 242
11.5 Applicable Regulatory Standards and Guidelines .................................................. 246
11.5.1 The IEEE Standard (IEEE Std C95.1—2005) ................................................... 246
11.5.2 The IEC Standard (IEC 60601-2-33) .............................................................. 252
11.5.3 The ICNIRP Guidelines .................................................................................. 259
11.5.4 The FDA Guidelines ....................................................................................... 260
11.6 Compliancy Statement ........................................................................................... 262

CHAPTER 12. CONCLUSIONS AND FUTURE WORK ............................................... 263
12.1 Future Work ........................................................................................................... 265

APPENDIX A. BIOLOGY OF THE HEART AND HEART VALVE DISORDERS ..... 267
Al The Heart ................................................................................................................. 267
A2 Heart Valve Disorders ............................................................................................. 271

A21 Mitral Valve Prolapse ....................................................................................... 272
A22 Mitral Stenosis or Obstruction ......................................................................... 272
A23 Mitral Regurgitation or Insufﬁciency ............................................ . .................. 272
A3 Treatments of Heart Valve Disorders ...................................................................... 273
A.3.l Observations and Monitoring ........................................................................... 273
A32 Medications ...................................................................................................... 274
A33 Surgical Repair ................................................................................................. 275
A34 Heart Valve Replacement ................................................................................. 275

viii

 

 

A.4 Types of Prosthetic Heart Valves ............................................................................ 277

A.4.l Mechanical Valves ........................................................................................... 277
A42 Ball Valves ....................................................................................................... 277
A43 Disc Valves ....................................................................................................... 278
A44 Tissue Valves ................................................................................................... 281
A.4.4.1 Animal Tissue Valves (xenografts, heterografts) .................................................. 281
A.4.4.2 Human Tissue Valves (homografts, autografts) and the Ross Procedure ............. 283
A.4.4.3 Decellularized homograft and xenograft aortic prosthesis (Synergraft) ............... 284
APPENDIX B. ANIMAL STUDY EXAMINATION PROCEDURE .............................. 286
B.1 The Animal Study Procedure .................................................................................. 286
311 Animal Support ................................................................................................. 286
3.1.2 Position and Orientation of the Test Animal .................................................... 286
8.1.3 Calibration and Zeroing of Measurement Instruments ..................................... 287
8.1.4 Positioning of the Gradiometer ......................................................................... 289
3.1.5 Examination and Data Acquisition ................................................................... 290
B.1.6 Animal Recovery .............................................................................................. 290

B2 Critical Considerations ............................................................................................ 290
B3 Recommended Data Capture Events ....................................................................... 291
BIBLIOGRAPHY ................................................................................................................. 294

ix

 

 

LIST OF TABLES

Table 3.1 Deﬁnition of Image Assessment Grades for Visual Interpretation of

Cineangiographs ..................................................................................................................... 33
Table 7.1Typical resistance and inductance of the 9 F gradiometer sensor coils (Measured at
1 MHz) .................................................................................................................................. 128
Table 7.2 Components of the 9 F (POT-9075) gradiometer assembly ................................. 129
Table 7.3 Total resistance and inductance of the 6 F gradiometer sensor coils (Measured at

100 kHz) ................................................................................................................................ 133
Table 7.4 Component list for the 6 F (POT-9083) gradiometer assembly ........................... 138

Table 8.1 Summary of the ﬁnite element analysis and physical measurement parameters and
results for the magnetic ﬂux density at the center of the Helmholtz excitation coil ............ 142

Table 8.2 Comparison of the peak-to-peak X signal voltages between the 108 and SLS cases
at various excitation frequencies ........................................................................................... 153

Table 8.3 Comparison of the peak-to-peak Y signal voltages between the 108 and SLS cases
at various excitation frequencies ........................................................................................... 155

Table 8.4 Comparison of the peak-to—peak R signal voltages between the 108 and SLS cases
at various excitation frequencies ........................................................................................... 155

Table 8.5 Comparison of the peak-to-peak 6 value between the 108 and SLS cases at various
excitation frequencies ........................................................................................................... 157

Table 8.6 Summary of In Vitro Double-Blinded Test Results of 19 BSCC Heart Valves with

a 9 F gradiometer (Trial A) ................................................................................................... 169
Table 8.7 Summary of In Vitro Double—Blinded Test Results of 19 BSCC Heart Valves with
the 9 F gradiometer (Trial B) ................................................................................................ 172
Table 8.8 Summary of In Vitro Double-Blinded Test Results of 34 BSCC Heart Valves With
”‘6 6 F gradiometer (Trial C) ................................................................................................ 177
Table 8.9 Summary of the mean, standard deviation, and decision thresholds for the cases
When all SLS data are used and when the outlier data points are excluded .......................... 179
Table 8.10 Summary of In Vitro Double-Blinded Test Results of 11 BSCC Heart Valves with
"‘6 6 F gradiometer (Trial D) ................................................................................................ 184
Table 10.1 Selected convective ﬁlm coefﬁcients. After [15 8] ............................................ 205

 

 

 

 

 

.

. . .

‘I

.“

l .
I.
'-
7 \
-h”
5' “‘
u
.d
nwr'.
i .
.
.
.- .-
.
g. ,
.
..
.
‘A
I» c
g...
' u
H”
.
.
n h
‘1.
‘-
'l
‘v‘

 

Table 10.2 Material properties used in the ﬁnite element model. After [159]-[162] .......... 207

Table 10.3 Summary of maximum temperature elevations of the heart valve after 180 and
900 seconds of continuous exposure, respectively, for the normal and worst-case heat transfer
scenarios ................................................................................................................................ 234

Table 11.1 Selected results from the RF inductive heating FEA from Chapter 9 ................ 244

Table 11.2 Basic restrictions on exposure for the various regions of the body (after [163],
Table l), with limits for the gradiometer-based examination system ................................... 248

Table 11.3 MPE for the head, torso and limbs: f = 3.35 kHz to 5 MHz (after [166], Tables 2
and 3) .................................................................................................................................... 250

Table 11.4 The maximum permissible values, 3,, , of the time derivative of the external

magnetic ﬁeld for the head, torso and limbs (f = 65 kHz). ................................................... 252
Table 11.5 Gradient output limits for the gradiometer-based examination system at the

Operating frequency of f = 65 kHz. (The n, values are obtained from [168].) ..................... 257
Table 11.6 IEC temperature limits for RF exposure (after [169], Table 104) ...................... 258
Table 11.7 SAR limits (after [167], Table 105) .................................................................... 258
Table 11.81CNIRP basic restrictions for a time-varying EM ﬁeld at f = 65 kHz ................ 260

 

LIST OF FIGURES

Figure 2.1 Illustration of the gradiometer-based approach for BSCC heart valve SLS
detection (components not drawn to scale) ............................................................................... 6

Figure 2.2 Induced EMF due to a time-varying magnetic ﬁeld, B. After [2] .......................... 7

Figure 2.3 Induced current in the outlet strut by the excitation ﬁeld in turn creates a magnetic
ﬁeld which perturbs the original excitation ﬁeld (magnetic ﬂux lines simpliﬁed for clarity;

components not drawn to scale) ................................................................................................ 9
Figure 3.1 ijrk-Shiley Convexo-Concave (BSCC) 60° heart valve. The 70° valve has a

similar design except the occluder disc opens at a 70° angle ................................................. 11
Figure 3.2 A BSCC heart valve with the woven Teﬂon fabric sewing ring ........................... 11

Figure 3.3 Drawings of a BSCC heart valve showing the disc in the open (top) and closed
positions (bottom). The disc opening angle 0 is either 60° or 70°. Adapted from [17] ........ 12

Figure 3.4 Scanning electron microscope (SEM) image of a fractured outlet strut viewed
from the outlet side. Adapted from [81] ................................................................................ 15

Figure 3.5 Inlet side view SEM image of the fractured outlet strut shown in ........................ 15

Figure 3.6 Chest radiograph of a patient with outlet strut fracture. The outlet strut has
separated and is ﬂoating near the ﬂange. The occluder disc (arrow) has embolized the aortic

arch. Adapted from [81] ......................................................................................................... 16
Figure 3.7 Relationship between outlet strut loading and disc rotation during valve closure
measured in a pulse duplicator. Adapted from [17] ............................................................... 18
Figure 3.8 Prior to April 1984, the manufacturing tolerance speciﬁed gaps at the base and tip
of the outlet strut. Adapted from [17]. ................................................................................... 19
Figure 3.9 Hook-to-well gap between the outlet strut tip and the ocluder disc well. Adapted
from [17]. ................................................................................................................................ 20
Figure 3.10 A simulated case of disc over-rotation. F0 is the vertical implact force on the
outlet strut. Adapted from [17]. ............................................................................................. 20

Figure 3.11 Loading of the inlet and outlet struts during the occluder disc over-rotation.
Adapted from [104] ................................................................................................................. 20

Figure 3.12 After April 1984, the new manufacturing speciﬁcations ensured that the disc
remained in contact with the base of the outlet strut when at rest. Adapted from [17]. ,,,,,,,, 21

xii

 
 

 

Figure 3.13 Finite element analysis of the outlet strut leg showing the stress distribution due
to tip loading by the occluder disc during over-rotation (o-max denotes the region of
maximum stress). Adapted from [17]. ................................................................................... 23

Figure 3.14 Cineradiographic tunnel view of a BSCC heart valve showing a case of “deﬁnite
SLS” assessment (arrow). Adapted from [81]. ...................................................................... 32

Figure 3.15 Cineradiographic oblique view of a BSCC heart valve showing a case of
"deﬁnite SLS" assessment (arrow). Adapted from [81]. ....................................................... 33

Figure 3.16 Orientation of the EMAT excitation system. Patient not shown for clarity.
Adapted from [121] ................................................................................................................. 45

Figure 3.17 A differential eddy current probe for detecting the outlet strut condition of a
BSCC heart valve. Adapted from [118] ................................................................................. 47

Figure 3.18 Differential eddy current probe detection approach. Adapted from [118]. ........ 47
Figure 3.19 Circuit diagram of the differential eddy current probe. Adapted from [118] ..... 47

Figure 3.20 Examples of differential eddy current probe signals from an in vitro test. The
large amplitude signal is from an 108 valve and the small amplitude signal is from a SLS
case. Adapted from [118] ....................................................................................................... 48

Figure 3.21 Induced current along a wire-loop model of a heart valve. The different
impedance valves (Zl's) represent various ﬂaw depths. Adapted from [116]. ...................... 50

Figure 3.22 Results in Figure 3.21 normalized with respect to the 21:106 (2 signal. Adapted
from [1 16]. .............................................................................................................................. 51

Figure 4.1 The heart valve is submerged in a uniform magnetic ﬁeld and oriented in a fashion
such that the ﬂange ring intercepts the least amount of the applied ﬁeld, while the outlet strut
loop intercepts the maximum amount of the ﬁeld. ................................................................. 54

Figure 4.2 Current is induced in the outlet strut loop due to the component of the applied
magnetic ﬁeld that is normal to the plane of the loop ............................................................. 55

Figure 4.3 If the heart valve is rotated ,6 degrees with respect to the applied magnetic ﬁeld,
the effective area of the outlet strut loop for intercepting the applied ﬁeld will be reduced. . 55

Figure 4.4 Off-axis magnetic ﬁeld due to the outlet strut (approximated by a circular current

loop). ....................................................................................................................................... 59
Figure 4.5 A gradiometer placed off-axis from the outlet strut loop ...................................... 61
Figure 4.6 A gradiometer placed on the axis of the outlet strut loop ...................................... 64

xiii

Figure 4.7 A conceptual representation of the gradiometer—based inspection system
(components not drawn to scale) ............................................................................................ 70

Figure 4.8 A circuit element representation of the gradiometer-based inspection system
components (shown with the ﬁrst gradiometer sensor coil only) in Figure 4.7. .................... 70

Figure 4.9 Scaled gradiometer output signal as a function of the distance, 21 , between the
gradiometer and the outlet strut loop ...................................................................................... 75

Figure 4.10 Scaled gradiometer output signal as a function of the separation, A2, between the
two sensor coils ....................................................................................................................... 76

Figure 4.11 Scaled amplitude of the gradiometer output signal as a function of the excitation
ﬁequency ................................................................................................................................. 78

Figure 5 .1 A typical conﬁguration for eddy current problems ............................................... 81

Figure 5 .2 Finite element mesh of a BSCC heart valve (the inlet strut and occluder disc are
omitted for clarity without loss of model accuracy) ............................................................... 85

Figure 5.3 Each excitation coil consists of nine turns of a 1/0 AWG wire. Due to the inter-
wire spacing and jacket thickness, the ﬁll-factor of this coil less than 0.57. .......................... 87

Figure 5.4 Excitation coil windings are modeled as a wire with a square cross-sectional area
................................................................................................................................................. 87

Figure 5.5 Finite element analysis of the Helmholtz coil showing the magnetic ﬂux density
and ﬁeld (arrows) distribution ................................................................................................ 89

Figure 5.6 Magnetic ﬂux distribution at the center of the Helmholtz coil along the x axis on
the y=z=0 line ......................................................................................................................... 90

Figure 5.7 Magnetic ﬂux density at the center of the Helmholtz coil along the Z axis on the
x=y=0 line ............................................................................................................................... 91

Figure 5.8 Finite element model of a BSCC heart valve submerged in a volume of uniform
magnetic ﬁeld .......................................................................................................................... 92

Figure 5.9 Real part of the (By, Bz) components in the x = 0 mm plane for both 108 and SLS
Cases. The vertical axis is the y-axis. In these cases, the real components of the ﬁelds are
Unperturbed. ............................................................................................................................ 93

Figure 5.10 Imaginary part of the (By, Bz) components in the x = 0 mm plane for an 103
Valve. The vertical axis is the y-axis. ..................................................................................... 94

Figure 5.11 Imaginary part'of the (By, Bz) components in the x = 0 mm plane for a SLS
Valve. The vertical axis is the y-axis. ..................................................................................... 95

xiv

Figure 5.12 Real (a) and imaginary (b) parts of the (Bx, By) components in the z = -5 mm
plane for an 108 valve. The vertical axis is the y-axis. ......................................................... 96

Figure 5.13 Real (a) and imaginary (b) parts of the (Bx, By) components in the z = -5 mm
plane for a SLS valve. The vertical axis is the y-axis. ........................................................... 97

Figure 6.1 System block diagram of the gradiometer-based prosthetic heart valve
examination system (components not drawn to scale) ........................................................... 99

Figure 6.2 Impedance matching circuit. Rr denotes a bank of power resistors, and Cr denotes
a bank of power capacitors. .................................................................................................. 102

Figure 6.3 Illustration of the magnetic ﬂux lines generated from a Helmholtz coil. After
[129]. ..................................................................................................................................... 105

Figure 6.4 Side (a) and top (b) views of the excitation coils. The excitation coils are mounted
on a modular wooden structure arranged in a Helmholtz conﬁguration. ............................. 108

Figure 6.5 Excitation coils are mounted with an inclination to facilitate the alignment of the
magnetic ﬁeld with the ﬂange ring of the heart valve. ......................................................... 109

Figure 6.6 Dimensions of the excitation coils and the mobile custom coil support system. 109

Figure 6.7 Custom examination table positioned between the excitation coils. The excitation

coils system is designed to accommodate both the patient and the custom examination table.
............................................................................................................................................... 1 10

Figure 6.8 A mobile support frame allows the positioning of the coils around the patient and
the custom examination table with no disturbance to the patient. ........................................ 110

Figure 6.9 Each platform caster has a leveling pad that can be raised and lowered by a
telesc0ping ratchet handle for parking the support structure. ............................................... 111

Figure 6.10 Custom examination table with a nonmetallic composite top that does not
interfere with the applied magnetic ﬁeld .............................................................................. 116

Figure 6.11 Tabletop dimensions and height adjustment range for the custom surgery bed 116

Figure 6.12 Two views of the ﬁnished custom surgery table. Two large wheel handles on the
Side are for adjusting the pitch and roll of the tabletop ......................................................... 117

Figure 6.13 A bracket is welded underneath the base of the custom examination table for
accommodating a wooden extension arm that provides additional support and stability ..... 118

Figure 6.14 A T-shape wooden extension is used to increase the stability of the custom
Surgery table when in use... ................................................................................................... 118

IFigure 6.15 Custom surgery table with a T-shape wooden extension for increased stability
. . ............................................................................................................................................. 119

XV

Figure 7.1 Scaled gradiometer output U(z1, Az) as a function of Az at values of 21 .......... 122

Figure 7.2 A 2D view of the data from Figure 7.1 showing the strongest output signal

occurring when 21 < 4 mm and Az 2 3 mm .......................................................................... 122
Figure 7.3 Advancement of the gradiometer assembly inside the heart (black arrow) ........ 125
Figure 7.4 A sealed catheter-encased gradiometer assembly is delivered intraluminally and
positioned in the mitral annulus. After [195] ........................................................................ 125
Figure 7.5 A guide catheter is pre-positioned in the left ventricle and the gradiometer
(circled) is advanced via the catheter into the mitral annulus. After [195] .......................... 126
Figure 7.6 Schematic diagram of the 9 F gradiometer assembly .......................................... 130
Figure 7.7 9 F gradiometer assembly .................................................................................... 130
Figure 7.8 Schematic diagram of the 9 F gradiometer sensor head (POT-9075) ................. 131
Figure 7.9 Close-up view of the 9 F gradiometer (POT-9075) sensor head ......................... 131
Figure 7.10 9 F gradiometer encased and sealed inside a guide catheter ............................. 132
Figure 7.11 Schematic diagram of the 6 F gradiometer assembly ........................................ 133
Figure 7.12 6 F gradiometer, POT-9083 ............................................................................... 134
Figure 7.13 Schematic diagram of the 6 F gradiometer sensor head (POT-9083) ............... 135
Figure 7.14 Sensor head of the 6 F gradiometer (POT-9083) .............................................. 135
Figure 7.15 Schematic diagram of a modiﬁed 6 F gradiometer sensor head with a tapered
cable anchor .......................................................................................................................... 136
Figure 7.16 Cardima, Incorporated, Naviport® guide catheter. After [154] ....................... 137
Figure 8.1 The Helmholtz excitation coil system has a uniform magnetic at its center within a
cylindrical volume of radius 0.25 m and 0.5 m in height ..................................................... 141
Figure 8.2 Magnetic ﬁeld distribution at the center of the Helmholtz coil. The results show a
Volume of uniform ﬁeld with approximately 0.74 % variation ............................................ 142
Figure 8.3 Clinical-scale prototype for the gradiometer-based examination system ............ 144

Figure 8.4 Test ﬁxture of the gradiometer-based examination system prototype. The heart
Valve is mounted in an elastromer holder and the gradiometer is encased in a guide catheter
- - . ............................................................................................................................................ 145

xvi

Figure 8.5 In-phase component output, X, from the LIA as a function from the outlet strut for
the 10S, SLS, and no valve cases ......................................................................................... 146

Figure 8.6 Quadrature output, Y, from the LIA as a function of the distance between the
gradiometer and the outlet strut for the IDS, SLS, and no valve cases ................................. 147

Figure 8.7 Magnitude, R, of the gradiometer output as a function of distance from the outlet
strut for the 108, SLS, and no valve cases ............................................................................ 147

Figure 8.8 Phase angle, 6, output as a function of distance from the outlet strut for the 105,
SLS, and no valve cases ........................................................................................................ 148

Figure 8.9 Model-generated X signal with respect to the changing distance between the
gradiometer and outlet strut for the IDS, XSLS and no heart valve cases. L = 10 mm. ...... 150

Figure 8.10 Estimated Y signal over time with respect to the changing distance between the
gradiometer and outlet strut for the IDS, XSLS and no heart valve cases. L = 10 mm. ...... 150

Figure 8.11 Estimated R signal over time with respect to the changing distance between the
gradiometer and outlet strut for the IDS, XSLS and no heart valve cases. L = 10 mm. ...... 151

Figure 8.12 Estimated 0 signal over time with respect to the changing distance between the
gradiometer and outlet strut for the 108, XSLS and no heart valve cases. L = 10 mm. ...... 151

Figure 8.13 Estimated relative changes in X for the 108 and SLS cases at 41.15, 45.58, 51.57,

54.85, and 60.45 kHz. L = 10 mm ....................................................................................... 153
Figure 8.14 Estimated relative changes in Y for the 108 and SLS cases at 41.15, 45.58, 51.57,
54.85, and 60.45 kHz. L = 10 mm ....................................................................................... 154
Figure 8.15 Estimated relative changes in R for the 108 and SLS cases at 41.15, 45.58, 51.57,
54.85, and 60.45 kHz. L = 10 mm ....................................................................................... 155
Figure 8.16 Estimated relative changes in 0 for the IOS and SLS cases at 41.15, 45.58, 51.57,
54.85, and 60.45 kHz. L = 10 mm ............ ' ........................................................................... 156
Figure 8.17 In situ movement of a BSCC heart valve obtained from a series of ﬂuoroscopic
images ................................................................................................................................... 158
Figure 8.18 Movement of the heart valve in relation to a stationary observation point ....... 159
Figure 8.19 Motorized stages programmed to simulate the relative movement between the
gradiometer and the outlet strut ............................................................................................ 159
Figure 8.20 Output signal X for a BSCC heart valve with SLS (A5023) ............................. 161
Figure 8.21 Output signal Y for a BSCC heart valve with SLS (A5023) ............................. 162
Figure 8.22 Output signal R for a BSCC heart valve with SLS (A5023) ............................. 162

xvii

Figure 8.23 Output signal 6 for a BSCC heart valve with SLS (A5023) .............................. 163

Figure 8.24 Output signal X for a BSCC heart valve with an IOS (87384) .......................... 163
Figure 8.25 Output signal Y for a BSCC heart valve with an 108 (87384) .......................... 164
Figure 8.26 Output signal R for a BSCC heart valve with an 103 (87384) .......................... 164
Figure 8.27 Output signal 6 for a BSCC heart valve with an [OS (87384) ......................... .. 165
Figure 8.28 Comparison of the LIA X and Y signals for the 108 (solid line) and XSLS
(dashed line) cases. ............................................................................................................... 165
Figure 8.29 Comparison of the LIA R and 6 signals for an 108 (solid line) and XSLS (dashed
line) cases .............................................................................................................................. 166
Figure 8.30 Output 6 signals from the BSCC heart valve study (Trial A), part I ................. 167
Figure 8.31 Output 6 signals from the BSCC heart valve study (Trial A), part II ............... 168
Figure 8.32 Output 6 signals from the BSCC heart valve study (Trial A), part III .............. 168
Figure 8.33 Output 6 signals from the BSCC heart valve study (Trial A), part IV .............. 169
Figure 8.34 Output 6 signals from the BSCC heart valve study (Trial B), part I ................. 170
Figure 8.35 Output 6 signals from the BSCC heart valve study (Trial B), part II ................ 171
Figure 8.36 Output 6 signals from the BSCC heart valve study (Trial B), part III .............. 171
Figure 8.37 Output 6 signals from the BSCC heart valve study (Trial B), part IV .............. 172
Figure 8.38 Output 6 signals from the BSCC heart valve study (Trial C), part I ................. 173
Figure 8.39 Output 6 signals from the BSCC heart valve study (Trial C), part II ................ 174
Figure 8.40 Output 6 signals from the BSCC heart valve study (Trial C), part III .............. 174
Figure 8.41 Output 6 signals from the BSCC heart valve study (Trial C), part IV .............. 175
Figure 8.42 Output 6 signals from the BSCC heart valve study (Trial C), part V ............... 175
Figure 8.43 Output 6 signals from the BSCC heart valve study (Trial C), part VI .............. 176
Figure 8.44 Output 6 signals from the BSCC heart valve study (Trial C), part VII ............. 176

Figure 8.45 Peak-tO-peak 6 variation for the samples examined in Trial C. The mean,
Standard deviation, and various thresholds for 6 are as indicated ......................................... 180

xviii

Figure 8.46 Peak-to—peak 6 variation for the samples examined in Trial C. The adjusted

mean, standard deviation, and thresholds for 6 are as indicated. .......................................... 181
Figure 8.47 Output 6 signals from the BSCC heart valve study (Trial D), part I ................. 183
Figure 8.48 Output 6 signals from the BSCC heart valve study (Trial D), part II ............... 183
Figure 8.49 Output 6 signals from the BSCC heartvalve study (Trial D), part III .............. 184

Figure 8.50 Peak-to-peak 6 variation for the heart valves examined in Trial D. The adjusted
thresholds from Table 8.9 are included for comparison. ...................................................... 185

Figure 9.1 Optimal position and movement range for the gradiometer with respect to the
outlet strut ............................................................................................................................. 188

Figure 9.2 Examples of variations in the gradiometer position with respect to the heart value

outlet strut: (a)-(c) translational, (e)-(f) alignment angle. ..................................................... 188
Figure 9.3 Probability density functions of the output signal parameter when a defect is
present (p(y|x1)), and without a defect (p(y|x0)). ................................................................... 189
Figure 9.4 Application of the functional model to obtain the output signal parameter
conditional pdf. Adapted from [180] and [181]. .................................................................. 193
Figure 9.5 Output phase angle, 6, over time for a valve with an 108 (solid line), and one with
a SLS (dashed line). .............................................................................................................. 195
Figure 9.6 Output phase angle, 6, as a function of z, the distance between the gradiometer and
outlet strut, for a valve with an IOS (solid line), and one with a SLS (dashed line). ........... 196
Figure 9.7 Original and model generated outputs for the 108 case ...................................... 196
Figure 9.8 Original and model generated outputs for the SL8 case ..................................... 197
Figure 9.9 Example of a Gaussian distribution for the input pdf (2) .................................... 197
Figure 9.10 PDF of 6105 for heart valve 71981 with an 10$ (2: 2.5 mm) ........................... 198
Figure 9.11 PDF of 6515 for heart valve A4865 with a SLS (2: 2.5 mm) ............................ 198
Figure 9.12 PDF for 6103 (solid line) and 63Ls (dashed line), (2 = 2.5 mm) .......................... 199
Figure 9.13 PDF for 6105 (solid line) and 631,3 (dashed line), (z = 12 mm) ........................... 200

Figure 9.14 Magniﬁed view of the overlapped region between the two output phase angle
pdfs (z = 12 mm)........... ...................................................................................................... 201

Figure 9.15 POD with respect to the distance between the gradiometer and outlet strut ..... 201

xix

Figure 10.1 Geometry of the axisymmetric ﬁnite element model. The vertical line at the
center is the axis of symmetry. (See Figure 10.2 and Figure 10.3 for more details.) .......... 208

Figure 10.2 Conﬁguration and dimensions of the heart and the heart valve ﬂange ring at the
center of the ﬁnite element model. ....................................................................................... 208

Figure 10.3 Cross-sectional dimensions of the ﬂange and suture rings used in the ﬁnite
element model. ...................................................................................................................... 209

Figure 10.4 Finite element analysis results showing the spatial temperature elevation proﬁle
of the BSCC heart valve, suture ring and the surrounding biological materials after 180
seconds of continuous excitation. ......................................................................................... 211

Figure 10.5 Finite element analysis results at t = 180 8 showing localized temperature
elevations around the outer “rim” of the ﬂange ring. ........................................................... 212

Figure 10.6 Magniﬁed view of the temperature elevation proﬁle on the BSCC heart valve at t
= 180 s. The location of the peak instantaneous temperature (Tm. = 310.002953 K) is as
indicated. ............................................................................................................................... 213

Figure 10.7 Finite element analysis results showing the BSCC heart valve temperature over
180 seconds of exposure to the gradiometer-based examination system (observed at the peak
temperature point as indicated in Figure 10.6). .................................................................... 214

Figure 10.8 Finite element analysis results showing the last temperature peak in a l80-second
simulation at t = 177 s, with Tm = 310.002983 K. .............................................................. 215

Figure 10.9 Spatial temperature elevation proﬁle of the BSCC heart valve, suture ring and the
surrounding biological materials after 177 seconds of continuous excitation. ..................... 216

Figure 10.10 Magniﬁed view of the spatial temperature elevation proﬁle on the BSCC heart
valve at t = 177 s. The point in the ﬁgure indicates the location of the peak instantaneous
temperature (Tm = 310.002983 K). ..................................................................................... 21.7

Figure 10.11 Finite element analysis results showing the BSCC heart valve temperature
(observed at the peak temperature point indicated in Figure 10.6) over 15 minutes of
exposure to the excitation magnetic ﬁeld ............................................................................. 218

Figure 10.12 Spatial temperature elevation proﬁle of the BSCC heart valve, suture ring and
the surrounding biological materials after 900 s of continuous excitation. To = 310 K. ...... 219

Figure 10.13 Magniﬁed view Of the spatial temperature elevation proﬁle of the BSCC heart
valve att=900 s. T0=310 K. ............................................................................................. 220

Figure 10.14 Spatial temperature elevation proﬁle of the BSCC heart valve, suture ring and
the surrounding biological materials after 15 minutes (t = 900 s) of continuous excitation
without blood circulation (convective cooling). ................................................................... 222

XX

 

 

 

v‘

 

 

 

 

 

1
l".
..

 

 

a.

 

 

 

 

Figure 10.15 Three dimensional view Of the proﬁle from Figure 10.14. ............................. 223

Figure 10.16 Temperature elevation from inductive heating (with no blood ﬂow) is spatially
localized to a region of 4 cm radius from the center of the ﬂange ring. ............................... 224

Figure 10.17 Magniﬁed view of the spatial temperature elevation proﬁle of the BSCC heart
valve at t = 900 s with no convection cooling. The peak of the instantaneous temperature is

located at the center of the heart valve cross-section. The maximum temperature rise is AT =
0.0652 K. (To = 310 K). ....................................................................................................... 225

Figure 10.18 Finite element analysis results showing the BSCC heart valve temperature over
180 seconds without convective and conductive heat transfers ............................................ 226

Figure 10.19 Measurement point, P, for the second special case (II) temperature study ..... 227

Figure 10.20 Finite element analysis results showing the BSCC heart valve temperature
without convection and conduction cooling over 900 seconds of continuous exposure to the
excitation magnetic ﬁeld. ...................................................................................................... 229

Figure 10.21 Magniﬁed view of the spatial temperature elevation proﬁle of the BSCC heart
valve without convection and conduction cooling at t = 900 8. To = 310 K. ....................... 230

Figure 10.22 Temperature proﬁle inside the heart valve at t = 900 s after a continuous
exposure to the excitation signal without conductive and convective cooling. To = 310 K.
(Note that the temperature variation is too small, < 10e-4 degrees, for proper labeling of the
temperature scale) ................................................................................................................. 231

Figure 10.23 Temperature elevation proﬁle at the beginning of the simulation (1 = 0.01 8)

showing the effects of numerical noise. The maximum temperature rise at this instance is AT
= 4.21x10’5 K. ....................................................................................................................... 232

Figure 10.24 Temperature of the BSCC heart valve with normal (solid line) and
special/worst case (broken line) cooling conditions over 15 minutes. To = 310 K ............... 233

Figure 11.1 Amplitude of the magnetic ﬂux density is the largest at the center of the plane
positioned mid-way between the two excitation coils. ......................................................... 238

Figure 11.2 The average value of the electric ﬁeld intensity, E, along path LE is used for SAR
calculation. ............................................................................................................................ 242

Figure 11.3 The magnitude of the electric ﬁeld, he], along path LE (z=0, p 00549-00599 m)
for the finite element model in Figure 11.2. ......................................................................... 243

Figure 12.1 A simpliﬁed cross-section view of the human heart. After [135] .................... 269

Figure 12.2 During systole, the ventricles contract. The AV vales are closed and blood in the
right ventricle is forced into the pulmonary trunk while blood in the left ventricle is forced
into the aorta. After [136] ..................................................................................................... 270

xxi

Figure 12.3 During diastole, the ventricles relax. The AV valves are pushed open and blood

ﬂows from the atria to the ventricles. After [136] ............................................................... 270
Figure 12.4 Mitral stenosis is the narrowing of the valve opening that reduces blood ﬂow
from the atrium to the ventricle. After [137] ....................................................................... 273
Figure 12.5 In mitral regurgitation, the improper closing of the valve causes blood to leak
back into the atrium from the ventricle during systole. After [137] .................................... 273
Figure 12.6 Starr-Edwards Silastic Ball Valves. Top: Mitral Model 1260. Bottom: Aortic
Model 6120. .......................................................................................................................... 278
Figure 12.7 BjOrk-Shiley titling disc valve ........................................................................... 279
Figure 12.8 Kay-Shiley TGCD series single leaﬂet caged disc valve. ................................. 279
Figure 12.9 Medtronic-Hall single tilting disc valve. ........................................................... 280
Figure 12.10 Omniscience single tilting disc valve. ............................................................. 280
Figure 12.11 St. Jude standard bileaﬂet valve. ..................................................................... 280
Figure 12.12 Carbomedics CPHV standard bileaﬂet valve. ................................................. 281
Figure 12.13 On—X bileaﬂet valve. ....................................................................................... 281
Figure 12.14 Hancock porcine valve. ................................................................................... 282
Figure 12.15 Carpentier-Edwards porcine valve (two views). ............................................. 282
Figure 12.16 St. Jude Toronto Stentless Porcine Valve (SPV) ............................................. 282

Figure 12.17 Freestyle Stentless Aortic Bioprosthesis: End view (left) and lateral view
(right). ................................................................................................................................... 283

Figure 12.18 Carpentier-Edwards pericardial valve: Lateral view (left) and top view (right).
............................................................................................................................................... 283

Figure 12.19 The Ross procedure. The numbers depict the surgery sequence. After [143]284

Figure 12.20 The custom surgery table for the gradiometer-based heart valve examination

system ................................................................................................................................... 287
Figure 12.21 Table top of the custom surgery bed is positioned between the excitation coils

through a side opening during examination .......................................................................... 288
Figure 12.22 Optimal position of the gradiometer inside the outlet strut loop ..................... 289

xxii

Figure 12.23 The gradiometer should be placed as close to the outlet strut loop as possible
without crossing the plane of the loop .................................................................................. 289

Figure 12.24 Optimal position of the gradiometer in relation to the outlet strut .................. 289

xxiii

BSCC
EM
EMF
IOS
OSF
RF
SAR
SLS

WBA

KEY TO ABBREVIATIONS

BjOrk-Shiley convexo-concave
Electromagnetic
Electromotive force

Intact outlet strut

Outlet strut fracture

Radio frequency

Speciﬁc absorption rate
Single-leg separation

Whole-body-average

xxiv

CHAPTER 1. INTRODUCTION

 

Drawing of the Heart and its Blood Vessels, from the Anatomical Notebooks
Leonardo da Vinci (1452-1519)
(Image courtesy of The Arts Renewal Center, www.artrenewal.org)

It is known that the heart, in every-day conversation, is synonymous

with love and life, but it also has many other associations. For example
‘have a heart' means to be merciful, and ‘have your heart in the right
place' means to be kind, while 'heavy heart' refers to sadness.

No other bodily organ elicits this kind of response. After all, when was the
last time you had a heavy pancreas?

Adapted from
David Henry Pinney (http://www.pinney.co.uk)
and http://www.howstuffworks.com/heart.htm

The BjOrk-Shiley Convexo-Concave (BSCC) prosthetic heart valve was one of the

commonly implanted prosthetic heart valves1 until its withdrawal from the market in 1986.
Between 1979 and 1986, about 86,000 BSCC valves were implanted worldwide. The valve
employs a pyrolytic carbon disc as an occluder to ensure unidirectional blood ﬂow from the
atrium to the ventricle of the heart, as well as from the ventricle into the aorta. The disc is
held in place by two struts. The strut on the inlet side is integral to the ﬂange ring while the
strut on the outlet side is tungsten inert gas welded to the ring. Occasionally, after years of
use, the weld at one of the two sites fractures, resulting in a single-leg separation (SLS) of the
strut [16]. This single-leg separation causes the load on the weld at the other end of the strut
to increase, thereby setting the stage for it to fail eventually. If the outlet strut is severed
from the ﬂange ring, the disc can escape, leading to loss of blood ﬂow control and fatality in
most cases. There is, therefore, a considerable amount of interest and a sense of urgency in

developing techniques to detect SLS failures.

This thesis describes a minimally invasive approach, called the gradiometer-based method,
for detecting SLS failures in BSCC heart valves. The method employs two large excitation
coils placed on either side of the valve to generate a uniform, time varying magnetic ﬁeld in
the neighborhood of the heart valve. The coils are oriented in a manner to ensure the current
induced in the ﬂange ring is a minimum, while maximizing the current induced in the outlet
strut. The ﬁeld generated by the induced eddy currents in the outlet strut perturbs the
uniform ﬁeld established by the external excitation coils. The perturbation in the applied

ﬁeld, or more speciﬁcally the perturbation in the gradient of the ﬁeld, in the vicinity of the

 

1 . .
See Appendix A for a survey on the heart, heart valve drseases and types of treatments.

strut is measured using a catheter-mounted gradiometer. An SLS failure causes a reduction
in the current and hence the perturbation in the magnetic ﬁeld. The gradiometer signal can,

therefore, be analyzed to indicate SLS failures.

Finite element simulation models have been used to demonstrate the feasibility of the
gradiometer-base detection method as well as to determine the system parameters. Proof-of-
concept experiments have been conducted to validate the models ﬁndings as well as to
validate the gradiometer design. Finally, a clinical-scale system that is suitable for animal
and human subject studies has been fabricated. In Vitro double-blind tests of this system has

achieved 100% accuracy in detecting SLS failures.

1.1 Organization of the Thesis
This thesis is organized as follows: Chapter 2 introduces the history of the BSCC prosthetic
heart valve, the outlet strut failure problem, and a number of proposed techniques by other
research groups for detecting single-leg fractures of the valves. Chapter 3 describes the
gradiometer-based method for SLS detection. The mathematical formulations describing the
principles of the examination approach are shown in Chapter 4. Chapter 5 discusses the
ﬁnite element analysis results of the gradiometer-based method, including the generation Of
the excitation magnetic ﬁeld and the interaction of the heart valve and the outlet strut with
this ﬁeld. Chapter 6 describes the hardware components and speciﬁcations of an
examination system based on the gradiometer-based method, and Chapter 7 presents two
gradiometer designs that are suitable for clinical use. Chapter 8 discusses the system

performance characteristics and presents the in Vitro test results. Chapter 9 presents a

technique for developing the probability of detection model for the system. Chapter 10
examines the effects of inductive heating on both the patient and the heart valve due to the
magnetic ﬁeld of the system. Issues with patient and operator safety, as well as compliance
with applicable international health standards are discussed in Chapter 11. Chapter 12

presents a few concluding remarks and ideas for future works.

For background and supplemental information, Appendix A gives a physiological description
of the heart and the cardiac functions, the nature of heart valve diseases and their typical
treatment, and a brief survey on the kinds of prosthetic heart valves. Lastly, Appendix B

recommends a procedure for conducting animal study with the gradiometer-based

examination system.

1.2 Beneﬁts of Work
Successful development of the gradiometer-based detection method will provide patients and
physicians with a highly reliable tool for diagnosing single-leg separation (SLS) in BjOrk-
Shiley Convexo-Concave (BSCC) prosthetic heart valves. Knowledge of an intact valve will
provide immense comfort to the patient and his or her family. Patients whose heart valve has
a single-leg separated outlet strut can evaluate their options more objectively with their

physicians, as well as to better prepare for emergency responses in case if a complete outlet

strut fracture occurs.

CHAPTER 2. THE GRADIOMETER-BASED METHOD

The gradiometer-based method is a minimally invasive catheter-based method. In this
approach, two large excitation coils are placed on each side of the heart valve outside the
patient’s body to generate a uniform time-varying magnetic ﬁeld around the valve as shown
in Figure 2.1. Eddy currents are induced in the various parts of the valve according to
Faraday ’8 Law. The ﬁeld produced by the external coils is oriented in a fashion that is
parallel to the valve ﬂange ring. This orientation ensures that a minimum amount of eddy
current is induced in the ring, and that the current induced in the outlet strut is dominant. The
ﬁeld generated by the induced eddy currents in the outlet strut perturbs the uniform ﬁeld
established by the external excitation coils. The perturbation in the ﬁeld (or rather, the
gradient of the ﬁeld) in the vicinity of the strut is measured using a catheter-mounted
gradiometer. The presence of an SLS reduces the induced current and the perturbation of the
applied uniform ﬁeld. The gradiometer signal, therefore, can be used as a basis to determine

the condition of the outlet strut.

2.1 Faraday’s Law and Maxwell’s Equations
In the gradiometer approach, the operating principle of the gradiometer as well as the eddy
currents induced in the heart valve are both described by the Faraday ’8 law of Induction.
Faraday’s Law states that a time-varying magnetic ﬁeld will produce an induced voltage,

called the electromotive force (EMF), in a closed circuit, which causes current to ﬂow. The

induced EMF, VEMF (in volts), is related to the time rate of change of the magnetic ﬂux

linkage by the circuit. This relation is expressed as

d/l__ E

V =-—_ N 2.1
EMF dt - dt ( )

where it is the magnetic ﬂux linkage, N is the number of turns in the circuit and ¢ is the ﬂux
through each turn. The induced voltage acts to oppose the effect of the applied magnetic
ﬁeld as indicated by the negative sign in equation (2.1). This is known as Lenz’s Law: the
induced current in the circuit ﬂows in the direction such that the induced magnetic ﬁeld

produced by the induced current opposes the original magnetic ﬁeld.

Excitation

\l 1,; is: (- tutu:

 

 

    

I 7 E

  

 

1
.4.‘ b1

l

L

 

 

 

 

 

 

 

 

 

Patient
l l Catheter-mounted
Gradiometer

Figure 2.1 Illustration of the gradiometer-based approach for BSCC heart valve SLS
detection (components not drawn to scale)

it

E/ L

dl I

 

 

 

 

 

Induced B
Figure 2.2 Induced EMF due to a time-varying magnetic ﬁeld, B. After [2]

In Figure 2.2, the amount of magnetic ﬂux linking the circuit (the circular loop) is given by

¢ = l5 B -dS (2.2)

where B is a time-varying magnetic ﬁeld and S is the surface area of the circuit bounded by

the closed path L. For a single-loop circuit (i.e., N = 1), (2.1) can be written as

d BB
VEMF =—EISBOdS——Isg.ds (2.3)

From Figure 2.2, the VEMF is also given by

VEMF = {LE ° ‘1' (2.4)

where E is a time-varying electric ﬁeld and d1 is the incremental path length. Applying

Stokes’s theorem to (2.4) and equating to (2.3) gives
JS(VXE)-dS=— —-dS (2.5)

Comparing the integrands yields

6B
V E = ——

which is one of Maxwell’s equations for time-varying ﬁelds. Equation (2.6) describes the

relationship between the induced electric ﬁeld in the circuit and the applied magnetic ﬁeld.

A detailed discussion on the Maxwell’s equations is beyond the scope of this thesis. The

reader is referred to references such as [ll-[3] for a comprehensive treatment of the subject.

In the gradiometer approach, the heart valve is submerged in a time-varying uniform
magnetic ﬁeld. The valve is oriented such that the ﬂange ring is parallel to the applied ﬁeld.
In this manner, the ﬂange intercepts the least amount of the applied ﬁeld as shown in Figure
2.1. Meanwln'le, the outlet strut protrudes away from the plane of the heart valve and forms a
closed loop with the ﬂange. This outlet strut loop intercepts the applied magnetic ﬁeld and
an EMF is induced in the loop according to (2.1). The induced EMF causes a current to ﬂow
in the outlet strut, which in turn creates a magnetic ﬁeld that perturbs the applied excitation
ﬁeld as illustrated in Figure 2.2. The net perturbation in the vicinity of the outlet strut is
measured using a gradiometer. The gradiometer consists of a pair of sensor coils separated
by a known distance. When subjected to a time-varying magnetic ﬁeld, a current is induced
in each coil, again, according to Faraday ’3 law. The difference of the coil signals provides a
measurement of the gradient of the ﬁeld in the vicinity of the outlet strut. In the case of an
108, the loop formed by outlet strut and the ﬂange is complete and the amount of induced
current is large. The perturbation of the uniform ﬁeld is correspondingly large and the
gradiometer output signal is strong. In the presence of a SLS, the loop formed by the strut
and the ﬂange is either partially or completely opened, resulting in little or no current ﬂow.
Consequently, the perturbation on the applied ﬁeld is small. The gradiometer output signal
level in this case will be very weak or non-existent. The gradiometer signal can, therefore,

be analyzed to indicate a SLS failure.

 

2.2 Implementing the Gradiometer-Based Method
The implementation of the gradiometer-based method begins with the development of a
mathematical model for the approach. Finite element analysis is also performed to
demonstrate system feasibility before construction commences. The components of the
clinical scale prototype—the gradiometer—based examination system—are then identiﬁed and
their operating parameters speciﬁed. Gradiometer prototypes will be evaluated in the clinical
scale system. Component designs and parameters are reﬁned until the system is Optimized.
In vitro tests will then commence with actual heart valve specimens. The in vitro test results

will be used to demonstrate the accuracy of the gradiometer-based method.

The development of the gradiometer-based examination system and the system test results

are described in the following chapters.

 

 

 

 

 

 

 

V V V V V

Figure 2.3 Induced current in the outlet strut by the excitation ﬁeld in turn creates a magnetic
ﬁeld which perturbs the original excitation ﬁeld (magnetic ﬂux lines simpliﬁed for clarity;
components not drawn to scale)

 

 

>4

CHAPTER 3. BJORK-SHILEY CONVEXO-CONCAVE

PROSTHETIC HEART VALVE

The original BjOrk-Shiley tilting disc valve was co-developed by Dr. Viking O. BjOrk of
Sweden and Shiley, Incorporated in USA [4]. It was ﬁrst introduced clinically in 1969.
This original model was reliable, durable, and had good hemodynamics and a low rate of
thromboembolism [5], [6]. This model was then modiﬁed to further decrease the
thromboembolism risk by further improving the hemodynamics of the valve [5], [6]. The
end result was the introduction of the convexo-concave models ([7]-[9]) with the tilting
disc occluder opening at 60° and 70° angles, respectively. Figure 3.1 shows a BjOrk-
Shiley Convexo-Concave (BSCC) 60° valve. The 70° valve had a similar design with the
exception that the disc opened at a 70° angle. Figure 3.3 shows drawings of a BSCC
Valve with the occluder disc in the open (top ﬁgure) and closed (bottom ﬁgure) positions.
The Opening angle is labeled 0 in this ﬁgure. The implantation of the BSCC 60° heart
Valve started in 1976 with the commencement of the clinical trials. The valve was
eventually approved by the United States Food and Drug Administration (FDA) in April
1979, and was implanted world-wide until 1986 when the manufacturer withdrew it from
the market [10]-[16]. The 70° valve was sold only outside of the US. between 1980 and
1983. Since only the 60° valve models were implanted in the US, the discussions in this
thesis will be centered primarily on this model, which will be referred to simply as the
BSCC valve in the remaining chapters. It is important to note that although the

di . ‘ . . .
Sclllssrons are focused on the 60° convexo-concave valve, the work described in this

t .
hesls is nevertheless equally applicable to the 70° valve.

10

¥

1,, .. Occluder disc

  
   

Outlet strut

Flange ring v \ X /

Figure 3.1 ijrk-Shiley Convexo-Concave (BSCC) 60° heart valve. The 70° valve has a
similar design except the occluder disc opens at a 70° angle

 

 

Figure 3.2 A BSCC heart valve with the woven Teﬂon fabric sewing ring

3.1 Outlet Strut Fracture of the BSCC Valve
The BSCC valve consists of a ﬂange (oriﬁce) ring, an inlet strut, an outlet strut, an
occluder disc (Figure 3.1, Figure 3.2) and a fabric sewing ring (Figure 3.2, Figure 3.3).
The inlet strut and the ﬂange are manufactured as an integral unit from a cobalt-based
Haynes-25 alloy bar stock. The outlet strut is fabricated from a wire of the same alloy
and is tungsten inert gas (TIG) welded1 to the ring. This welding method melts the ends
of the outlet strut legs and the neighboring ﬂange material to form solid welds without

using any ﬁller material [29]. The sewing ring is made from woven Teﬂon fabric.

 

l
ungsten Inert Gas welding (TIG-welding) is also known as Gas Tungsten Arc Welding
( TAW).

11

¥

 
   
   
  

Woven Teflon
fabnc

i
.
,,
'i
, .
VA.“

L—lnlet strut “"‘ Flange

Outlet strut
Mitral
sewing ring

      

  
  
 

Disc (orifice)
Dacron anchoring sutures
Mitral Disc Well
sewing ring
Pyrolite ®
-- .~ Coating

 

 

. L Radiopaque
Graphite Substrate Tantalum Ring

Figure 3.3 Drawings of a BSCC heart valve showing the disc in the open (top) and closed
positions (bottom). The disc opening angle 0 is either 60° or 70°. Adapted from [17]
The occluder disc, made by CarboMedics, Incorporated (Austin, Texas), has a graphite
Col'e covered with a Pyrolite® coating. The disc is held in place by the inlet and outlet
Struts. The occluder disc has an embedded tantalum ring to provide radiopacity [17].
This marker ring has a gap to allow for thermal expansion during the Pyrolite coating
p r 0Cess. The disc is inserted into position between the struts by manually bending the
outlet strut and setting the hook-to-well gap distance on the convex surface of the disc.

The valve is then checked to ensure proper clearance around the perimeter of the disc

12

when opened and closed. The opening translation and rotation of the disc are limited by
the hook-shaped central tip of the inlet strut. On closing, pressure gradient across the
valve causes the disc to return to and rest on the inlet strut’s convex surface, thus
preventing backﬂow into the inlet oriﬁce. This ensures unidirectional blood ﬂow through
the valve. The valve was manufactured in odd millimeter sizes from 17 mm to 33 mm.
The two smallest sizes (17 and 19 mm) valves were only available in the 70° design. The
three largest size (29, 31 and 33 mm) models all had the same 29-mm ﬂange ring and

disc size. They were labeled according to the outside diameter of the fabric sewing rings.

3.1.1 Loading of the Outlet Strut

The human heart beats an average of 37.8 million times a year. During each cardiac
cycle, the occluder disc impacts the arresting struts and the ﬂange ring with a force as
much as 2200-2600 grams [l7], [19]-[22], [70]. Stress loading and other factors
Sometimes result in fatigue fracture [24]-[27] of one of the two outlet strut legs, a
c0ndition known as a single-leg separation (SLS) of the strut. Figure 3.4 and Figure 3.5
Show two scanning electron microscope (SEM) images of a fractured outlet strut as view
from the outlet and inlet side, respectively. This single-leg separation causes the impact
10ad on the other leg of the strut to increase, thereby setting the stage for it to fail
evﬁntually. When both legs of the outlet strut break, the occluder disc escapes. This
leads to a sudden massive valvular incompetence—a loss of blood ﬂow control and
p r eSsure to maintain circulation by the heart—as well as embolization by the occluder
disc [30]-[36]. The result is acute heart failure which requires immediate surgical

luteI‘vention [37], [38]. Figure 3.6 shows a chest radiograph of a patient whose heart

13

valve has fractured. In this ﬁgure, the outlet strut is ﬂoating near the valve ﬂange, and
the occluder disc has embolized the aortic arch. The mean time between the onset of
symptoms and fatality depends on the location of the valve—a few minutes in the case of
an aortic valve, and one to two hours for a'mitral valve [10]. Approximately 86,000
valves were implanted worldwide between 1976 and 1986, with about 4000 of the 70°
model. By the end of 2005, there have been a total of 656 outlet strut fractures reported

to the manufacturer [39]. Meanwhile, Shiley, Inc. has reported 272 cases of conﬁrmed

fatalities1 related to outlet strut fractures, with an additional 167 alleged fracture-related
deaths [39]. The total alleged and conﬁrmed fatalities constitute approximately 66%, or
two-third, Of all the outlet strut fracture cases reported. This number is believed to be
underestimated as many cases may have gone unreported or misdiagnosed. Misdiagnoses
of outlet strut fractures (OSFs) are mostly due to the fact that the symptoms of a valve
failure resemble closely to other forms of acute heart failures. A valve failure-related
fatality can only be detemrined by performing an autopsy. As of July 2004, there is an
eStimated 22,200 surviving BSCC patients [39] world-wide. There is, therefore, a
cOnsiderable amount of interest and a sense of urgency in developing patient care

Strategies to prevent future valve-related fatalities.

1\
t This is the total number of alleged and actual events reported to Shiley as death related
0 the valve.

14

 

 

Figure 3.4 Scanning electron microscope (SEM) image of a fractured outlet strut viewed
from the outlet side. Adapted from [81]

 

Figure 3.5 Inlet side view SEM image of the fractured outlet strut shown in
Figure 3.4. Adapted from [81]

15

 

Figure 3.6 Chest radiograph of a patient with outlet strut fracture. The outlet strut has
separated and is ﬂoating near the ﬂange. The occluder disc (arrow) has embolized the
aortic arch. Adapted from [81]

3.2 Care for the BSCC Heart Valve Patients
COIlective efforts have been focused on three areas to reduce the number of fatalities due
to Outlet strut fracture in BSCC heart valve patients. The ﬁrst is to understand the
mechanics of OSFs in BSCC heart valves, [24]-[82]. The second is to identify factors
that may increase the risk of OSFs [47]-[68] and to recommend courses of actions [64]-
[68] to both BSCC heart valve patients and attending physicians. The third effort is to

de"elop screening techniques to detect single-leg separations (SLSs) ([71]-[118]) in situ

be fOre they evolve into complete outlet strut structures.

3.2.1 BSCC Heart Valve Outlet Strut Fracture Mechanisms

Early investigations of outlet strut fractures focused on the opening events of the valve.
As the occluder disc opens, the central hook of the outlet strut is designed to limit the
opening rotation and translation of the disc. The outlet strut arrests the opening disc and
sustains an impact force in excess of 2200 grams. The force also exerts a bending
moment on the strut. Meanwhile, high-speed photographs and accelerated wear tests
showed [17] that the disc closing impact loads are largely transferred to the inlet strut.

Consequently, the valve Opening forces were the focus of investigations in the early

1980’s.

Studies of outlet strut loading suggested that increasing the opening angle from 60° to
70° would reduce the strut opening load by almost 50% [17]. This fact, along with
hemodynamic advantages, prompted the introduction of the 70° valves by Shiley.
Ironically, the 70° valve models turned out to be more susceptible to OSF than the 60°
rI‘lodels of the same sizes [47], [59]. Meanwhile, as outlet strut fractures continued to
Occur, more valve specimens became available from clinical explants. Data collected
fI‘om examinations of the explanted clinical valves, along with the fact that the 70° valves
Were more prone to fracture, had shifted the focus of investigation to searching for and
Stndying the previously unanticipated transient and irregularly occurring closing events.
Metallurgical analyses [25]-[29] of the unused valve inventory, in Vitro studies in pulse
cluplicators ([19], [20], [22], [74], [97] and [104]) and in vivo studies with sheep under
both sedentary and exercising conditions had led to a better understanding of the valve

Qlosing dynamics and the outlet strut fracture mechanics [21], [86].

17

 

 

3.2.2 The Mechanics of Valve Closure
Examination of explanted BSCC valves showed, in some cases, multiple wear marks on
the outlet strut tip and leg bases and on the inner surface of the ﬂange ring. This

suggested that the occluder disc potentially has more than one closing position. To

understand and observe any transient closing events, Structural Acoustics, Incorporatedl
(SAI) had mounted miniature strain gages on the leg bases of the struts to measure
dynamic loading. Figure 3.7 shows the position of the disc in relation to the outlet strut
loading at the instance of closure measured in a pulse duplicator. During closure, the disc
initially rotates beyond its mean stable closed position. The over-rotated disc then comes
into contact with and loads the outlet strut. The impact loading of the outlet strut

Continues as the disc damps out to a ﬁnal closed position, coming to rest on the inlet strut.

 

  
  
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘ V U
.. OutletStrut M U
- $039.11” I I
o -
'U a
,3 .1
ET.
1% :1 DISC Rotation Mean Stable Closed Position
'5 III ._ _ __ __r, e r
m ..
qlJLlllllljlllLllJLUllllllJLlJllllllllll.lJlllJJll
0 2 4 6 8 10 12 14 16 18 20

Time (ms)

Figure 3.7 Relationship between outlet strut loading and disc rotation during valve
closure measured in a pulse duplicator. Adapted from [17].

\

1
Structural Acoustics, Incorporated, Raleigh, North Carolina, USA.

18

One cause of the outlet strut loading may be attributed to the manufacturing tolerance
standard. During valve assembly, the outlet strut was manually bent into place after it
was welded to the ﬂange ring. The metallic parts of the heart valve were then polished
and checked for tolerance compliance. Prior to 1984, the outlet strut was adjusted such
that there were gaps (Figure 3.8) between the disc and the base of the strut as well as
between the hook and the well of the disc (hook-to-well gap, Figure 3.9). An example of
outlet strut loading is shown in Figure 3.10. In this case, when the occluder disc over-
rotates, the inlet strut tip acts as a fulcrum for the disc to impact the outlet strut tip with a
vertical force F0 for less than 2 ms. This impact force creates a bending moment on the
outlet strut. This scenario was observed by SAI in their in Vitro study. Figure 3.11 shows
the load experienced by the inlet and outlet struts during a closure event with disc over-
rotation. In this ﬁgure, the inlet strut load begins to rise approximately 1 ms before over-
rotation occurs. The inlet strut load then drops dramatically as the disc over-rotates and
the outlet strut tip becomes loaded. After approximately 1.5 ms, the disc rebounds from
the outlet strut and in turn impacts the inlet strut. The oscillation of the disc continues as

it comes to rest at its ﬁnal closed position, suiking the inlet and outlet struts alternately in

the process.

 

  

 

 

\

Gap at Base and Tip

Figure 3.8 Prior to April 1984, the manufacturing tolerance speciﬁed gaps at the base and
tip of the outlet strut. Adapted from [17].

19

Hook to Well Gap

 

 

.— =: :1): =3 :2 := :2 1
. --==* ~ 1
S / t
Figure 3.9 Hook-to-well gap between the outlet strut tip and the ocluder disc well.
Adapted from [17].
TI and Disc Contact
/ P

    

 

 

 

Gap at Base of Outlet Strut

Figure 3.10 A simulated case of disc over-rotation. F0 is the vertical implact force on the
outlet strut. Adapted from [17].

 

 

 

 

 

l

1 r Outlet Load
ii 3 ' l
=3 4 —: .A J. A AA A. A. i
E Inlet Load

 

 

 

l1 .
a/jWWU—_

Jlllllll_LII_l LIIlllLlLlllJllIIIIIJIIIJJLILIJLIJJIJ

 

 

LLJ

 

Figure 3.11 Loading of the inlet and outlet struts during the occluder disc over-rotation.
Adapted from [104].
20

In April, 1984, Shiley changed the tolerance speciﬁcations and modiﬁed the procedures
for setting and verifying the hook-to-well gap. The new speciﬁcations eliminated the gap
between the disc and the strut base, allowing the rim of the disc to remain in contact with
the base of the outlet strut when the disc Was at rest (Figure 3.12). With this new
conﬁguration, a disc over-rotation would cause it to impact the outlet strut leg bases
rather than the strut tip, thus eliminating the vertical force, FC, in Figure 3.10 and the
subsequent bending moment exerted on the outlet strut. These modiﬁcations appear to be

effective as no outlet strut fracture has been reported in a BSCC valve manufactured after

April 1984.

Although the effect of the hook-to—well gap is a candidate for the cause of outlet strut
failures, the effect of the gap on valve closure dynamics is not yet completely understood.

Investigations are on-going [46] to evaluate the effect of hook-to-well separation on tip

loading of the outlet strut.

 

Gap at Tip

Contact at Base

Figure 3.12 After April 1984, the new manufacturing speciﬁcations ensured that the disc
remained in contact with the base of the outlet strut when at rest. Adapted from [17].

21

3.2.3 Dynamics of the BSCC Valve

With an unmodiﬁed hook-to-well closure clearance, ﬁnite element analyses performed by
SAI [17] showed maximum stress concentration on the inlet side near the base of the
outlet strut during tip loading as shown in Figure 3.13. Although this stress is
experienced for less than 0.5 ms, the loading force can reach between 2200 g to 2500 g

under sedentary conditions. This amount of loading exceeds the outlet strut’s endurance

limit of approximately 430x103 kPa (~62000 p.s.i). If critical loading were kept
continuous, valve fracture could occur in as few as three months (107 cardiac cycles)

[17].

The dynamics of BSCC heart valves in the mitral position had been studied extensively in
pulse duplicators [18], [20], [22], [23] and in sheep [19], [21] to determine the impact
forces on the outlet strut during valve closure, and to investigate the physiological factors
that affect these forces. Of particular interest were the valve responses under sedentary
and exercise conditions. Impact forces were measured using miniature strain gages
mounted on the bases of the outlet strut legs, and the left ventricular pressure was
measured using a catheter-mounted manometer positioned in the left ventricle of the
Sheep or pulse duplicator chamber. In the sheep study, an ECG was also recorded for

both synchronization and determination of the instantaneous heart rate.

In the heart valve dynamics studies, the systolic left ventricle (LV) pressure signal, LVP,
Was low-pass ﬁltered and differentiated to give the LV pressure rise, dP/dt (where P is the

pressure measured in mmHg, and t is time in seconds). LV dP/dt was used as an

22

k

   

indicator of the level of physiological activity: at the time of valve closure, an
anesthetized subject typically has a dP/dt of 450-750 mmHg/s; under sedentary
conditions, dP/dt ~ 750-1300 mmHg/s; and, in exercise conditions the dP/dt may range
from 2500-5200 mmHg/s. Note that these dP/dt valves are rough estimates—they are

highly dependent on the subject and its general physiological condition.

 

Figure 3.13 Finite element analysis of the outlet strut leg showing the stress distribution
due to tip loading by the occluder disc during over-rotation (o-max denotes the region of
maximum stress). Adapted from [17].

23

From pulse duplicator studies, Rau et a1. [20] observed that the impact forces on the
outlet strut legs vary linearly with the disc closing velocities. Meanwhile, Guo et a1. [18]

derived an equation for the leaﬂet closing velocity in a bileaﬂet valve. Based on these

results, Schreck et al. [21] suggested that

l
(impact force) ,3 ~ LV dP/dtclosure- (3.1)

and

2
(impact force) ~ LVPclosure- (3-2)

Equations (3.1) and (3.2) had been veriﬁed by in situ data collected from sheep studies.

In general, impact forces on the outlet strut legs vary greatly with cardiac activities.
Impact forces are signiﬁcantly higher under exercise conditions than those at rest. In
pulse duplicator studies, Chandran et a1. [22] reported an impact force of approximately
2000 grams (~ 4.4 lbs) with a disc closing velocity of ~ 2 m/s (~ 6.56 ft/s). Rau et al.
reported impact loads reaching 2800 grams at a LV dP/dt of 4000 mmHg/s, with closing
Velocities up to 2.0 m/s. In animal studies, Schreck et al. reported impact forces ranging
from 1780 grams (anesthetized) to 3250 grams (exercising) in sheep. These results
Suggest that a higher level of physical activity increases the loading of the outlet strut,
Which may increase the likelihood of strut fracture. Younger heart valve patients tend to
assume more active lifestyles, and their hearts are more likely to function at higher

maximum heart rates, increased stroke volumes, and more rapid pressure changes (dP/dt)

24

than those of older patients. These factors collectively subject a heart valve to more

frequent heavy loading, thus increasing the chances for an outlet strut failure.

3.2.4 Valve Closure Delay

As the differential pressure between the atrium and ventricle reverses during the start of
systole, the higher pressure in the ventricle causes a reverse ﬂow of ﬂuid through the
prosthetic heart valve. In a tilting disc valve such as the BSCC valve, the valve does not
close immediately to stop the back-ﬂow. Rather, the valve is closed sometime after the
onset of the reverse ﬂow when the disc is pushed by the current to the closed position.
As a result, the instant of valve closure varies as much as 20 ms comparing to a stationary
cardiac cycle, both at rest and under exercise [19]. Rau et a1. [20] observed that the
maximum amplitude of the disc over-rotation was larger with increased closure delay.
The arresting forces for the disc on the outlet strut legs was also larger with increased
delay—with a 20 ms closure delay approximately doubling the impact forces on the
outlet strut. This suggests that valve closure delay is highly correlated with the disc
Closing velocity, left ventricle pressure during closure, and the impact loads on the outlet
Strut legs. Incidentally, when one strut leg is separated, the disc may be arrested from
contact with the base, rather than the tip, of the outlet strut. This condition results in
small impact loads on the strut that are due to the acceleration forces of the disc and are

not correlated to the pressure at closure [20].

In 1995, Rau et al. reported that the outlet strut loading was symmetrical for both strut

legs in their studies. In the following year, Chandran et al. [21] reported that the impact

25

loads on the outlet struts were asymmetric, with the loading on one leg up to 25% larger
than that on the other. This result may explain the single-leg separation preceding a

complete outlet strut fracture.

Based on the relationships between component loads, closure delay, disc velocity, and

pressure peaks, non-invasive approaches may be developed to evaluate the condition of
the heart valve. One class Of such approaches to determine the state of the outlet strut is
by examining the closing sounds of the valve. These acoustic approaches are described

later in this chapter.

3.3 Metallurgical Analyses of the Heart Valve and the Outlet Strut Weld Sites
The high impact loading of the outlet strut raised suspicions on the integrity of the strut
legs and their welding sites to the ﬂange ring. Metallurgical analyses of the alloy and the
weld sites suggested that valve failures were due to material fatigue [24]-[29]. Wenzel et

a1. [29] reported that in their study, 22% of ﬁrst—to—fail leg separations and 17% of all
fractures initiated outside of the weld. Fatigue suiations were also observed in all the
fractured valves in the specimen collection, with evidence indicating that every fracture
originated from the inﬂow side. In all the complete OSF cases, the fracture surface on
one leg was heavily faceted and showed branching cracks while the facture surface on the
other leg appeared rounded and burnished. This suggested that one leg of the outlet strut
had failed before the other [25]. Welding metallurgy study concluded that no
metallurgical feature of any weld investigated could be identiﬁed as contributor to the

fracturing of the outlet struts [29].

26

Meanwhile, Xiao et a1. [28] suggested an electrochemical failure mechanism of the valve
in which the strut leg fractures at the less noble weld area (which includes the welding
point and the surrounding heat-affected zone on the strut leg and the ﬂange) due to stress
corrosion cracking and erosion corrosion in the oxygen-rich, chloride—containing arterial
blood. He suggested that the pyrolytic carbon occluder disc and Haynes 25 alloy valve

material formed a bimetallic couple, essentially a dynamic electrochemical device.

During the opening and closing of the valve, a current on the order of 106A ﬂows

through the metallic parts of the valve. This current and the potential difference between
the disc and the ring lower the pH of the ﬂuid immediately surrounding the less noble
weld region and break down the oxide ﬁlm that covers and protects the alloy. In this
slightly acidic environment, material from the weld region erodes away, resulting in

reduced strength of the weld region and renders it susceptible to mechanical cracking.

Although Xiao et a1. presented a highly plausible explanation into the failure of the
BSCC valve, their conclusions disagree with ﬁndings from other researchers in one key
area: by Xiao’s electrochemical model, the downstream part of the less-noble weld area
Should be most affected due to its proximity to the rapidly-ﬂowing blood electrolyte.
This implies fractures should originate from the outlet side of the welds. Nevertheless, as
mentioned earlier, Wenzel et a1. [29] presented evidence indicating that every fracture in

their study originated from the inﬂow side.

27

3.3.1 Risk Factor Analysis and Prophylactic Valve Replacement

Identiﬁcation of the BSCC heart valve failure risk factors can raise the awareness of the
high risk patients and potentially increase their survival chances in case of an outlet strut
fracture. Epidemiological studies in Europe with selected cohorts had provided some
estimates for the strut fracture risk. Results of these studies were reported from:

Sweden, 1989 [55]; the Netherlands, 1992 [47]; the Netherlands follow-up report, 1996
[40]; the United Kingdom, 2001 [53], and a UK follow-up study from 2002-2003 [54], of
which the results have not been published. Meanwhile, Walker et a1. [41] had also
conducted a United States and Canada case study on the nature of the valve fractures in

relation to the manufacturing characteristics and patient physiological conditions.

Results of applying actuarial methods to the epidemiological data have shown that large
valves (2 29 mm) have consistently been found at greater risks [47]-[53]. Mitral valve
implants are at least 4 times higher in fracture incidents than aortic implants [41], [47],
[55]. Valves with 70° openings are 6 times more likely to fracture than 60° valves, and
patients with 60° valves have a better survival rate than those with 70° valves in case of a
Strut failure [47]. Valves manufactured between 1981 and 1984 also have an elevated
failure risk [40], [41], [59], with those manufactured between 1981 and June 1982 having
the highest failure rate [50], [65]. In a United States and Canadian cohort study of 150
outlet strut fracture cases, Walker et a1. [41] suggested that the hook deﬂection test and
load deﬂection test used in the manufacturing quality control process might have
indirectly indicated the risk of outlet su'ut fracture in certain valves. The hook deﬂection

test was used from May, 1980 to January, 1982. In this test, a 7.0 kg weight was

28

suspended from the outlet strut and the movement of the strut was measured. In
February, 1982, the load deﬂection test replaced the hook deﬂection test for
manufacturing tolerance control. In the load deﬂection test, a 5.0 kg load was gradually
applied (loading), removed (unloading), and reapplied. The amount of deﬂection was
recorded against the applied load for two cycles of loading and unloading. The intent of
this test was to verify that there was no yield of the outlet strut at 4.0 kg of loading, or
less. However, Walker et a1. observed an increased risk associated with increased
cumulative deﬂection and the amount of permanent deformation of the outlet strut
resulting from the testing procedures. They suggested the increased outlet strut

ﬂexibility, as indicated by the deﬂection test results, as a risk factor for failure.

In general, younger patients tend to be of higher risk as they are more likely to be active.
In fact, patients under age 50 at the time of implant typically encounter a higher rate of
outlet strut fractures and a higher cumulative risk of fatality than those above 50 years of

age [47], [50].

For a BSCC heart valve patient, it is sometimes difﬁcult to weigh the beneﬁts and risks
of an elective valve replacement surgery since an outlet strut fracture is not a guaranteed
event for all BSCC heart valves. Although a SLS is a frequent precursor to OSF, not
every SLS case evolves into an OSF. The development duration from a SLS to an OSF,
if it indeed occurs, may last for two or more years [17], and is highly case dependent. In
cases where the patient is older or is suffering from chronic diseases, increased surgical

morbidity risk may prohibit a valve replacement surgery. On the other hand, for a

29

younger and otherwise healthy patient, a prophylactic valve replacement operation may
increase his life expectancy as well as quality of life. Many researchers have attempted
to develop guidelines for prophylactic replacement of the BSCC valves [55]-[63], while
others have suggested patient care recommendations [64], [65]. Van Gorp et a1. [62]
devised a clinical prediction rule for the 30-day mortality in BSCC valve replacement
surgeries. In 1997, the Bowling-Pﬁzer Supervisory Panel issued a set of guidelines for
assessing patients with ijrk-Shiley convexo-concave heart valves for elective
explantation [66]. The guidelines were revised in 2002 [67] and again in 2003 [68]. The
Supervisory Panel document provided guidelines for comparing the risk of valve fracture
with the risk of reoperation to replace the valve. The risk of fracture is estimated based
on factors such as valve size, implant position, speciﬁc manufacturing characteristics,
patient gender and current age. This risk is estimated differently between the 60° and 70°
valves. The risk of reoperation, on the other hand, was derived from the published cohort

studies data and actual BSCC valve reoperation experiences.

Along with the guidelines, the Supervisory Panel also recommends that all BSCC heart
Valve patients should regularly consult their physicians. A patient and his relatives and
Close ﬁiends should have a good understanding of the symptoms related to an OSF, as
well as knowledge of the nearest medical facility with signiﬁcant experience in
cardiovascular surgery in case if an emergency valve replacement is needed. The best
preventive measure for the patient is early SLS detection and subsequent aggressive

monitoring. This thesis presents a minimally invasive electromagnetic approach that is

30

best suited for diagnosing SLS conditions. In case Of an OSF, prompt recognition and

surgical intervention can be lifesaving.

3.4 Previous Approaches in BSCC Heart Valve SLS Detection
Since an outlet strut fracture of a BSCC heart valve is typically preceded by the
separation of one of the two outlet strut legs, there have been world-wide efforts [71]-
[124], with a sense of urgency, to develop reliable in vivo techniques to detect SLS
failures in BSCC valves. Such techniques all aim to provide improved patient-level data
to facilitate informed diagnostic and therapeutic decisions. At present, cineradiography is
the only clinical diagnostic tool for detecting SLS in BSCC heart valve patients despite
the method’s less than satisfactory success rate [75]-[84]. Other approaches under
development are based on modalities such as echocardiography, acoustics, ultrasound,
electromagnetic acoustic transduction, and several electromagnetic methods based on

eddy current phenomenon.

3.4.1 Radiographic Detection Approach

High-resolution, high energy (80-125 kVp) cineradiography has been used to evaluate the
outlet strut condition of the BSCC heart valve [75]-[88]. Radiographs from two imaging
proﬁles are obtained for each valve: a tunnel or en face view (the valve inclined 30°
from the horizontal plane of the ﬂange) as shown in Figure 3.14, and the left or right 60°
oblique view as shown in Figure 3.15. If necessary, radiographs from the complementary
60° oblique view may also be used. The cineangiographic images are captured at

approximately 30 frames per second for 3 to 4 seconds. The images are then visually

31

evaluated by an expert review panel for SLS cues: gap and offset on the outlet strut. Gap
refers to the Opening when the outlet strut leg separates from the ﬂange, and offset refers
to the deviation of the strut leg from its central axis of insertion (e.g., see Figure 3.5).

The expert reviewers rate each radiograph on a 0 to 5 grade according to Table 3.1.
Although the review panel typically consists of board-certiﬁed experts in cardiology,
cardiothoracic surgery, and radiology, the assessment of radiographs tend to be both
reviewer and valve speciﬁc. In one phantom model study [84] of 6 manufactured SLS
and 3 IOS valves, 3 of the fractured valves were either not detected or rarely detected.
The study also reported a high false positive rate and increased human fatigue errors after

viewing more than 120 radiographs.

 

Figure 3.14 Cineradiographic tunnel view of a BSCC heart valve showing a case of
“deﬁnite SLS” assessment (arrow). Adapted from [81].

32

 

Figure 3.15 Cineradiographic oblique view of a BSCC heart valve showing a case of
"deﬁnite SLS" assessment (arrow). Adapted from [81].

Table 3.1 Deﬁnition of Image Assessment Grades for Visual Interpretation of

 

 

 

 

 

 

 

 

Cineangiographs
Image Assessment . .
Descrr tron
Grade p

0 Suboptimal examination: insufﬁcient image contrast or
strut base display

1 Apparently normal

2 Minimally suspect: fractured appearance in 1 or 2 frames
in one View only

3 Suspicious: fractured appearance in several frames in one
view, or in 1 or 2 frames in two views

4 Probable fracture: fracture appearance in multiple frames
in 1 view, or several frames in 2 views

5 Deﬁnite fracture: fractured appearance in multiple frames
in more than 1 view

 

 

 

33

In a clinical study [81] at the William Beaumont Hospital at Royal Oak, Michigan, 315
BSCC heart valve patients were examined with a cineradiograph system. Each patient is
examined 2 or 3 times over a 9-month period. The radiographic assessments were later
conﬁrmed in 21 cases by physical inspection of the valves, available due to explantation
or autopsy. Ten of the eleven valves radiographically rated as “probable” or “deﬁnite
SLS” were conﬁrmed, with one “probable” case being a false-positive. On the other
hand, two patients who had negative radiograph assessments experienced strut
fractures—both patients died. It was not known whether SLS conditions existed in either

or both of the cases during the Cineradiographic examinations

In a follow—up study [82], the trials at William Beaumont Hospital were expanded to the
Stanford University Medical Center and the Western Inﬁrmary in Glasgow, Scotland.
From both studies, a total of 964 BSCC valves were imaged (828 mitral valves, 136
aortic valves). Twenty-six SLS valves were correctly identiﬁed. In addition, there were

4 false-positive and 1 false-negative assessments.

In order to reduce reviewer-dependency and inter-reviewer variability, many quantitative
approaches ([77], [80], [83], and [85]) have been developed for the automatic objective
and reproducible assessment of BSCC heart valve outlet strut cineradiographs. These
approaches employ image processing techniques such as temporal registration, object
translation and rotation correction [77], segmentation and principal axis transformation
[80], with quantitative metrics such as decrease in pixel intensity (DIPI) ratio, gap half-

width [83], normalized pit depth (NPD) and the depth-sigma ratio (DSR) [85]. In [83],

34

logistic regression was used to identify a statistical model for the prediction of SLS in the
post-processing of radiographic images. This approach has accuracy equal to that of a
panel of expert subjective reviewers. Meanwhile, Vrooman, et a1. [85] developed a
software package for processing BSCCheart valve Cineradiographic images. This
package had a reported accuracy of 100% with limited phantom studies. The complete

performance data of this package in clinical tests were not provided in the report [85].

The results from the phantom model studies and clinical trials suggest that although the
feasibility of SLS detection by cineradiography has been demonstrated, the accuracy of
the method has yet to be deﬁned [82]. The resolution and capacity of present x-ray
technology maybe inadequate for guaranteed strut fracture detectability [84]. Current
clinical Cineradiographic systems may detect as little as 25% of the single-leg separated
BSCC valves [86]. Visual radiograph assessment also has high inter-observer variability.
Extensive training in radiograph interpretation is, therefore, required for accurately
identifying SLS conditions [81], [86]. Although automatic quantitative assessment
algorithms and software packages appear promising, further development and

veriﬁcations are needed before such algorithms may be applied clinically.

At present, the Penn State Milton S. Hershey Medical Center in Hershey, Pennsylvania,

USA is the only Cineradiographic imaging center for BSCC heart valve patients. This

screening program is sponsored by the Bowling-Pﬁzer Supervisory Panel [89].

35

3.4.2 Echocardiography, Acoustics, Ultrasound and Electromagnetic Acoustic

Transduction Based Approaches

3.4.2.1 Echocardiographic Detection Approach

Two-dimensional echocardiography had been demonstrated in diagnosing complete
BSCC heart valve outlet strut fractures [71], [72]. In such cases, the dislodged echo-
dense occluder disc is often seen moving rapidly and chaotically in the left ventricle.
Beyond a diagnostic tool for OSFs, efforts to use echophonocardiography to detect outlet
strut single-leg separations had not achieved much success [73], [74]. The most recent
reported project involving the application of echocardiography on BSCC heart valves
was the work conducted by the Cleveland Clinic [89] to characterize three-dimensional
motion of the valve by computed tomography and echocardiography. This work was

supported by the Bowling-Pﬁzer Supervisory Panel.

3.4.2.2 Acoustic Detection Approaches

In a mechanical heart valve prosthesis such as the BSCC heart valve, the periodic
Opening and closing of the occluder disc generates a clicking sound that is audible over a
distance external to the patient. It has been postulated [90]-[122] that a signiﬁcant
alteration of valve properties such as obstruction of the occluder or fracture of the valve
would change the frequency content and intensity of the Opening and closing sounds.
Based on this assumption, much work has been focused on acquiring and analyzing the
transthoracic prosthetic heart valve sounds for the non-invasive diagnosis of the
functioning and. structural integrity of the valve. For instance, Nygaard et a1. [95]

investigated the optimal position for recording in situ heart valve sounds and determined

36

that the precordial area of the fourth and ﬁfth left intercostal space on the chest near the
sternum would allow the transmission of prosthetic heart valve sounds from the heart to
the chest surface with the lowest attenuation and a minimum interference from
respiration. Others [106] had also used the second intercostal space adjacent to the right
stemal border for recording aortic valve sounds. These recording locations and the
immediate surrounding areas on the chest have been commonly used for acquiring in situ
heart valve sounds described in the literature. Several of these reports are described in

the following sections.

The acoustic approaches for detecting SLS in BSCC valves may be divided into three
categories: the low frequency, resonant frequency, and acoustical data recognition
approaches. The low frequency approaches ([90]-[93], [96], [105] and [103]) examine
the frequency response of the heart valve from zero to approximately 1.5 kHz to identify
the condition of the outlet strut. The resonant frequency approaches ([94], [97], [98],
[99], [106] and [107]) determine the condition of the outlet strut by detecting its ﬁrst
natural resonant frequency. The acoustical recognition approaches apply signal
processing and pattern recognition techniques to process and classify a heart valve sound

recording as from either an 108 valve or one with SLS.

Stein et a1. [90] developed a low frequency approach by examining the opening sounds of
two BSCC heart valves in a pulse duplicator and suggested that the dominant frequency
of the intact valve was 234 Hz and that for the fractured valve (with a through-and-

through cut) was 78 Hz. Durand et al. [96] demonstrated a low frequency (0-500 Hz)

37

analysis approach combined with a pattern classiﬁer such as the K-means or Bayesian
classiﬁer for classifying the opening sounds both in a pulse duplicator (6 105, 6 SLS
valves) and in sheep (11 10$, 16 SLS valves). They achieved 83% correct classiﬁcation
for the in Vitro tests and 82% accuracy in a sheep study. In a separate report [105],
Durand et al. demonstrated anther low frequency outlet strut condition diagnosis
approach based on the ratio of the high frequency area to the low frequency area of the
averaged and amplitude—normalized opening sound spectrum. A cut-off frequency of 500
Hz was shown to generate ratios that provided the most robust separation of the two
classes of valves. They tested 4 SLS and 24 IOS valves and achieved 96.4% correct
classiﬁcations, with one SLS valve incorrectly classiﬁed. Meanwhile, Walker and
Scotten [92], [93] reported the feasibility of identifying fractured valves from their
closing sounds, and Sava et al. [103] reported that the main part of the energy in the in

vivo closing sounds was located below 2 kHz.

Eberhardt et a1. [69] and Chia [70] had separately performed ﬁnite element analyses of
BSCC heart valves with different outlet strut conditions. Their work suggested that the
fundamental resonance mode for an intact outlet strut was between 7000-8000 Hz. They
had also reported that the fundamental vibration frequency of a fractured minor strut with
a hinged joint was approximately 4000 Hz, while a strut with a non-contacting fracture
had a resonant frequency between 2000-2700 Hz. It is, therefore, feasible to determine

the condition of the outlet strut by detecting its fundamental vibration frequency.

38

With knowledge of the estimated resonant frequency ranges, Plemons et a1. [94]
demonstrated a technique for identifying IOS valves in time-windowed signals by
examining the area after the lst 95% of the energy spectrum. In another report [98],
Plemons and colleagues concluded from observations of sheep experiments that the IDS
resonant frequency remained constant under changing physiological conditions (e.g.,
exercise versus resting) and observation times, whereas the occluder disc tonals showed
signiﬁcant variations. They suggested that this feature of the [08 valve maybe used to
distinguish its resonant frequency from interfering disc tonals. In a companion report
[99], Plemons et a1. analyzed the closing sounds recordings from 24 conﬁrmed clinical
cases using spectral techniques. They selected six characteristic features of the ﬁrst
resonant peak of the intact valve in the time-windowed, ensemble averaged spectrum of
the recordings. Their technique correctly classiﬁed 13 of 14 intact valves, and 10 of 10

fractured valves.

Dow et al. [106] extended Plemons’ work and developed a time-frequency representation
(TFR) technique for detecting the presence or absence of the intact outlet strut tonal
component around the 7-8 kHz band. The technique applies overlapping time-windows
of approximately 5.5 ms to each closing event and calculates the spectral density function
for each'of these signals. These calculations are then repeated for all qualifying events in
a recording session and an ensemble average is obtained. The peak density within the
target frequency band is then selected as the TFR for this valve. In a test of 32 valve

recordings (16 intact, 16 SLS valves), Dow’s technique correctly identiﬁed all SLS

39

valves, and 11 of 16 IOS valves (accuracy = 84.4%). Dow suggested the technique as a

prescreening tool for radiographic diagnosis.

Without restricting the frequency range of interest, the acoustical recognition approach
applies signal processing and pattern recognition techniques in classifying the heart valve
sound recordings. For example, Eberhardt et a1. [96], [104] examined transthoracic
recordings of BSCC heart valve closing sounds from patients and applied statistical
techniques to process the recordings. They applied the Volterra expansion to the
coefﬁcients of the processed data to identify characteristic features of the data, and used

an artiﬁcial neural network to classify the features as belonging to either an IOS or SLS

valve. In blinded and leave-one-out1 tests, Eberhardt’s technique classiﬁed 33 SLS and

IOS “gold standard” valves correctly [104].

Scarbrough et al. [100], on the other hand, applied a short-time Fourier Transforrn-based
approach to classify the closing sounds from the phonocardiograms of 22 BSCC (12 IOS,
10 SLS) valves. The approach was shown to classify all but two of the processed signals

correctly. The misclassiﬁed signals belonged to an IOS and SLS valve, respectively.

Candy et al. [101], [102] suggested two elaborate acoustic signal processing systems for
classifying heart valve Opening sounds recordings. The ﬁrst technique represented the
data in terms of spectrograms that displayed the individual sounds as a function of heart

beat number and temporal frequency. This spectral technique extracted potentially

 

1 . . . . .
In a leave-one-out test, the valve being tested is not used 1n the trarnlng data set.

40

discriminatory features of each signal that were subsequently processed by a statistical
classiﬁer for valve condition identiﬁcation. The second technique was a parametric
approach in which a heart valve sound generation model was constructed. Heart valve
closing sounds were synthesized and collected into a database. The data were then used
to develop both an optimal classiﬁer and an adaptive classiﬁer implemented with a
probabilistic neural network. In a test with 7 IOS and 10 SLS valves, the optimal
detector was able to correctly classify 15 of 17 valves (88.2%), misidentifying one IOS
and one SLS valve. The adaptive probabilistic neural network was able to classify 14 of

17 valves correctly (82.4%), misidentifying 3 SLS valves.

Acoustic diagnosis approaches generally suffer from the following practical issues. First,
the heart valve sound is inﬂuenced by its surrounding biological structures which exhibit
frequency dependent sound transmission and resonance properties. Second, respiration
induces variance in transthoracic sound propagations from the valve to the chest surface.
Third, optimal sensor locations need to be determined. And ﬁnally, the time or frequency

domain processing windows used in the algorithms need to be deﬁned.

Although acoustic approaches present potentials in diagnosing SLS failures in BSCC
heart valve patients, the overall accuracy must be improved before they can be adopted
clinically as the implications for a patient with ether a false-positive or false-negative

diagnosis can be serious or even life-threatening.

41

3.4.2.3 Ultrasonic Detection Techniques

Active acoustic methods, including ultrasonic and electromagnetic acoustic transduction
techniques [108]-[122], have been proposed to improve the reliability of detecting BSCC
heart valve outlet strut fractures. ShaWkat et al. [108] devised a technique to assess the
condition of the heart valve by analyzing its ultrasonic “clicks,” or signatures, of the
detected signal. Rambod et a1. [109]-[111] proposed an Ultrasound Burst-Spectrography
(UBS) system in which bursts of two ultrasonic pressure waves at approximately 3.1
MHz were focused onto the outlet strut. The pressure waves exert forces on the strut and
cause it to vibrate. The frequency of one of the ultrasonic waves is swept from 3.101—
3.110 MHz such that the vibration frequency of the outlet strut varies from 1—10 kHz.
The responses are measured by a hydrophone positioned in the vicinity of the heart valve.
When the vibration frequency coincides with the fundamental resonance frequency of the
outlet strut, a strong response is detected. The frequency of this response may then be
used to infer the condition of the outlet strut based on the results of Chia [70] and
Eberhardt [69]. The UBS approach had demonstrated 100% classiﬁcation accuracy in
aortic-model pulse duplicator studies of 11 IOS, 6 explanted SLS and 2 manufactured

SLS valves.

Although these results appear to merit further development, Rambod’s approach is, at the
time of this writing, experimental and has yet to be developed into a clinical system. The
most important obstacle of this approach is the in vivo focusing of the ultrasonic waves

onto the outlet strut. The movement of the heart within the thoracic cavity during cardiac

cycles and the hemodynamics inside the venuicle pose the challenging tasks of targeting

42

and tracking the outlet strut by the ultrasound waves, as well as maintaining the proper

focal distance between the transducer and the outlet strut.

Van N eer, et al. [112] proposed another ultrasonic fracture detection approach based on
the assumption that a broken outlet strut leg will have increased movement compared to
either the intact leg or the ﬂange. In Vitro studies of this approach were conducted in a
pulse duplicator using a 10 MHz ultrasonic transducer. Focused ultrasound waves were
directed at a particular leg of the valve. The movement of the outlet strut was obtained
by correlating the received signals with the excitation signal to produce the time delays of
the echoes. These time-of-ﬂight differences of the received echoes were then used to
indicate the amount Of movement of the target strut leg. Van Neer reported that the
amplitude of movement between both legs of an IOS valve was approximately 9.2 pm i
0.1 pm. In the case of an SLS valve, the amplitude of movement of the fractured leg was
12 pm i 1.6 pm while that for the intact leg was 8.6 pm i 0.1 pm. These results showed
that although there was a difference in movement amplitudes between an intact and a
fractured outlet strut leg, the difference was so small such that a deﬁnitive determination
on the outlet strut condition is difﬁcult. The damping of the fractured leg’s movement
also decreased the accuracy of the approach. Also, it is generally challenging to focus the
ultrasonic wave onto a speciﬁc location on the valve. There are also potentially high
levels of movements from other sources within the body that manifest themselves as

noise in the received signal.

43

3.4.2.4 Electromagnetic Acoustic Transduction Based Approach

Udpa et a1. [119]-[123] proposed a beat frequency electromagnetic acoustic transduction
(EMAT) based technique to actively excite the outlet strut for determining its
fundamental resonance frequency. The technique uses two pairs of excitation coils (coil
A and B) each excited by an alternating current source. The coils are placed around a

patient and oriented in the manner shown in Figure 3.16 (the patient is not shown in this

ﬁgure for clarity). Coil pair A, excited at frequency f], induces currents in the heart valve

with the same frequency. Coil pair B, excited at frequency f2, generates a magnetic ﬁeld

in the vicinity of the valve that interacts with the induced current in the heart valve. A

Lorentz force is produced and acts on the heart valve. The amplitude of the force is

proportional to the ﬂux densities of coils A and B and the frequency f]. The force has
two frequency components at f1 +f2 and f 1 - f2, respectively. The frequenciest and f2
are chosen in a manner such thath +f2 is signiﬁcantly above the dominant resonant

modes of any of the valve components, while the beat frequency f] - f2 is close to the

resonant mode of an intact or fractured outlet strut. In practice, coil A is operated in a

tone burst mode, while the frequency f2 is swept such that the beat frequency varies over

a range of approximately 10 kHz, coinciding with the range of expected outlet strut
modal frequencies for IOS and SLS valves. Transthoracic vibration recordings are made
during a short period soon after the excitation tone burst is switched off. Signal
processing algorithms are then applied to the recordings to extract the dominant

frequency and determine the condition of the outlet strut.

44

In a prototype unit, the beat frequency EMAT based approach demonstrated 100%
accuracy in classifying 34 BSCC heart valves (14 IOS, ll explanted SLS and 10
manufactured SLS valves). Although these results appear promising, the approach has
yet to be developed into a clinical-scale system. Successful completion of the clinical
unit may lead to a noninvasive alternative to the current prescreening and diagnostic

systems for SLS failures in BSCC heart valves.

Coil Pair A Operating at
Excitation Frequency f1

Head
Va Ive

 

Coil Pair B Operating at
Excitation Frequency f2

Figure 3.16 Orientation of the EMAT excitation system. Patient not shown for clarity.
Adapted from [121].
3.4.3 Electromagnetic Approaches
Evaluation techniques ([113]-[119], [l22]-[124]) exploiting the electromagnetic behavior
of a BSCC heart valve have emerged as potential tools for detecting a single-leg

separated outlet strut. Two of the techniques, proposed by Amin [118] and Lepelaars et

45

a1. [113]-[117], respectively, are described in the following sections. The remaining
chapters of this thesis describe the third and the most promising method called the

gradiometer-based examination approach.

3.4.4 Differential Eddy Current Approach

Amin [117] applied differential eddy current testing techniques to evaluate the condition
of the outlet strut in a BSCC heart valve. (A detailed treatment of differential eddy
current testing is beyond the scope of this dissertation. Readers are referred to references
such as [125] and [126] for more information.) In this approach, a differential eddy
current probe (Figure 3.17) is positioned in the vicinity of the outlet strut as shown in
Figure 3.18. Figure 3.19 shows the electrical circuit diagram of a differential eddy
current probe, which functions as both the excitation source and the sensor. Figure 3.20
shows signal examples from an IOS valve and one with SLS, respectively, when the
probe is positioned and aligned properly as shown in Figure 3.18. The condition of the
outlet strut is inferred from the signal amplitude from the probe. In a laboratory test, the
differential eddy current probe approach achieved 100% classiﬁcation accuracy with 5

IOS and 10 SLS valves.

One disadvantage of the differential eddy current probe approach is that it is highly
sensitive to the position and orientation of the sensor in relation to the outlet strut. As is
typical with eddy current testing, this approach is affected by the distance and the

alignment angle between the probe and the outlet strut. The effects of probe movement

46

inside the heart and the constantly varying distance between the probe and the heart valve

due to the cardiac cycle must be addressed before this approach can be applied clinically.

~8 m m ~3mm

 

 

 

 

 

I <3mm

 

311 turns, 42 AWG

Figure 3.17 A differential eddy current probe for detecting the outlet strut condition of a
BSCC heart valve. Adapted from [118].

Excitation Coil &
Differential Probe

 

Figure 3.18 Differential eddy current probe detection approach. Adapted from [118].
Coil 1 Coil 2
. l
Differential
Probe

Figure 3.19 Circuit diagram of the differential eddy current probe. Adapted from [118].

47

 

Figure 3.20 Examples of differential eddy current probe signals from an in vitro test. The
large amplitude signal is from an IOS valve and the small amplitude signal is from a SLS
case. Adapted from [118].

3.4.5 Catheter-Based Antenna Approach

Lepelaars et al. [113] proposed a catheter-based antenna for detecting complete and
partial fractures in BSCC heart valves. In this approach, a pulsed current is injected into
a small transmitting antenna placed in front of the heart valve. The magnetic ﬁeld
generated by this antenna induces a current in the ﬂange ring as well as in the loops
formed between the inlet and outlet struts with the ring. The transmitting antenna is then

used as the receiving antenna to detect these currents. When the valve is intact, the

48

receiving antenna signal shows a stationary behavior at late times (the time period after
the initial over-shoot and oscillations due to the pulsed excitation). When the valve is
fractured, the antenna output signal is small as there is no continuous current ﬂow in the
valve, and that the transient current is terminated after the initial oscillations. A partial
valve fracture will result in an exponential decay superimposed on a constant current,
with the decay rate proportional to the severity of the fracture. Lepelarrs et a1. had
performed detailed analytical and numerical studies [113]-[116] on simpliﬁed models
that use thin-wire loops to represent the heart valve and the transmitting-receiving
antenna. In these models, electrical impedances are inserted into the wire-loops
representing the heart valve to simulate various ﬂaw depths in the valve. An intact valve

is represented by an impedance of 0 0, while a completely broken valve is represented by
an impedance Of 106 9. Figure 3.21 shows the analytical modeling results with various
ﬂaw depths (impedance valves). Figure 3.22 shows the results from Figure 3.21

normalized with respect to the 21:106 9 reference signal.

Although the initial results from Figure 3.21 and Figure 3.22 appear promising, the
catheter-based antenna approach is still in early development stage. Some of the key

issues that must be addressed include:

1. The impedance representation of fractures in a BSCC valve must be veriﬁed;
2. More realistic models must be developed to replace the thin-wire models used;
3. The siZes of the thin-wire models should match those of actual BSCC heart valve

components (the ﬂange ring and the inlet and outlet struts);

49

4. The performance of the approach taking into consideration the overlapping
natural of the ﬂange ring and inlet and outlet strut loops;

5. The speciﬁc indication of a single-leg separation in the outlet strut;

6. The optimal location of the antenna must be determined;

7. Clinical aspects of the system such as the size versus sensitivity of the antenna,

and the effect of the heart’s motion on the detection system must be examined.

Until these issues are addressed satisfactorily, the catheter-based antenna approach is not

ready for clinical applications.

 

 

 

 

 

 

 

 

 

 

20
T "A“?
A l
< 10‘ I 21:09
E, E 21:29
0
§: 21:89
Z1=1M£2
0 j I T t
0 5 10 15 20

Time (ns) -—+

Figure 3.21 Induced current along a wire-loop model of a heart valve. The different
impedance valves (Zl's) represent various ﬂaw depths. Adapted from [116].

50

 

——+

 

 

 

 

 

 

2r ..

é

:' Z1'=8Q'i
, J
O 5 10 15 20

Time (ns) —>

Figure 3.22 Results in Figure 3.21 normalized with respect to the 21:106 0 signal.
Adapted from [116].

3.5 A Novel Approach for SLS Detection in BSCC Heart Valves
The radiographic, acoustic and electromagnetic approaches described above all suffered
from a lack of accuracy, with cardiac motions frequently introduce additional errors. The
following chapters of this thesis describe a novel electromagnetic approach, called the
gradiometer-based method, for detecting single-leg separations in BSCC heart valves.
This approach is highly accurate and is not affected by the movement of the heart (and

the heart valve).

51

CHAPTER 4. MATHEMATICAL MODEL

A mathematical model for the gradiometer-based method has been developed to facilitate
the design and characterization of system components and to provide deeper insights into
the examination technique. The model expresses the output of the gradiometer as a
function of physical parameters such as the characteristics of the excitation magnetic ﬁeld
and the location and orientation of the heart valve in relation to both the excitation ﬁeld
and the gradiometer. The model also relates the gradiometer output signal to the
condition of the outlet strut. In the following discussions, a general case is ﬁrst
developed for the scenario when the gradiometer is positioned arbitrarily in front of the
outlet strut loop. Then, the results are presented for the special case when the
gradiometer is located on the axis of the outlet strut loop and the sensor coils are
positioned parallel to the plane of the loop. This special case and the underlying
orientation and positional assumptions simplify the problem and permit the development
of a closed-form solution for the gradiometer signal. The simpliﬁed model is used in
Chapter 7 to provide a starting point for designing the gradiometer and understanding
how the various physical parameters affect system performance. The mathematical
model shows the gradiometer output as a function of the outlet strut condition, and thus
demonstrates the feasibility of the gradiometer-based method for characterizing the outlet

strut integrity.

52

4.1 Induced Current in the Outlet Strut

1n the gradiometer-based method, the heart valve is submerged in a time-varying uniform
magnetic ﬁeld, BH, created by a Helmholtz coil. The valve is oriented such that the

ﬂange ring is parallel to the applied ﬁeld as shown in Figure 4.1. In this fashion, the
ﬂange intercepts the least amount of the magnetic ﬁeld and minimal eddy current is
induced in the ring. Meanwhile, the outlet strut forms a closed loop with the ﬂange and
protrudes away from the plane of the heart valve at a 42° [151] anglel. This outlet strut

loop intercepts the applied magnetic ﬁeld and an electromotive force (EMF), vS , is

induced in the loop according to Faraday ’8 Law:

(I
VS(I)=—ENSWS

d _.
‘ ”ESIBH ass v (4.1)
S

In (4.1), the number of turns, N S, is equal to 1 for the strut loop. 1/15 is the magnetic ﬂux
linking the loop, and SS is the projected or effective area of the strut loop that intercepts
the applied ﬁeld. The orientation of this ﬁeld is shown in Figure 4.1. In Figure 4.2, the

component of BH that is normal to the plane of the strut loop is

 

1This angle is for a 60° BSCC valve. The outlet strut angle is different between a 60°
and 70° valve. Although the discussions in this thesis focus on the 60° valves, the design,
results and conclusions presented for the gradiometer-based detection system herein are

equally applicable for the 70° valves.

53

. : 4r ,.
'3 ’ pr
'1‘. r'“
------ :- P‘"°"‘-.
..... wi' 1..
r ':
o AV 1

 
   

 

 

 

 

 

 

*------------ ---------

17 if

V V

Figure 4.1 The heart valve is submerged in a uniform magnetic ﬁeld and oriented in a
fashion such that the ﬂange ring intercepts the least amount of the applied ﬁeld, while the
outlet strut loop intercepts the maximum amount of the ﬁeld.

B’H =|13H|sin(ot)a'H (4.2)

Substituting (4.2) into (4.1) and integrating the contributions from the incremental

regions d SS over the effective strut loop area SS yields:

—S
dt

=-—i|B "S 'sin(a)
dt H S

d .
= ‘EBHSS srn(a) (4.3)

where BH = [B Hl is the amplitude of the applied magnetic ﬁeld at the outlet strut,

S S = IS 5| is the effective area of the strut loop, and a is the angle between the outlet strut

plane and the direction of the applied magnetic ﬁeld. Equation (4.3) is valid when the

heart valve is oriented as shown in Figure 4.1. In this manner, the outlet strut loop

54

presents the maximum effective area for intercepting the applied magnetic ﬁeld. If the
outlet strut is rotated ,6 degrees with respect to the applied ﬁeld as depicted in Figure 4.3,
the effective area will decrease and the induced EMF in the strut loop will subsequently

be reduced by a factor of cos(,B) such that

vs(t)=—%BHSssin(a/)cos(ﬂ) (4.4)

B’H
0

BH

Figure 4.2 Current is induced in the outlet strut loop due to the component of the applied
magnetic ﬁeld that is normal to the plane of the loop

.[ 13H

  

 

 

 

 

Figure 4.3 If the heart valve is rotated ,6 degrees with respect to the applied magnetic
ﬁeld, the effective area of the outlet strut loop for intercepting the applied ﬁeld will be
reduced.

55

The current, iS , in the strut loop due to the induced EMF is given by Ohm ’s Law:

is(t)= v50) (4.5)

 

 

. ‘ d BH SS sin(a)cos(,6)

1s (t) = 7 (4.6)
I RS

where RS is the resistance of the outlet strut loop. This resistance is a function of the

conditions and material characteristics of the closed-loop formed by outlet shut and the
ﬂange. The relation between the outlet strut condition and the strut loop resistance will

be discussed later in this chapter.

4.2 Magnetic Field Due to the Helmholtz Coil System

To obtain BH (the magnitude of the Helmholtz coil magnetic ﬁeld, BH), let the excitation

current be a sinusoidal signal, iH , with amplitude I H and frequency fH Hz, such that
1.1.10): [H SIII(CUHI) A (4.7)

where (UH = 2an. With a radius rH and N H turns in each coil, the magnitude of the

magnetic ﬁeld at the center of the excitation coil system is given by the Helmholtz coil

equation [129]:

H_Ngf‘o H.H

SfrH

8#N
0 H
B Ht=() —5_—5—_x/_rH I Hsin(a)Ht) (48)

 

56

The direction of this ﬁeld is parallel to the axis of the coils. Substituting (4.8) into (4.6)

yields

, __i 8ﬂ0NHIH£§ . .
so- a.(— 5% RS sweetness/3)] (4.9)

In practice, the angles aand ,6 are functions of time synchronized with the cardiac cycle.

In which case, (4.9) can be expressed as

. d 8ﬂ0N H I H SS . .

l t =—— ———srn tsrn at cos ﬁt 4.10

so (,1 5% RS (can) (0) (1» ( 1
On the other hand, in situ movement study1 reveals that the heart valve motion due to the
cardiac activity involves primarily translational movements that are orthogonal to the
applied magnetic ﬁeld. In this case, aand ,8 may be assumed stationary and (4.9) may

be simpliﬁed to:

. 8 N1 S
'S(’)=_ .UO HH_S_

Ss/Em Rs wgsin(a)008(ﬂ)008(wrlt) (4.11)

4.3 Magnetic Field Due to the Outlet Strut Loop Current
The induced current in the strut loop creates a magnetic ﬁeld, Bs. To describe this ﬁeld,
the outlet strut loop is approximated as a single-wire circular coil. The radius rs of the
circular coil is selected such that the area enclosed by the coil is equal to the projected

area of the outlet strut loop:

 

1Results of the in situ heart valve movement study from cinegrams are discussed in
Chapter 9.

57

SS :ﬂ.rsz (4.12)

With the circular coil approximation, the magnetic ﬁeld observed at an off-axis point P
(Figure 4.4) may be Obtained by determining the vector potential of the loop. Since

V - B S = 0 , the magnetic ﬂux density may be expressed as the curl of the magnetic vector

potential, As , such that

BS =VXAS (4.13)

From Biot-Savart ’s Law, the vector potential due to a differential current element, di, on
the circular loop L5 is [184]

, , ,uois d1 ,uois rsue’ .-
A r, =-— —=——— -———-—d
3( Z) 42: Ll: W 47: Ll W 40 (4°14)

In (4.14), the variables r’ and z’ belong to a local coordinate system in which the z’ axis
is normal to the plane of the outlet strut loop and the x’-y’ plane coincides with the strut

loop plane. The length W is the distance between the wire loop and point P. The

cylindrical coordinate variables (r’, (0’, z’) are given by r’ = ‘lx'z + y'2 . (0’ = tan—1 -:—,

and z’= z’. The origin of this local coordinate system is located at the center of the strut
loop as shown in Figure 4.4. In this ﬁgure, the vector potential due to the differential
current at point m is largely cancelled by the contributions from the differential current at

point 11 due to symmetry about the (o’ = 0 (x’-z’) plane. The resulting net potential at P is

perpendicular to that plane, or more generally, in the direction of 13¢. Considering the

contributions from all the symmetric differential-current pairs on the loop, the resulting

58

magnetic vector potential, Ag, of the loop will also be in the a), direction, with a

magnitude, AS, given by

fj- 2rs cos (0'

I

r r 1
A50“ ,2 ) = #073 1 d(0 (4.15)
0 (rs2 + r'2 + z'2 — 2rsr’cos (0')§

 

Equation (4.15) can be rewritten as [184]-[187]

l
. — 2
AS(r’, z’) = $78erle [[1 —£2-]K(k)— 156)] (4.16)
z’ PO". 0. z’)
A.
,Q
I ll

\
\

 

 

 

-F------

‘0

V

 

 

Figure 4.4 Off-axis magnetic ﬁeld due to the outlet strut (approximated by a circular
' current loop).

59

where

 

I
k =[ 4’5" ]2 (4.17)
2 2

(rs + r') + z'

and K( k) and E( k) are the complete elliptic integrals [188], [189] of the ﬁrst and second

kind, respectively. The two integrals are deﬁned as

 

% .
K(k)= j 1 d6 (4.18)
O(l—kzsin26)2
E
l
E(k): 2[(r—Itzsin219Ft1t9 (4.19)
0

From the expression of AS and equations (4.17)-(4.19), the components of the magnetic

ﬁeld, B3, due to the outlet strut current (as approximated by a circular current loop of

radius r5 ) are given by [184], [185]

 

 

2 I2 12

i ’ r + r + Z
83,, = #0 8 Z S , E(k)- K(k) 4.20
2,; l (r __ r )2 + r2 ( )

r'[(r +r')2 + ’45- S 2

3 Z
. 2 12 [2
, _ #018 1 rs ‘ r ‘ Z
853 - 2n 1 [(rs - r')2 + 2'2 E(k)+ K(k)l (4.21)
[(rs + r’)2 + 2’2];

Equations (4.20) and (4.21) describe the magnetic ﬁeld due to the induced current is in

the outlet strut for a location off-axis from the strut loop as shown in Figure 4.5. For a

gradiometer placed at this position, the total magnetic ﬁeld experienced by each of the

60

two gradiometer sensor coils is the sum of the magnet ﬁelds due to the Helmholtz coil
and the outlet strut current:

Btoral = BH + BS (4-22)

lrl 3.

Gradiometer

 

 

 

 

Figure 4.5 A gradiometer placed off-axis from the outlet strut loop

4.4 Induced EMF in the Gradiometer Coils
The gradiometer consists of two bobbin coils aligned co-axially and separated by a ﬁxed
distance. Let the coil closest to the outlet strut loop be coil 1, with N1 number of turns,
radius r] and a cross sectional area S 1 such that S1 = 7r - r12. Similarly, denote the coil
farther away from the strut loop as coil 2, with N 2 number of turns, radius r2 , and a
cross sectional area S 2. In practice, the two gradiometer coils are identical in
construction and geometry. The two coils are also electrically connected in opposition

such that the output of the gradiometer, v g , is the difference of two coil outputs, v1 and

v2 , respectively.

61

The magnetic ﬂux lines linking the outlet strut and the gradiometer coils are

WI = [BS] dSI (423)
Si

”’2 =SjB32 «182 (4.24)
2

From Faraday ’8 Law, the induced EMF in the ﬁrst gradiometer coil due to the linkage

ﬂux from (4.23) is given by:

d I (4.25)
t 31

Similarly, the induced EMF in the second gradiometer coil due to the linkage ﬂux in

(4.24) is
d
v t = — — N
21 l dt 21” 2
_ d I ._ (4.26)
_ -N2 :1; [1352 - 6182
S
The gradiometer output is, therefore, given by
vg (t): v1(t)-v2 (t) (4.27)

These matching EMFs terms are canceled upon the subtraction in (4.27) and

consequently do not contribute to the output of the gradiometer.

62

To obtain the output, vg , for this non-trivial case, the location and geometry of each

gradiometer sensor coil must be known. Equations (4.20) and (4.21) are then solved for
every point that lies on the plane of the area, S 1 or S 2 , enclosed by the respective sensor
coil. The linkage ﬂuxes are then obtained from (4.23) and (4.24), and the corresponding
induced electromotive forces are obtained from (4.25) and (4.26). Finally, the

gradiometer output signal is Obtained from (4.27).

4.5 Special Case: Gradiometer on Axis of Outlet Strut Loop
From Section 4.3, if a gradiometer is placed at a point (r’, (p ’, z’) in front of the outlet
strut loop, the magnetic ﬁeld experienced by the gradiometer coils due to the current in
the strut loop is given by (4.20) and (4.21). For the special case when the gradiometer is
placed on the z’-axis of the outlet strut loop as shown in Figure 4.6, and assuming the

sensor coil areas, S l and S 2 are small comparing to the strut loop, the equations may be

combined and simpliﬁed as

- 2
#01er .

3 32'
2(r32 + [2);

 

leren = (4.28)

If the gradiometer is advanced to the center of the strut loop, 2’ also becomes zero and the

ﬁeld experienced by the gradiometer is given by

 

#ois ,.
B3lr'=0,z’=0 = 2,3 W (4.29)

63

   

Gradiometer

 

 

 

Figure 4.6 A gradiometer placed on the axis of the outlet strut loop

In practice, the gradiometer is positioned by a cardiologist during the prosthetic heart
valve inspection. The cardiologist will adjust the orientation and alignment of the

gradiometer as close to that shown in Figure 4.6 as possible.

4.6 Induced EMF in the Gradiometer Coils
For the case when the gradiometer is positioned on the axis of the outlet strut loop as
shown in Figure 4.6, (4.25) and (4.26) may be simpliﬁed as
v1 (t) = —N1 1 BS 81 (4.30)
d1 1

and

d (4.31)
V2(t)='N2;Bsz-92

The magnitude of the outlet strut’s magnetic ﬁeld experienced by each of the two

gradiometers coils located at zl and Z2 , respectively, is

2.
#0’3 18

 

BS:

 

 

1 _3_ (4.32)
2(’82 + Z12 )2
and
B #0 r3218
52 _ 3 (4.33)
2(’82 + Z22)E
The induced EMF in each gradiometer coil is, therefore,
”ﬂo'sz '12 dis (t)
V10) = “N1 3 (4.34)

2 2 dr
2(’8 +21 F

and

”60’82’22 disl’)
3 (4.35)

V210: ‘Nz dt
2(rs2 + z22)5

 

Since the two gradiometer coils are identical, substituting (4.34) and (4.35) into (4.27)
yields

N ﬂﬂorzrz d' (t) 1 1
): 8 S 8 1S _

+ 3 (4.36)

2 dt 2 23 2 2
(rs +21F (’8 +22)5

where N g = N1 = N 2 is the number turns in each gradiometer coil and r¢g = r1 = r2 is

 

 

vg(t

the radius of the gradiometer coil. The current in the outlet strut, is , is given by (4.9).
For this on-axis case, if the angles a and ,6 and the resistance R3 of the strut loop remain

constant over time, carrying out the differentiation in (4.9) gives

65

13(1) = - 8'UONHIH E—S—sin(a)cos(ﬂ)§t—sin(amt)

 

 

 

53m Rs (4 37)
“’11 8A0NH1HSS - '
= — SlIl a COS COS 1
RS 5% () (a) (an)
Substituting (4.37) into (4.36) yields
vg(t)=
N n 2 2
8 ”0’5’8 1 + 1 “’H 8“ONH’HSSs...(e)ees(p)i805(8),,»
2 2 2 3 2 2 3 RS S‘B'H dt
_("8 +22 )5 (rs +21 )5_

 

 

(4.38)

Rearranging and simplifying this equation gives the gradiometer output for the on-axis
case as
2 2 2

v8 (t)— Rs 53/3711 sin(a)cos(,6) - sin(a)Ht)

 

 

 

 

(4.39)

Let the separation between the two gradiometer coils be Az. The distance 22 can then be

expressed in terms of 21 as:

22 = Z1 + AZ (4.40)

Substituting (4.12) and (4.40) into (4.39) yields

66

vg(r)=

247: 2IV 1 N rZrZS
601i #0 HH ggS s. 1 1 _
srnacos — srn t

(rs?- +605 (.32 +(z. +442 2

 

 

(4.41)

Equation (4.41) gives the output of the gradiometer located on the axis of the outlet strut

loop. Note that in (4.37) and (4.41) the angles a and ,6, the effective areas S S and

S g (= 7r - rg2 ), and the strut loop resistance R5 are assumed to be non-varying over time.

These equations show that the outlet strut current and the gradiometer output voltage are

both sinusoidal signals at the excitation frequency.

4.7 Outlet Strut Condition and Strut Loop Resistance

In equation (4.5), the resistance RS of the outlet strut loop is a function of the material

characteristics and condition of the loop. Assuming that the conductance of the strut loop
is homogeneous throughout the ﬂange ring and the outlet strut, the resistance Of an intact

outlet strut loop may be represented by a lumped resistance, R0. For an outlet strut that is
either completely or partially fractured, the separation may be represented by an electrical

resistance, RL, corresponding to the degree of discontinuity at the location of the fracture

[1 13]-[l 17]. The outlet strut loop resistance RS for a general case can then be expressed

as

R =R +R (4.42)

67

The current induced in the outlet strut loop becomes a function of the load resistance

(which in terms is a function of the condition of the outlet strut):

 

. _ d 8,u N 1‘ s . .
15(t)——E[ 535:1“ (R0 +SRL)srn(wHt)srn(a)cos(,6)]

__8#0NH1H 58
53/341 (R0+RL

)wH cos(wHt)sin(a)cos(/J’) (4'43)
In (4.43), the condition of the outlet strut is modeled by the load resistance RL. For a

complete fracture (i.e., a SLS case), RL = 00. This effectively creates an open circuit and
no current may ﬂow in the outlet strut loop. For an intact outlet strut, RL = 0 , R S = R0 ,

and (4.43) becomes (4.9).

The development of a comprehensive model between RL and the degree Of discontinuity
in a fractured outlet strut is described in [113]-[117]. For the development of the
gradiometer-based examination system, only the cases where RL = 0 and RL = 00 will

be considered.

4.8 An Alternative Approach: Mutual Inductance Consideration
A conceptual representation of the gradiometer-based examination method is shown in

Figure 4.7. In this ﬁgure, the gradiometer output, v , is the difference between the two

8

gradiometer sensor signals, v1 and v2 , respectively. The output of each gradiometer coil

is the result of ﬂux linkages among a system of four coils: the top and bottom coils of the

Helmholtz system, the outlet strut loop and the gradiometer sensor coil itself. Figure 4.8

68

shows a circuit element representation of the coil systems with the ﬁrst gradiometer
sensor coil. In this ﬁgure, the various coils and current loops are represented as inductors

(L). The time domain output signal, v1, is obtained from the frequency domain

expression

V1 =-J'€0HM1,sls +ijMl,H11H +jwliMl,H21H (4.44)

where I S and I H are the frequency domain representations Of the outlet strut current and

excitation coil current, respectively. The output, V1, is the frequency domain

representation of VI, and to is the excitation frequency. The mutual inductance, M 1,8» is
deﬁned as the ﬂux linkage 41,8 = [VIII/1,8 on the ﬁrst gradiometer sensor coil due to the

outlet strut loop current, 15:

M15: ’ = ' (4.45)

 

Similarly, the mutual inductance M 1.11] is deﬁned as the ﬂux linkages of the ﬁrst

gradiometer sensor coil per unit In such that

M 1 H1 = = ' (4-46)

 

69

 

 

 

Helmholtz Coil _ :
System —————————_E_:———_____

 

 

Gradiometer

Sensor Coils .-./"" V8
Outlet Strut -. ‘ ,-

Loop

 

 

 

 

 

Figure 4.7 A conceptual representation of the gradiometer-based inspection system
(components not drawn to scale)

 

 

 

 

S
r'“““””’”‘”"’”’[. ”HI kc 4:
l
S M1 S 1 1
.- M1,H2 /. Q
0
1
Lee I H

I l
l l
l I
Figure 4.8 A circuit element representation of the gradiometer-based inspection system
components (shown with the ﬁrst gradiometer sensor coil only) in Figure 4.7.
Similar to (4.44), the frequency domain expression for v2 is

70

V2 = -J'wHM2,sls +ijM2,HlIH +ijM2,H21H (4.47)

where M 2,8 , M 2,H1 and M 2,112 are the mutual inductances due to the ﬂux linkages to

the second sensor coil from the outlet strut loop and the top and bottom coils of the
Helmholtz system, respectively. Since the gradiometer is submerged in a uniform ﬁeld
created by the Helmholtz coil, the ﬂux linkages from the top and bottom coils to the

gradiometer sensor coils are equal. As a result, the gradiometer output may be written as:

Vg =V1-V2 =-J'0}HM1,sls +ijM2,SIS

 

 

= (M 2,8 - M 1,8 )ij Is (448)
The corresponding time domain signal is
Vg (I) = (M 2,8 - M 1,8 ldizft) (4.49)
To determine M 1,8 , the deﬁnition of mutual inductance may be used:
M1,s = N}? (4.50)

The ﬂux ((11 passing through the ﬁrst gradiometer sensor coil due to the current in the
outlet strut loop is given by (4.23). For the case when the gradiometer is positioned on
the axis of the outlet strut loop (as shown in Figure 4.6), the magnitude of the outlet strut
magnetic ﬁeld experienced by the ﬁrst gradiometer sensor coil is given by (4.32). The

magnetic ﬂux linking the outlet shut loop and the sensor coil is

71

 

M 1,SiS (4.51)

Comparing (4.50) and (4.51), the mutual inductance M 1,8 is given by

2
1110’le ( 2)
MLS = 3 III]
2(7‘82'1'212)E

 

(4.52)

Similarly, the mutual inductance between the outlet strut loop and the second sensor coil

of the gradiometer is given by

 

2
ﬂorsN2 ( 2)
M23: 3 7272

 

 

, (4.53)
2(7‘82 + Z22)E
Using equations (4.52) and (4.53) and recalling the gradiometer speciﬁcations
N8 = N1 = N2 and r8 = r1 = r2 , (4.49) can be written as
N gzrﬂﬂor 1 1 at t
v, (r): S g S( ) (4.54)

2 (suit (a mi d’

The current is in the outlet strut loop is given by (4.37). The time derivative Of this

current is

72

dis(’) _ (0121 8/10NH1HSS

 

 

= sin 0' cos ﬂ sin I (4.55)
d, RS 5% <) (Ham
Substituting (4.55) into (4.54) yields
(0% 47Z'ﬂ3NgNHIHr52I‘gZSS 1 1

 

 

v (t) = sin(a)cos(,6 — sin(a)Ht)
g R 3 3
S 5‘5,“ (rs2 + leﬁ (r82 + 222};

(4.56)

Equation (4.56) is identical to (4.39). This result is expected since, by deﬁnition, the
mutual inductance is the flux linkage of the sensor coil due to the outlet strut loop
current. As this linkage flux is also used in the derivation of (4.39), both developments
are expected to arrive at the same result. Essentially, the mutual inductance approach
amounts to an alternative presentation of the same physical phenomenon as described by

the Faraday ’s Law in (4.25) and (4.26).

4.9 Factors Affecting the Gradiometer Signal Output
Equation (4.37) shows that the amplitude of the strut current is proportional to the

excitation frequency, (UH , and inversely proportional to the resistance, RS , of the strut

loop. From (4.41), the amplitude of the gradiometer output signal is:

o proportional to the square of the excitation frequency, (0%;
o inversely proportional to the strut loop resistance, Rs ;

o proportional to the number of turns, N g , in each gradiometer sensor coil;

0 proportional to the effective area of each gradiometer sensor coil, S g = 7: r; ;

73

0 related to the strength of the applied uniform excitation magnetic ﬁeld, BH , from

the Helmholtz coil system. This ﬁeld is in turn

0 a function of the number of turns, N H , in each excitation coil;
0 proportional to the amplitude of the current (1H ) through the coils;

o inversely proportional to the radius, rH , and separation of the coils.

4.10 Spatial Dependency of the Gradiometer Output Signal
The gradiometer output signal is also a function of its distance, 21 , from the outlet strut
as well as the separation, Az, between the two gradiometer sensor coils. To visualize
these dependencies, Equation (4.41) is expressed as a product of a time-dependent

function W(t) and a position-dependent function U (21 , AZ) such that

 

vg =W(t)-U(z1.Az) (457)
where
2 47: 2N I N 2 25
W(t)=wH #0 H H grgrs Ssin(a)cos(ﬂ)sin(a)Ht) (4.58)
RS 5611.1
and
-2 2
U(z1,Az)= (r32 + 2.12) 2 -(r82 +(z1 + Az)2)2 (4.59)

Equation (4.59) describes the scaled gradiometer output signal, U(z1,Az) , as a function
of both 21 and Az. Figure 4.9 and Figure 4.10 show U(z1 ,Az) as a function of zl and

A2, respectively.“ Experimental results showing the examination system output as a

74

function of z] are shown in Chapter 8, and the development of a probability of detection

model with respect to variations in 2:1 is presented in Chapter 9.

-3 Scaled Gradiometer Output Signal U(z1,Az) vs. 21

6 _, "‘ “'1‘” ‘ ' 1 I T 1 T 1
l U(z,,Az), AZ=2.3 mm Q

I T—ﬁ

l __ . __ __ _ ___ ___.

’ r
| j ‘
/’ R
, ,
, \ ‘
\
‘\
\

 

Arbituary Units

 

 

 

r.
L.
|
r.
,—
l
o

l
o 1 2 3 4 '5 6 "7 8 W9 10
21(mm)

Figure 4.9 Scaled gradiometer output signal as a function of the distance, zl , between the
gradiometer and the outlet strut loop

75

-3 Sealed Gradiometer Output Signal U(z1,Az) vs. A2

9,-.- . T ﬁ—ﬁ. . . . —.

 

 

 

 

 

O)

Arbituary Units

1 '._ % ;____-___-_. _ _ _ _ _ [.4
' U(z1,Az), 21:2 mm 1

 

1
- ' ‘- ‘ - ._.. 1
L _ _ _ __- . .__—.._.J_ _ .L 7 I

o" 2' 4 6 a “139—312 14 16 18 20
AZ(mm)

Figure 4.10 Scaled gradiometer output signal as a function of the separation, Az, between
the two sensor coils

In Figure 4.9, the output signal, U (z1,Az) (Az =2.3 mm), increases in the ﬁrst 1-2 mm
when the distance between the gradiometer and the outlet strut is increased. The signal
subsequently decreases with distance, and is reduced by 3 dB from its maximum value
when the gradiometer is approximately 3 mm from the outlet strut loop. These results
suggest that the gradiometer should be placed within 5 mm of the outlet strut (i.e., 21$ 5
mm) during an examination. Finite element analysis of the gradiometer output with
respect to its distance from the outlet strut is presented in Chapter 5, and the experimental

results are given in Chapter 8. A probability of detection model that takes variabilities in

21 into consideration is presented in Chapter 9.

76

Figure 4.10 shows the scaled gradiometer output signal as a function of Az (with z] = 2

mm). The results from this ﬁgure show that if the ﬁrst sensor coil is placed in the
vicinity of the outlet strut loop, the gradiometer output signal will increase as the second
sensor coil is moved away from the ﬁrst. This is because when the second coil is near to
the ﬁrst, the induced signals in both coils are very similar and, therefore, the difference
signal is small. As the separation between the coils, Az , increases, the induced signal in
the second coil decreases and the difference signal subsequently increases until it levels
off towards a valve that is a function of the ﬁrst sensor coil signal. The gradiometer

output as functions of the distance, z 1, from the outlet strut, and the separation, Az,

between the two sensor coils, is examined further in Chapter 7 for designing the

gradiometer.

It is important to recall that when simplifying (4.20) and (4.21) into (4.28), the area of the
sensor coils was assumed to be small comparing to the outlet strut loop. If this
assumption is not satisﬁed, (4.20) and (4.21) must be evaluated and applied in (4.30) and
(4.31) to obtain the output of the sensor coils. Nevertheless, equation (4.41) and the
simpliﬁed mathematical model developed in this chapter provide a valid and convenient
starting point for designing the gradiometer. Equation (4.41) also provides insights into
the effects of various physical parameters on the gradiometer output as described in

Section 4.9. For example, Equation (4.41) shows that, with all other parameters kept

, is a quadratic function of

 

 

constant, the amplitude of the gradiometer output signal, v g

cc (0%] ). This quadratic dependency can be

 

 

the excitation frequency, a)” (i.e., v g

77

observed from Figure 4.11, which shows the scaled amplitude of the gradiometer signal
from fy = 10-100 kHz (0);; = 2 1t fH). The results in this ﬁgure suggest that the excitation
frequency should be set as high as possible within the limits of the equipment. The

experimental results for the gradiometer output as a function of excitation frequency are

presented in Chapter 8.

Scaled Amplitude of Gradiometer Signal vs. Excitation Frequency
120 l T T T l r Ff l '
‘ l l ‘ I I l ‘

 

; . 1 1 i r 1 I

‘ 1 1 1 1 1 1 1
100 .......... ,__1--_,__1____--.; ______ 1 ______ 1 ....... .L ______

l I 1 | l l l

1 1 1 . 1 I .

    

801

Arbituary Units
0)
O
l

 

 

20 — 7

, . . i

0 .. .. ,2, l_ ._ _, l - ..__ ,4 i ,. ‘ ___ _ ' . _ .J
1 2 3 4 5 6 7 8 9 1o

Excitation Frequency (Hz) x 104

Figure 4.11 Scaled amplitude of the gradiometer output signal as a function of the
excitation frequency

Equation (4.43) shows that, if the outlet strut fracture is modeled as an electrical
resistance and the valve is subjected to an externally applied uniform magnetic ﬁeld, the

induced current in the strut is directly related to the outlet strut condition. This outlet

78

strut current manifests itself as perturbation in the applied magnetic ﬁeld. When a
gradiometer is used to measure this perturbation, the results from (4.25)-(4.27) can be

used to indicate the condition of the outlet strut.

Equation (4.41), though developed for the special case when the gradiometer is
positioned on the axis of the outlet strut loop, expresses the gradiometer output signal as a
function of the applied excitation ﬁeld, the geometry of the gradiometer coils, the
orientation of the gradiometer in relation to the outlet strut, and the condition of the outlet
strut. These results are all taken into considerations when designing the gradiometer

assembly.

79

CHAPTER 5. NUMERICAL MODELING

A three-dimensional ﬁnite element analysis has been performed on the gradiometer-based
heart valve examination system to investigate the feasibility of the approach and for
optimizing system parameters. The analysis results show the magnetic ﬁeld perturbations
in the vicinity of the heart valve when subjected to an applied excitation ﬁeld. The
results also show that the amount of perturbation due to an intact heart valve is large
when compared to that due to a single-leg separated valve. These results demonstrate
that the gradiometer-based method is a promising approach for detecting SLS in BSCC

heart valves.

5.1 The Finite Element Model

The ﬁnite element model is based on a quasi-static vector-scalar (A -V) potential
formulation [129], [130]. Figure 5.1 shows the conﬁguration for a typical eddy current

problem. The solution domain is divided into a conducting region, (21 , and a non-
conducting region, £22. In $21, the magnetic vector potential A and electric scalar
potential V are used to represent the electromagnetic ﬁeld. In S22 , only A is used. The
excitation current is represented in the form of a current density, is , and is assumed to

exist in the non-conducting region only. The governing equations are derived from the

Maxwell’s equations:

80

 

 

 

 

 

 

Non-conducting
Region
9 2 Conducting
Region
ﬂ; EEZED 91
EEZED

 

 

Figure 5.1 A typical conﬁguration for eddy current problems

where

electric ﬁeld intensity (V/m)

electric ﬂux density (C/m2)
magnetic ﬁeld intensity (A/m)

magnetic ﬂux density (T)

4.. 6:11:31 Cl ml
11

= current density (A/m2)

pV = volume charge density (C/m3)

(5.1)

(5.2)

(5.3)

(5 .4)

Under quasi-static conditions, the displacement current, IS , is assumed to be negligible:

a6_

——0.
6:

Applying the magnetic vector potential formulation

E=VXA.

81

(5.5)

(5.6)

with the Columb gauge ([131])
v K = 0

and a sinusoidal excitation of frequency a)
813'

——=— 'wE
a: 1

allows substituting (5.6) and (5.8) into (5.2) to give
VXE = —wa = —jm(VxA)
from which
Vx(E+ ﬂux): 0.
Hence,
E = —ij — vv
where V is the electric scalar potential function.

Also, from the constitutive equations we have

n=i§
,u
1 _
=—V A
#1 x1
and
T=0E

where O’iS the conductivity and u the permeability of the material in the region.

Substituting (5.5) and (5.11)-(5.13) into (5.2) yields
in(VxX)= 0(— ij—VV)
,u

The problem is therefore deﬁned by the following governing equations:

82

(5 .7)

(5.8)

(5.9)

(5.10)

(5.11)

(5.12)

(5.13)

(5.14)

inVxX+jwaX+oVV=0 (5.15)
,u

In 821:
v.a(jw7£+vv)=0 (5.16)
inVxK—i 61%
In S22: # S .

In these equations, a) is the angular frequency of the excitation signal. At the interface

F12 between the conducting and non-conducting regions, the boundary condition

J-ﬁ=a(-ij—vv)-ﬁ=0 (5.18)

is imposed.

Solving equations (5.15)-(5.17) yields the magnetic vector potential function, A , and
electric scalar potential function, V, from which the magnetic ﬂux density, E , can be

obtained.

The ﬁnite element method does not solve these partial differential equations directly.
Rather, the method seeks the weak form solution. Using Galerkin’s rule, the discretized

weak form representation is [122]:

§ ﬁlmy x 31} )% (V xiii j )+ (viii )% (V1111. )+ jwaNiNj ]dg2}Aj

+ZUHNiVNJ-dQ]Vj

= ”J-Ni-jsdg (5.19)

83

and

ZKjwa H vﬁiﬁjdojAj]+[ H VNiOVdeQ]Vj = 0 (5.20)
1

where NJ— represents the basis function associated with the nodes of the elements. This

discretized weak form representation is a linear system.
The ﬁnite element analysis of the system involves the following steps:
0 discretize the solution domain into ﬁnite elements,
0 evaluate the elemental equations,
0 assemble the global linear system,
0 solve the linear system of equations,

0 process the nodal values to obtain the physically measurable quantities.

Figure 5.2 shows the ﬁnite element mesh of a BSCC heart valve generated using the

COMSOL Multiphysicsl software package. In this model, ﬁrst order tetrahedron
elements are used. Solving the linear system given in (5.19) and (5.20) produces a
solution for A and V at the ﬁnite element nodes. The values of the potential functions
and ﬁelds at any arbitrary point in the solution domain can then be obtained through

interpolation.

 

l Formerly known as FEMLAB, by COMSOL, Inc. of Burlington, MA
(www.COMSOL.com).

84

5.2 Simulation Results
At the center of the gradiometer-based inspection system is a uniform magnetic ﬁeld
encompassing the heart valve under test. To maximize the volume of this uniform ﬁeld,
a Helmholtz excitation coil system is uSed. In a clinical-scale setup, the excitation coils
are separated by 1 m in order to accommodate the patient and the examination table.
Each excitation coil also has a l-meter radius following the Helmholtz conﬁguration.
When constructing the ﬁnite element model for the system, the element size has to be
sufﬁciently small particularly within the ring and its vicinity. Unfortunately, this requires
the use of a large number of elements, making the computational effort prohibitively

large.

~0.015

 

0.015 0,015

Figure 5.2 Finite element mesh of a BSCC heart valve (the inlet strut and occluder disc
are omitted for clarity without loss of model accuracy)

85

In order to reduce the computational requirements, the analysis of the gradiometer-based
method is performed in two steps. First, the excitation coils are modeled to obtain the
ﬁeld distribution at the center region. This provides estimates of the ﬁeld strength and
the size of the uniform magnetic ﬁeld Volume. Next, a second analysis is performed to
model the heart valve centered in the uniform magnetic ﬁeld. The ﬁeld distribution
obtained from the ﬁrst ﬁnite element analysis is applied as the source to the second
model to excite the heart valve. The perturbation of the applied magnetic ﬁeld is then

calculated for various outlet strut conditions.

5.2.1 The Excitation Coils

A 3-D ﬁnite element model has been constructed for the Helmholtz coil system. In this

model, each excitation coil consists of 9 turns (3 layers of 3 turns each) of a 1/0 AWG1
wire (Figure 5.3). The diameter of the wire is 11 mm and the mean radius of each
excitation coil is 1 m. The two excitation coils are separated by 1 m. The two coils are
also electrically connected in series, and oriented such that the current ﬂows through both

coils in the same direction. From Figure 5 .3, the cross-sectional area of the 3x3 winding
stack is approximately 1.52x10.3 m2 (39 mm x 39 mm). The empty spaces between the
wires in the winding stack and the insulation jacket thickness (~ 1 mm) make the ﬁll-

2 . . . . . .
factor of each excrtation corl less than 0.57. The empty spaces and the wrre insulatlon

also complicate the 3-D ﬁnite element model. To simplify the model and reduce the

 

1 . .
Amencan wrre gauge

 

97r(0.0055)2 _

0.56
(0.039)2

2 . .
F111 factor = conductor area / wrndow area used =

86

mesh size without loss of accuracy, each excitation coil is modeled as a square-shaped

wire (Figure 5.4) with a cross-sectional area equivalent to the total conductor area of the

3x3 winding stack. With a total winding cross-sectional conductor area of 91t(0.0055)2 =

8.553 x 10.4 m2, the cross-sectional width and height of the square wire are both 29.245

 

 

 

Figure 5.3 Each excitation coil consists of nine turns of a 1/0 AWG wire. Due to the
inter-wire spacing and jacket thickness, the ﬁll-factor of this coil less than 0.57.

29.245 mm

 

 

Figure 5.4 Excitation coil windings are modeled as a wire with a square cross-sectional
area

87

The magnetic ﬂux density and ﬁeld distribution predicted by the 3-D ﬁnite element
analysis of the Helmholtz coil system is shown in Figure 5 .5. Figure 5.6 shows the
magnetic flux density proﬁle along the x axis at the center of the coils, and Figure 5.7
shows the magnetic ﬂux density proﬁle along the z axis. With a sinusoidal input current

of 11 A (peak-to—peak) and a frequency of 62.5 kHz, the simulation results show a peak-

to-peak magnetic flux density of 1.47x10.4 T (1.47 gauss) at the center of the coils. The

simulation results also show a less than 0.5% variation in the magnetic ﬂux density
within a cylindrical volume 0.5 meters tall and a radius of 0.25 m centered between the
coils. The experimental conﬁrmation of these simulation results are presented in Chapter

8.

5.2.2 Field Perturbations Due to Outlet Strut Conditions

The ﬁnite element model of a heart valve submerged in a cylindrical volume of uniform
magnetic ﬁeld is shown in Figure 5.8. The magnetic ﬁeld distribution was obtained
previously from the Helmholtz coil simulation. In this model, the center of the valve
ﬂange ring is located at the origin of the x-y plane, while the positive z-axis extends from
the ﬂange plane into the outlet side of the valve. The outlet strut is positioned in the “+y”
direction. Figure 5.9 shows the real part of the y and 2 components of the magnetic ﬂux
density in the x = 0 mm plane for both the IOS and SLS cases (the vertical axis is the y-
axis). In these cases, the real components of the ﬁelds are unperturbed. The
perturbations become obvious in the imaginary parts, as shown in Figure 5 .10 and Figure

5.11. From these two ﬁgures, the level of perturbations due to an IOS valve is higher

than that from a SLS case. Figure 5 .12 shows the real and imaginary parts of the x and y

88

components of the ﬂux density in the z = -5 mm plane for an IOS valve, while Figure
5.13 shows those from a SLS case. The results show that, in the vicinity of the valve (5
mm in front of the ﬂange plane in this case), the degree of perturbation in the magnetic
ﬁeld is higher for an IOS valve than that for a single-leg separated valve. This
demonstrates the feasibility of outlet strut SLS detection by measuring the magnetic ﬁeld
perturbations in the vicinity of the heart valve. The following chapters describe a

gradiometer-based examination system utilizing this principle for SLS detection.

Magnetic ﬂux density, norm [T] Max: 2.525e-3

 

 

‘ ‘1.5

 

 

 

 

   

 

Min: 4.33e-7

Figure 5.5 Finite element analysis of the Helmholtz coil showing the magnetic ﬂux
density and ﬁeld (arrows) distribution

89

Magnetic ﬂux density, norm [T]
1 5 x10'4

 

 

A
'00

_L
N

Magnetic ﬂux density, norm ['l',
2 :2
(n 0'!

d
I
—L

 

1 .05

 

 

1 1 . 1 z z i i
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8
x
Figure 5 .6 Magnetic ﬂux distribution at the center of the Helmholtz coil along the x axis
on the y=z=0 line

 

90

Magnetic flux density, norm [T]

 

1.5

 

145. ......................................................... . ...................................................................................................
I c I o 0 o l O
.

. . . .
.............................................................................................................................................................

_L
In.

..............................................................................................................................................................

_I.
'0)
U"

.............................................................................................................................................................

Mametic flux density, norm [T
3

n...... o.......u............-........--........o.......-.n..............-.........-.-.o--............................v................v..._ ..................
. .

d
ix)

. . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . t
. . . . . . .
j. ......................................................... . ...................... . ............................................................................. .4
. . . . . . .
. e 0 v e i v I
. , . . . . .
. . . . . . .
. . . . . . .
. . . n . e
. . . v

 

 

 

1 .1 i i 1 i 1 1 2
-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8

2

Figure 5.7 Magnetic ﬂux density at the center of the Helmholtz coil along the Z axis on
the x=y=0 line

91

0.05 _
W

0.041;- (r \p

 

 

0.03q _____ \u.

0.02
q

0.01 \l

 

 

 

 

 

 

 

 

 

Figure 5.8 Finite element model of a lBSCC heart valve submerged in a volume of
uniform magnetic ﬁeld

92

 

 

 

 

 

 

 

 

 

 

 

A A A A A A A A A
10 ~ A A A A A A A A A
A A A A A A A A A
A A A A A A A A A
5 h A A A A A A A A A
A A A A A A A A A A
E 0 ~ A A A A A A A A A
V A A A A A A A A A
_5 _ A A A A A A A A A
A A A A A A A A A
A A A A A A A A A
-1 0 1.
-10 -6 6 5 10
(mm)

Figure 5.9 Real part of the (By, Bz) components in the x = 0 mm plane for both IOS and
SLS cases. The vertical axis is the y-axis. In these cases, the real components of the
ﬁelds are unperturbed.

93

10" \ x \ x J \ \ x \ t
\ \‘ \ \ \\ \ \ \ \ 3
5 \ -° \\ \ \ \ \ \ 1 1
' K / ,1 1 l i \ i ‘ ‘
E ' I ,/ 1 1 1 1 1 1 1 1
0* 1 1 1 1 1 1 1 1 1 1

g 1
I I 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1

-5-

1 1 1 1 1 1 1 I 1 1 ‘
1 1 1, 1 \ y '/ / / I 1
'10 i i \ \ x J / ’ l l

 

 

-10 -5 o 5 10
(mm)

Figure 5.10 Imaginary part of the (By, Bz) components in the x = 0 mm plane for an IOS
valve. The vertical axis is the y-axis.

94

10

1. 1;I//J\\\\\
(J’//$x\\\t
1 ii

5_ 1 11111\1
11111111111
11111111111

 

 

-5_ 1 1 1 1 1 1 i, i i / i
1 1 \ \ 1 1 j y 1 1 1
1 1 x \, \ i / y I 1 1
-10” 1 1 \ \ \\/ / 1 1 1
-10 -5 0 5 10
(mm)

Figure 5.11 Imaginary part of the (By, Bz) components in the x = 0 mm plane for a SLS
valve. The vertical axis is the y-axis.

95

1o~ , /’//< ‘1/ l

14’ \ ’ l
A
5 \\/1 \\£/\‘
5,1\ \‘I‘ /
’ / x, f\/
E‘ --
£0
p/ x. /\—-‘1’
...ft
-5 /
\I‘ ;\f /\<'-
,,\..o 1 \

 

 

 

 

-10 -5 o 5 10
(mm)
(a)
10-
5.
’5‘
3 °* ’
I ’ ‘ \
'SF . /1 r\\ l .
1/
-10-
-10 -5 o 5 10
(mm)
(b)

Figure 5.12 Real (a) and imaginary (b) parts of the (Bx, By) components in the z = -5 mm
plane for an IOS valve. The vertical axis is the y-axis.

96

 

 

10.. s / ,. . _. . . 1 \ \ ‘
\ i 1 1 , r
f
5’- ‘ I \ x I O
f ‘ ,. l
E
.6, °”
4 ‘ v - K I
- \ a . . . \ r
-5.
\ " ‘ ‘ a ’ ‘“ \
\ / ,_. x
,77‘ I /
-10* / / I “ ’ 1 1 1 "'
-10 -5 0 5 10
(mm)
(a)
10-
I 1 I 1 I I
5’- \ l ’l l 1 1 1 I 1 i f
f f I r \ . i \ l x
E 0” W, x 'e’
1% I 4 r 1 1 J x
J I 1 J ‘ ‘ l 3 \
-5r ‘
’ I I J ‘ 1 1 1
1 1
-10~ 1

 

 

-10 -5 0 5 10
(mm)

(b)
Figure 5.13 Real (a) and imaginary (b) parts of the (Bx, By) components in the z = -5 mm
plane for a SLS valve. The vertical axis is the y-axis.

97

CHAPTER 6. SYSTEM COMPONENTS

The gradiometer-based prosthetic heart valve examination system consists of two
subsystems, the excitation and measurement subsystem, both controlled by a personal
computer (PC). Figure 6.1 shows the block diagram of the examination system. The
excitation coils, part of the excitation subsystem, are used to establish a region of uniform
magnetic ﬁeld. A heart valve is placed in this region and the gradiometer, part of the
measurement subsystem, measures the gradient of the perturbations in the uniform ﬁeld.
The measurement subsystem then processes the gradiometer signal and provides an

indication relating to the condition of the outlet strut of the valve.

During the examination, the gradiometer-based evaluation system uses a custom-built
mobile frame to support the excitation coils, and a modiﬁed surgery table to support the
patient. The support frame adds mobility to the excitation coils with minimal
inconvenience to the patient. The modiﬁed surgery table provides multiple degrees of
freedom for positioning the patient to align the heart valve with the applied magnetic
ﬁeld. Since the table top is inserted between the excitation coils during the examination,
the top is made with no metallic parts to ensure that it does not interfere or distort the

applied ﬁeld.

This chapter describes the components of the gradiometer-based evaluation system. The

characteristics and performance of this system are presented in Chapter 8.

98

   
  
  

Excitation

Measurement Subsystem

Subsystem

Coils

Gradiometer

Figure 6.1 System block diagram of the gradiometer-based prosthetic heart valve
examination system (components not drawn to scale)
6.1 The Excitation Subsystem
The excitation subsystem consists of a signal generator, power ampliﬁer, an impedance
matching circuit and a pair of excitation coils. The signal generator creates a sinusoidal
excitation signal that is ampliﬁed by the power ampliﬁer. The output of the ampliﬁer is
applied to the excitation coils through an impedance matching circuit. The impedance
matching circuit ensures that the maximum power is transferred from the ampliﬁer to the
excitation coils. The excitation coils establish an alternating magnetic ﬁeld with a
uniform region at the center. The heart valve is placed in this region and the applied

magnetic ﬁeld induces eddy currents in the outlet strut that in turn perturb the applied

99

ﬁeld. The measurement subsystem then detects and processes the perturbation to

determine the condition of the outlet strut.

6.1.1 Signal Generator

The signal generator is capable of producing a sinusoidal waveform with a maximum

peak-to-peak output voltage of up to 20 V. An Agilent1 33220A arbitrary waveform
generator (AWG) serves as the excitation source. The operating frequency of the system
is typically 62.5 kHz, but may vary between 40.0 to 65.0 kHz. Equation (4.39) shows
that the gradiometer output increases with increased excitation frequency. This

characteristic is examined in Chapter 8.

6.1.2 Power Ampliﬁer
The power ampliﬁer must supply a minimum of 1050 watts to the impedance matching
circuit and the excitation coils at an equivalent impedance between 2.0-8.0 Q. The

ampliﬁer must have a bandwidth of at least 80.0 kHz, and capable of supplying an output
current of 15.0 Arms or larger, with a maximum output voltage of 70.0 Vrms or higher.
The ampliﬁer2 in the gradiometer-based examination system is constructed from two

Crown M-600 power ampliﬁers interlocked serially into a master-slave (push-pull)

conﬁguration known by the manufacturer as the M-2000 setup. The M-2000 can provide

 

1 Agilent Technologies, Incorporated, Palo Alto, CA.

2 Alternate power ampliﬁer suggestions: (1) AE Techron 8604 gradient ampliﬁer, AE
Techron, Incorporated, Elkart, IN; (2) QSC PL 236a power ampliﬁer, QSC Audio,
Incorporated, Costa Mesa, CA.

100

 

 

up to 140 Vrms at over 2000 W output power (into a 4-52 load) and a bandwidth of up to

100.0 kHz. The M-2000’s line voltage (110V AC) power requirement makes the

ampliﬁer system portable and suitable for use in any examination facility.

6.1.3 Impedance Matching Circuit
An impedance matching circuit (Figure 6.2) is used to ensure that the total load
impedance of the circuit and the excitation coils is matched to the output impedance of

the ampliﬁer at the excitation frequency. The circuit consists of a power resistor bank,

Rr, and a bank of power capacitors, Cr. The resistor bank consists of ﬁve 300W 1 Q

. 1 . . . . .
power resrstors connected 1n parallel, provrdlng an equ1valentre513tance of 0.2 Q. The
resistor bank is crucial in maintaining the stability of the excitation circuit by minimizing
power oscillation between the capacitors and the excitation coils when their impedances

are matched at the excitation frequency. The power capacitor bank consists of thirty ﬁve

2000 VAC 250.0 nF capacitorsz, providing a capacitance of approximately 7.143 nF
when connected in series. The exact number of capacitors used is a function of both the
excitation frequency and the inductance of the excitation coils. For a given operating

frequency, f, the number of capacitors required is such that the impedance of the

 

1 Part number 18-136-1AR/WBRKT, 300W, Milwaukee Resistor Corporation,
Milwaukee, WI.

2 Part number SCRN 246R, 2000 VAC, 250 nF, Cornell Dubilier Electronics,
Incorporated, Liberty, SC.

101

capacitor bank, X Cr, is matched to that of the excitation coils, X L, according to Equation

 

(6.1):
27rf X = 1
L 27If XCr
1
X = 6.1
Cr (27rf)2 XL ( )

 

From O—% l _0—
power Cr Rr R
ampliﬁers
L
z. ,.

 

 

 

 

 

 

Impedance Excitation
matching circuit coil

Figure 6.2 Impedance matching circuit. R1- denotes a bank of power resistors, and Cr
denotes a bank of power capacitors.

6.1.4 Excitation Coils

A large Helmholtz coil is employed to generate the uniform magnetic ﬁeld for the
gradiometer-based examination system. Each excitation coil has a mean radius of l m.
The two coils of the Helmholtz system are electrically connected in series and spaced 1 m
apart. The input excitation signal causes an alternating current to ﬂow in the coils which

in turn generates an alternating magnetic ﬁeld with a uniform volume at the center.

102

Two sets of excitation coils have been constructed, one with a 4/0 AWG wire and the
other with 1/0 AWG wire. Each excitation coil consists of 9 turns of the respective wire.
Both sets of excitation coils have the same Helmholtz conﬁguration and similar
electromagnetic properties. The excitation coils constructed with the 1/0 AWG wire has
a lower DC resistance, which amounts to a lower overall load impedance when connected

to the impedance matching circuit. For example, at an excitation frequency of 60.5 kHz,

the overall impedance1 of the excitation circuit with the 4 AWG wire excitation coils is
9.54 0, while that for the circuit with the 1/0 AWG wire coils is 5.52 Q. A smaller load
impedance allows for a closer match to the ampliﬁer output impedance, resulting in a
more efﬁcient power transfer from the ampliﬁer to the excitation circuit. The following

discussions pertain to both sets of excitation coils.

6.1.5 Helmholtz Coil

The Helmholtz conﬁguration produces a spatially uniform magnetic ﬁeld in a cylindrical
region extending between the centers of the two coils and approximately 1/4 to 1/2 of
their diameter as illustrated in Figure 6.3 and obtained numerically in Figure 5.5. For a
pair of N—tum coils each of radius R and a current I, the magnetic ﬁeld at any point on the
axis of an ideal Helmholtz coil may be calculated by assuming each coil of the Helmholtz
pair is a simple current loop made from a wire of inﬁnitesimal diameter and applying the
Biot-Savart Law to obtain the magnetic ﬁeld along the centerline of each loop. The total
magnetic ﬁeld is then obtained by superimposing the ﬁelds from the two loops. Hence,

the on-axis magnetic ﬁeld of the Helmholtz coil is given by [129], [185]

 

The overall impedance includes contributions from the connection terrmnals, Wires, and
connectors.

103

,u I 1 1
- 2‘; 3 + (6.2)

 

 

 

 

 

z . . . . .
where y = R, z 18 the distance on the ax15 from the center of the two cods , and B 18 the

magnetic ﬁeld in Tesla (T). The direction of the magnetic ﬁeld is along the axis of the

two coils. The constant ,uo = 41: x 10'7 Wb A’1 m'1 (or H/m) is the permeability of free

space. For the special case when 2 = 0, the magnetic ﬁeld at the center of the coils may

be estimated by the Helmholtz coil equation [129]

N
B-S’UO I

_ 51/5R

 

(6.3)

Although Equation (6.3) may be used to predict the magnetic ﬂux density, the equation
does not provide any dimensional information on the uniform magnetic ﬁeld volume
between the coils. To obtain this information, more sophisticated treatments such as the
ﬁnite element analysis described in Chapter 5 and the methodologies given in [176] and
[177] may be used. Chapter 8 presents the volume of the uniform magnetic ﬁeld

obtained by physical measurements.

104

 

 

Figure 6.3 Illustration of the magnetic ﬂux lines generated from a Helmholtz coil. After
[129].
The gradiometer-based examination system employs two 9-tum excitation coils each with

a mean radius of 1 m. The input is an ll-A peak-to-peak sinusoidal current. With these

parameter values, Equation (6.3) predicts a magnetic ﬁeld of approximate 0.89x10- T

(0.89 gauss). This value is smaller than that obtained from the ﬁnite element analysis
results shown in Figure 5.5-Figure 5.7. The disagreement between the two values is due
to the fact Equation (6.3) was derived with the assumption that the excitation current was
carried in a wire of inﬁnitesimal diameter. In reality, the excitation coil wire has a ﬁnite
cross-sectional area as shown in Figure 5.3. The contribution to the overall magnetic
ﬁeld from the current ﬂowing in the ﬁnite-thickness wire must be accounted for. This
had been done in the ﬁnite element analysis and the results were reported in Chapter 5.
Physical measurement of the magnetic ﬂux density in the Helmholtz coil with an input

current of 11 A (peak-to-peak) has also conﬁrmed a center ﬂux density of approximately

105

1.45 gauss as predicted by the ﬁnite element model. This and other experimental results

for the gradiometer-based examination system are described in details in Chapter 8.

The excitation coils are each wound on a 2-meter diameter mounting ring. The mounting
rings are separated by l m with support beams as shown in Figure 6.4 (the cross and
radial beams shown mounted inside the excitation coils in this ﬁgure are used for
specimen testing. Details and results of these tests are reported in Chapter 8). The entire
coil mount structure is modular and can be easily disassembled and transported in parts
no larger than 4’ by 2’. When assembled, the coil mount is capable of supporting more

than 150 lbs of wire.

6.2 Excitation Coils Mobile Custom Support System
The excitation coils are supported on a mobile custom support frame. During the
examination, the excitation coils are mounted on a 45° inclination (see Figure 6.5) to
facilitate the alignment of the magnetic ﬁeld with the ﬂange plane of the heart valve. This
orientation minimizes the amount of eddy current induced in the ﬂange ring and ensures
that the current induced in the outlet strut is dominant. The dimensions of the coil
support framework are shown in Figure 6.6. Fine tuning of the alignment may be made
by adjusting the roll and tilt of the custom examination table described later in this

chapter.

The excitation coils support system is designed to accommodate both the patient and the

custom examination table (Figure 6.7). The support system is constructed on a mobile

106

platform (Figure 6.8) to allow movements of the excitation coils while minimizing the
disturbance to the patient. The platform casters each has a leveling pad (Figure 6.9) that
can be raised and lowered using a telescoping ratchet handle for securing the support

structure in place once it is properly positioned.

6.3 The Measurement Subsystem
The measurement subsystem consists of the gradiometer, lock-in ampliﬁer and a data
acquisition system (DAQ). The signal from the gradiometer is ampliﬁed and processed
by the lock-in ampliﬁer. The outputs of the lock-in ampliﬁer are then digitized by the
DAQ and transmitted to the PC where the system software displays, processes and
analyzes the gradiometer signal to determine the outlet strut condition of the heart valve.
Two gradiometers have been prepared for use with the gradiometer-based examination
system. The designs of the gradiometers are described in the next chapter. The

remaining Measurement Subsystem components are described in the following sections.

6.3.1 Lock-in Ampliﬁer

The signal from the gradiometer is processed by a 100 kHz bandwidth dual phase lock-in
ampliﬁer (LIA). Lock-in detection1 is a phase-sensitive demodulation technique. The
dual phase ampliﬁer uses two phase sensitive detectors to detect and process the
gradiometer signal with reference to the input excitation signal. The principle of lock-in

detection is described below.

 

l . . . .
Lock-in detection 1S also commonly known as quadrature detection, quadrature
demodulation, phase quadrature technique, or phase quadrature heterodyne demodulation

107

 

Figure 6.4 Side (a) and top (b) views of the excitation coils. The excitation coils are
mounted on a modular wooden structure arranged in a Helmholtz conﬁguration.

108

 

Figure 6.5 Excitation coils are mounted with an inclination to facilitate the alignment of
the magnetic ﬁeld with the ﬂange ring of the heart valve.

 

 

 

2.12 m——>

 

Figure 6.6 Dimensions of the excitation coils and the mobile custom coil support system

109

 

Figure 6.7 Custom examination table positioned between the excitation coils. The
excitation coils system is designed to accommodate both the patient and the custom
examination table.

 

.._ ,y-lf-I. .‘F ““1..." 1' f’p‘l}.::". Vﬁ‘i‘v I ‘
Figure 6.8 A mobile support frame allows the positioning of the coils around the patient
and the custom examination table with no disturbance to the patient.

110

    

....ar. 1.1"" ;; V:- ,‘ (‘.“;«_;-l-b~\;' ,--.' 1.. ,

Figure 6.9 Each platform caster has a leveling pad that can be raised and lowered by a
telescoping ratchet handle for parking the support structure.

6.3.1.1 Principle of Lock-In Detection
Lock-in detection is performed with respect to a reference frequency, (UR , which is the
excitation signal in this case. Let the input signal be VS (1‘) = Psin(a)st + 03) where P, (as
and 6's are the signal amplitude, frequency and phase, respectively. In a lock-in
detection system, two lock-in signals are generated from (UR . The ﬁrst signal is
VR1(t) = VR sin(a}Rt + 6R) with amplitude VR and phase 6R. The second signal VRZ (t)
is VR1(t) phase shifted by 90°:

VR2 (I) = VR1(t +90°)

= VR sin(th +6R +90°)

VR2 (t) = VR cos(a)Rt + 6R) (6-4)

The input signal, Vs (t) is directed into two quadrature multipliers, A and B, respectively.

In multiplier A, the input signal is multiplied with VRI (t) and the output, X, is the sum of

111

two signals, one at the difference frequency (£05 — a) R ) and the other at the sum

frequency ((05 + (UR):

X = PVRI 8111((Ust +65 )sin(a)Rt +61?)

1 l
=§PVR1 003((603 'wR)’)+BS ”911%; PVRi 905((‘05 +wR)‘)+BS +61?) (6.5)

A low-pass ﬁlter is then applied to X to remove the high frequency component. If the
reference frequency is the same as the signal frequency, i.e., a) R = 605 , the component at
the difference frequency will be a DC valve. In the case if the input contains signals of
other frequencies as well as noise, the low-pass ﬁltering will remove all but the signal at
(”R . Also, setting the reference signal amplitude, VRl , to 2, leads to the ﬁltered
multiplier output

X = Pcos (65 —6R) (6-6)

which is a DC value proportional to the signal amplitude, P.

In multiplier B, the input signal is multiplied with VRZ (t). A subsequent low-pass ﬁlter
produces the output

Y = Psin ((95 — 6R) (6-7)
The two multiplier outputs, X and Y, represent the signal as a vector relative to the lock-in
reference signal. The output X is known as the “in-phase” component while Y is the
“quadrature’l’ component. Let the phase difference between the input and the lock-in

reference signal be 6, i.e., 6 = 65 — 6R. (When 6 = 0, the in-phase component measures

112

the signal while the quadrature component is zero.) To remove the phase dependency,

the magnitude, R, can be used:

R=(X2+Y2)y2 (6'8)

Substituting (6.6) and (6.7) into (6.8) shows that R is:
R = p (6.9)
which is the amplitude of the gradiometer signal. Also, the phase, 6, between the input

and the lock-in reference signal is given by

6: tan—1(Y/X) (6-10)

6.3.1.2 Lock-in Detection of the Gradiometer Signal
From the previous section, the relationship between the input and output signals of the
LIA is summarized as follows. Let the LIA input signal, VS (1‘), be the signal from the
gradiometer. Using the excitation signal from the signal generator as the reference, the
LIA outputs from (6.6) and (6.7) can be written as
X = R cos(6’) (6.11)
Y = R sin(6) (6.12)
where 0 is the phase angle between VS (t) and the reference signal, and R is given by

Equation (6.8). The outputs of the LIA, X and Y, are then digitized and analyzed to

determine the condition of the heart valve. When the gradiometer is placed directly in

front1 of the outlet strut, “large” values of X and Y indicate an IOS valve while “small”

 

1 In reality, the gradiometer will not be stationary in front of the outlet strut. Instead, the
relative distance between the two will be time-varying in a cyclic manner due to the
cardiac cycle. More discussions on the effect of the varying distance between the
gradiometer and the outlet strut are given in Chapters 4 and 8.

113

values indicate a SLS case. (The relative magnitudes of the “large” and “small” signals
are system- and gradiometer-dependent, and are obtained by pre-calibrating the
examination system with an IOS and a SLS valve, respectively, during system setup. A
detailed discussion of the signal level in relationship to the strut condition is given in
Chapter 8.) Large LIA outputs are results of a strong gradiometer signal at the excitation
frequency. This strong signal indicates a signiﬁcant perturbation in the gradient of the
uniform excitation ﬁeld, which is indicative of an IOS valve. On the other hand, small
and rapidly ﬂuctuating LIA outputs indicate a weak gradiometer output signal at the
excitation frequency. This implies a minimal perturbation in the applied magnetic ﬁeld,
which in turn indicates little or no eddy current is circulating in the outlet strut—a result
indicative of a single-leg separation in the outlet strut. Chapter 8 explains how to

interpret the outlet strut condition based on the LIA outputs. The gradiometer-base

system uses an ITHACO, Incorporated, Model 3961 lock-in ampliﬁer].

6.3.2 Data Acquisition System

The Data Acquisition System consists of a digital oscilloscope for real-time signal
monitoring and a PC-based data acquisition card. The digital oscilloscope displays the X
and Y output signals from the lock-in ampliﬁer, and can optionally store these signals.
Two channels from a 14-bit, 2 MS/s, 4-channel simultaneous sampling data acquisition
card is used to digitize the X and Y output signals from the lock-in ampliﬁer. The third

DAQ channel monitors the current passing through the excitation circuit by sampling the

 

1 Alternative lock-in ampliﬁer selection: Stanford Research Systems SR 830 DSP dual
phase lock-in ampliﬁer, Stanford Research Systems, Incorporated, Sunnyvale, CA
94089, www.thinksrs.com.

114

output of a current probe. The fourth channel records the electrocardiogram (ECG)
signal of the patient during the examination for future reference. The Gradiometer-based

examination system uses a HP 54710A/54713A digital oscilloscope by Agilent
Technologies, Incorporated1 (formerly Hewlett-Packard, Incorporated), the NuDAQ

DAQ-2010 PC-based data acquisition card by ADLINK Technology, Incorporatedz, and
the Agilent 1146A AC/DC current probe also by Agilent Technologies for data

monitoring and recording.

6.4 Custom Examination Table
A vintage surgery table had been modiﬁed into a custom examination table for patient
support during the heart valve examination. The original table had a metallic top that
would interfere with the excitation ﬁeld when inserted between the excitation coils. The
tabletop and other metallic sections of the table that would affect the examination results
were stripped off and a new top has been fabricated using non—conducting composite
materials. The new top contains high density foam and Luan set lengthwise at
approximately 6” intervals and wrapped around by ﬁberglass and resin. This creates a
lightweight and yet strong and non-conducting bed that is slightly thicker than 5 cm.
Figure 6.10 shows the modiﬁed vintage surgery table with the composite table top. The
table top is 2.48 m (97.5 in) long and 0.52 m (20.5 in) wide as shown in Figure 6.11. The
ﬁnished examination table is shown in Figure 6.12. In this ﬁgure, a custom tailored black
vinyl cover is ﬁtted to provide the table with an even and smooth surface and to allow for

easy cleaning and maintenance.

 

1 Agilent Technologies, Incorporated, Palo Alto, CA 94306, www.agilent.com
2 ADLINK Technology, Incorporated, Irvine, CA 92618, www.adlinktech.com

115

 

Figure 6.10 Custom examination table with a nonmetallic composite top that does not
interfere with the applied magnetic ﬁeld

 

2.48 m (975')
0.52 m (20.5")
——(
0.84-1.12 m

(33.25"-44.0")

5n,- ;.. .. q
'1 :
nut-mm. 1 in m :1: wall

Figure 6.11 Tabletop dimensions and height adjustment range for the custom surgery bed

The examination table can be raised and lowered from 0.84 m (33.25 in) to 1.12 m (44 in)
relative to the ground by a hydraulic mechanism. The top can also be tilted in the head—
to-toe fashion (pitch). Additional modiﬁcations have been made to provide a 10—degree
side-to-side tilting (roll) freedom in either direction. Two large wheel handles on the side

of the bed are used to adjust the pitch and roll of the tabletop. This positioning ﬂexibility

116

allows precision positioning of the patient to align the ﬂange plane of the heart valve with

the applied magnetic ﬁeld during examination.

To ensure safety and stability, a custom bracket has been welded underneath the base of
the table (Figure 6.13). This bracket accommodates a wooden forward T-shape extension
(Figure 6.14) that provides additional support and stability for the table when in use. The
extension is 1.85 m (73 in) long and 1.22 m (48 in) wide. Figure 6.15 shows the custom
surgery table with the T-shape wooden extension in place. During positioning and

transportation of the table, the arm extension maybe removed for mobility.

 

(b)

Figure 6.12 Two views of the ﬁnished custom surgery table. Two large wheel handles on

117

the side are for ad ustin the oitch and roll of the tableto -

         

Figure 6.13 A bracket is welded underneath the base of the custom examination table for
accommodating a wooden extension arm that provides additional support and stability.

 

Figure 6.14 A T-shape wooden extension is used to increase the stability of the custom
surgery table when in use

118

 

u“. 1..“ . . ’2. .-; . j .; :;.._.{‘.>.~,; -. .; : _ , $21119); .,: 3 .~

 

Figure 6.15 Custom surgery table with a T-shape wooden extension for increased
stability

A clinical-scale prototype of the gradiometer-based examination system has been
constructed from the components described in this chapter. The design of the
gradiometer is presented in the next chapter. The operating parameters and settings of

these components are described in Chapter 8.

119

CHAPTER 7. GRADIOMETER DESIGNS

Measurement of small magnetic ﬁelds with a single sensor often poses difﬁculties due to
the presence of large interfering ﬁelds such as the earth’s ﬁeld or other man-made
sources. The task becomes even more challenging if either the source of the small
magnetic ﬁeld or the senor, or both, is non-stationary. A small change in the relative
orientation between the source and the sensor in the presence of interference ﬁelds is
likely to produce a signal that is signiﬁcantly larger than the actual ﬁeld variations under
investigation. To circumvent the effects of the interference ﬁelds, a gradiometer may be

used.

A gradiometer consists of two identical and perfectly aligned coils, or sensors. These two
sensors are electrically connected such that their outputs are subtracted from one another.
When submerged in a uniform ﬁeld, the two sensors produce identical outputs and the
gradiometer output will be zero, effectively removing the inﬂuence of the uniform ﬁeld.
If the positional relation between the two sensors remains ﬁxed, the gradiometer may be
rotated and translated within the uniform ﬁeld without producing any orientation-

dependent output.

In the gradiometer-based prosthetic heart valve inspection system, the heart valve is
submerged in a uniform magnetic ﬁeld and a gradiometer is placed in the vicinity of the
outlet strut. A current is induced in the outlet strut and this current in turn creates a
magnetic ﬁeld. The induced magnetic ﬁeld perturbs the applied uniform ﬁeld and the

gradiometer measures the gradient of resulting ﬁeld in the vicinity of the strut. Since the

120

induced current is a function of the outlet strut integrity, the gradiometer output signal,

therefore, can be used to characterize the condition of the outlet strut.

7.1 Gradiometer Output Signal
The mathematical model of the gradiometer-based system described in Chapter 4
provides a starting point for designing the gradiometer. The design goal is to maximize
the gradiometer’s signal level. Although the gradiometer output is a function of many
physical and system parameters, the gradiometer design begins with the factors in
Equation (4.41) that are related to the geometry, position and separation of the sensor

coils.

From Equation (4.39), the physical qualities of the gradiometer that affect its output
signal are the number of turns in each sensor coil, the area of each coil, and the separation
between the two coils. By keeping all other factors constant, Equation (4.57) expresses

the gradiometer output vg as a scaled signal U(z1,Az) that is a function of both its

distance, 21 , from the outlet strut and the separation, Az, between the two sensor coils.
For values of zl from 0—10 mm, Equation (4.57) is evaluated for Az from 0—20 mm. The

results are shown in Figure 7.1. Figure 7.2 displays the same results in the 2D plane.
From this ﬁgure, the strongest output signal (the brightest region in the ﬁgure) occurs
when the gradiometer is 4 mm or less away from the outlet strut and the separation
between the sensor coils is approximately 3 mm or more. These results, in conjunction
with other physical constraints, are taken into consideration when designing the

gradiometer.

121

Scaled Gradiometer Output Signal U(2,,Az)

  

uary Un'

 

 

21 (mm) 0 0 A2 (mm)

Figure 7.1 Scaled gradiometer output U(z1,Az) as a function of Az at values of zl

Scaled Gradiometer Output Signal U(z1,Az) -3

   
 

|H||||||lllllllllllllllllllllllllllllllllllllllllllll||||||l||llllllllllllllllllllllllllllllllllll|||||||l|||l||||||lllllllll||||l|||lllllllllllllllllllllllllllllllllllllllllllll
lllllllllllllllllllllllllll|I||l||||lllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll ~ .
5 llllllllllllllllllllllllllllllllllll||||||l||Illllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll|||l|||||llllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll v
lllllllllllllllllllllllllllllllllllllllllllll[lllllllllllllllllIlllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll
4 lllllllll|||||||||lllllllll|||l||l|||||l||||llllllllllllllllllllllllllllllllllllllllllllll||||||l||llllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll
|l|||||||llllllllllllllllllllllllllllllllllll|||||l||||||l||l|H|||I||||l|l||l||||||l|l|||lllll||Illllllllllllllllllllllllllllllllllﬂlllllllllllilllllllllllm
lllllllllH||||||||||||l||l|llllllllllll I 7 ,
llllllllllllllllllllllllll 1 1
lllllll llllll

21 (mm)

   

      

 

Az (mm)
Figure 7.2 A 2D view of the data from Figure 7.1 showing the strongest output signal
occurring when 21 < 4 mm and A2 2 3 mm

122

7 .2 Physical Constraints
Besides the signal level and quality, there are three primary physical constraints that
affect the design of the gradiometer. First, as the gradiometer is placed in the mitral
annulus (in the vicinity of the outlet strut) during the prosthetic heart valve examination,
the gradiometer sensor coils and signal leads must be sufﬁciently small so that they can
be delivered and positioned inside the left ventricle of the heart via conventional cardiac
catherization procedures. Second, the lead wires of the sensor coils must be shielded
from the excitation ﬁeld to minimize interference. Third, the gradiometer assembly must
be of proper stiffness—neither too soft nor rigid—to aid the delivery and positioning of
the sensor coils. Based on these constraints, two gradiometer assemblies have been
constructed that are suitable for clinical use. In the ﬁrst assembly, the gradiometer sensor

coils, the lead wires and a thin layer of electromagnetic shielding material are encased

and sealed inside a 9 F1 pre-curved guide catheterz. The second assembly delivers a 6 F
gradiometer with a shielded specialty cable into position via a pre-curved 6 F cardiac
guide catheter. The design and speciﬁcations of these gradiometer assemblies are given
below. Characterization and performance tests results of the gradiometer-based

examination system with both gradiometers are presented in Chapter 8.

 

1 “French,” abbreviated as either “F” or “Fr,” is a unit of length measure typically used in
the medical community (e. g., in catheter sizes). 3 F = 1 mm.

2 Size for a catheter is both manufacturer and type dependent. For a conventional

catheter the size often refers to the outer diameter while that for a guide catheter refers
to the lumen (inner) diameter.

123

7.3 Delivery Approaches
To satisfy the physical design constraints, the gradiometer size and the delivery and
positioning approaches are ﬁrst determined. For precise positioning of the gradiometer in
the left ventricle, conventional cardiac catherization techniques from interventional
cardiology are employed. The gradiometer is delivered either intraluminally or via a
guide catheter into the ventricle. Circulatory access is obtained via either a femoral or
carotid cutdown followed by cannulation. Due to the limited arterial lumen clearance,
the outer diameter of the gradiometer assembly is limited to approximately 9 P, which
limits the catheter lumen size to less than 8 F. The gradiometer assembly is advanced
inferiorly into the left ventricle via the aorta and through the aortic valve until it reaches
the apex of the heart (Figure 7.3). The distal end of the assembly is then turned to follow
the inner walls of the ventricle and advances superiorly towards the mitral annulus.
Figure 7.4 and Figure 7.5 show two approaches for delivering and positioning the
gradiometer assembly. In Figure 7.4, a sealed catheter-encased gradiometer assembly is
advanced intraluminally into the ventricle and positioned into the mitral annulus. In
Figure 7.5, a guide catheter is pre-positioned in the left ventricle using conventional
catherization technique. A gradiometer is then advanced through the guide catheter and
positioned into the mitral annulus. The designs and beneﬁts of these two gradiometer

assemblies are discussed below.

124

Aorta

“U (main artery
taking blood
Pulmonary X to the body) Pulmonary artery

 

valve , \, which takes blood
_ I/ , ,, 7 to the lungs
Superior vena cava r /\ ,. ,-»
(main vein ' Pulmonary veins bring
which brings ' 7 blood into the heart
blood into 6 fromthe lungs
the heart from
the head and Left atrium
neck) ‘\ ‘
Aortic vat/e
Right atrium

7

Interior vena cava

. Mitral valve
Left ventricle

Figure 7.3 Advancement of the gradiometer assembly inside the heart (black arrow)

(main vein

which brings
blood into

the heart from Tricuspid
the body) valve

  
   

Right ventricle

 

REV-u. _ ..r
Figure 7.4 A sealed catheter-encased gradiometer assembly is delivered intraluminally
and positioned in the mitral annulus. After [195]

125

 

Figure 7.5 A guide catheter is pre-positioned in the left ventricle and the gradiometer
(circled) is advanced via the catheter into the mitral annulus. After [195]
7.4 Gradiometer Assemblies
In a gradiometer, each of the two sensor coils consists of a length of wire wound in a
multi-layer helical manner around the length of a cylindrical epoxy former. When
subjected to a time-varying magnetic ﬁeld, an EMF is induced in each coil according to
Faraday ’5 law. The output of the gradiometer is proportional to the number of turns and
the area of each sensor coil at a given excitation frequency and ﬁeld strength. The design
goals are, therefore, to maximize the number of turns and the diameter of the coils while
satisfying the physical size constraints. To allow for proper maneuvering and to avoid
damages to the blood vessels and the heart, the size (outer diameter) of the gradiometer

assemblies are limited to 9 F [150] or less.

During the examination, the gradiometer signal cable is also subjected to the applied

excitation magnetic ﬁeld. To minimize the coupling of the applied ﬁeld into the signal

126

cable, two small-diameter specialty cables each containing a pair of shielded twisted
signal wires are used. Two gradiometer assemblies, each using one of these cables, have
been designed. The ﬁrst is a 9 F assembly with the gradiometer sealed inside a catheter.
The second is a 6 F two-piece assembly consisting of a cardiac guide catheter and a 6 F

(standalone) gradiometer. Details of these two catheter assemblies are given below.

7.4.1 The 9 F Gradiometer Assembly

Figure 7.6 shows the schematic diagram and Figure 7.7 shows a picture of the 9 F
gradiometer assembly. In this system, a gradiometer is encased and sealed in a 9 F guide
catheter to form an integral unit. Figure 7.8 shows the schematic diagram of the
gradiometer sensor head. The sensor head is 3.0 mm long and consists of two sensor

coils each made from 88 turns of a 44 AWG wire. The gradiometer output is the

difference signal from the two sensor coils. The typical1 resistance and inductance of the
sensor coils are given in Table 7.1. Each sensor is approximately 2.5 mm in diameter and
0.35 mm tall, and the two coils are separated by 1.62 mm. The epoxy cylinder (the

sensor head) embedding the coils are made from Ultem 1000 polymer. Ultem 1000 is a

thermoplastic polyetherimide high heat polymer designed by General Electricz, Inc. for
injection molding applications. The polymer is inert to bio-chemicals and suitable for

biological applications. It can also withstand medical sterilization processes.

 

l . . . .
Due to manufacturing variations such as the denSity and ﬁll-factor of the turns, the

inductance of each coil varies slightly. The data given in Table 7.1 are typical valves.

2 GE Plastics, Incorporated, Pittsﬁeld, MA, www.geplastics.com

127

Table 7.1Typical resistance and inductance of the 9 F gradiometer sensor coils
(Measured at 1 MHz)

 

Sensor Coil Resistance ((2) Inductance (pH)

 

Distal 10.375 7.775

 

Proximal 10.414 7.793

 

 

 

 

 

The 9 F gradiometer has been fabricated by Xactex, Corporation1 and assigned the part
number POT-9075 (Figure 7.9). In this gradiometer, the sensor signal cable is made from
a pair of 36 AWG wires, twisted and shielded inside an aluminum foil. The signal leads
are approximately 1 cm long. The ﬁrst 25.4 mm of this cable from the sensor head is
covered with a heat-shrink tube as shown in Figure 7.8. This section of wires has a
maximum diameter of 2 mm, which allows it to ﬁt inside a 9 F guide catheter that has a

typical lumen size of approximately 2.6 mm. The signal cable is followed by 512 cm of

an Alpha Wire Company2 number 5 8401 cable as shown in Figure 7.6. This cable

contains a foil-shielded twisted pair of 24 AWG wires. The catheter is a Medtronic,

Incorporated3, ZumaTM 9 F, 100 cm, J CL 4.5 SH guiding catheter (catalog number

ZM9JCL458H). This catheter has an outer diameter of 3.02 mm (9.06 F) and an inner

diameter of 2.62 mm (7.86 F). The gradiometer is sealed inside the guide catheter using

 

l Xactex, Corporation, Pasco, WA 99301, www.xactex.com
2 Alpha Wire Company, Elizabeth, NJ 07207-0711, www.alphawire.com

3 Medtronic, Incorporated, Minneapolis, MN 55432-5604, www.medtronic.com

128

a small amount of Aron Alpha1 series 232 medical-grade adhesive. The tip of the

gradiometer is also coated with this adhesive to protect the sensor coil. Figure 7.10

shows a close-up view of the 9 F gradiometer-in-catheter assembly. Table 7.2

summarizes the components for the 9 F gradiometer assembly.

Table 7.2 Components of the 9 F (POT-9075) gradiometer assembly

 

 

 

 

 

 

 

Component Manufacturer and Part Number
Gradiometer Xactex, Corporation, Part No: POT-9075
Data Cable Alpha Wire Company, Part No.: 58401
Medtronic, Incorporated, ZumaTM 9 F Guiding
Catheter
Catheter JCL 4.5 SH (Cat. no.: ZM9JCL458H)
Adhesive Aron Alpha, Series 232

 

 

1 Aron Alpha, a division of Elmer’s Products, Incorporated, Columbus, OH 43215,
http://www.krazyglue.com/industrial/industrial.asp

129

 

 

 

 

  
  
     
    
   

 

100 cm L 512 cm (16.8 it)
Foil Shielding I
0.0. = 2 mm -
3'3"“; gﬁ; Foll-Shloldod
. 9F Catheter ' ' Muted-Pair Cable
0.0. = 3.02 mm OD. = 259 mm BNC
ILD. = 3:521 mm ID. = 1 .83 mm Connectors
Shrink Tube Gndlomotor 6th — m
OD. =1 mm (Bobbin coils.
Length = 25.4 mm 44 .
85 turns each.
0.0. = 2 .5 mm)

Figure 7.6 Schematic diagram of the 9 F gradiometer assembly

.‘.2_

Figure 7.7 9 F gradiometer assembly

130

 

Bobb' '15 ~ 3:: mm
In cor =5 . mm
(44AWG,85 8 mm 53.0mn

tums each) '
, ~ 0.82 mm

 

 

  
   

Ultem 1000 polymer

 

254 mm (1 .o')

Shrink tube
(1p =5 2.0 mm)

 

 

2m (leads, 36 AWG, tightly twisted)
A

 

B
Figure 7.8 Schematic diagram of the 9 F gradiometer sensor head (POT-9075)

_vcf‘lidftw “1.1,.” 414-; d , l V

Figure 7.9 Close-up view of the 9 F gr diometer (POT-9075) sensor head

131

 

Figure 7.10 9 F gradiometer encased and sealed inside a guide catheter

7.4.2 The 6 F Gradiometer Assembly

Figure 7.11 shows the schematic diagram of the 6 F gradiometer assembly. The
gradiometer is shown in Figure 7.12 without the guide catheter. In this system, the
gradiometer is delivered via a pre-positioned 6 F guide catheter to the proper location in
the mitral annulus. Figure 7.13 shows the schematic diagram of the gradiometer sensor
head. The sensor head is 6.0 mm long and consists of two sensor coils each made
from107 turns of a 46 AWG wire. Table 7.3 gives the overall resistance and inductance
of the sensor coils. Each sensor is approximately 1.7 mm (5.19 F) in diameter and 0.5
mm in length, and the two coils are separated by 1.8 mm. Similar to the 9 F design, the
sensor head of the 6 F gradiometer is also made from Ultem 1000 polymer. Figure 7.14

shows a close-up view of the 6 F gradiometer sensor coils.

132

Table 7.3 Total resistance and inductance of the 6 F gradiometer sensor coils
(Measured at 100 kHz)

 

Resistance (Q) 18 038

 

Inductance (pH) 18.62

 

 

 

 

 

 

  
 

 
 

36 AWG Wire

 
  

     
  

 

TWiSted Pair 065 C3133; Junctlon Shielded
ID. 1 .78 mm Connector Twisted-Petr Cable
Length = 100 cm BNC
”1 0'2 5 cm Connectors
I ‘ Gradiometer

(Bobbin coils, 46 AWG.
107 turns each.
0.0. = 1.7 mm)

Figure 7.11 Schematic diagram of the 6 F gradiometer assembly

The 6 F gradiometer signal cable is made from a 150 cm long Cooner Wirel CZl215-2F2
wire, followed by 450 cm of a custom cable as shown in Figure 7.11. The CZ1215-2F
cable consists a braid—shielded twisted pair of 36 AWG wires coated with a clear PVC
jacket. The custom cable is made from shielding a Cooner Wire CW3735 cable with a
92166 S/S braid and covering the ﬁnal cable with a white 3520 SPC shrink tube. The
CZ1215-2F cable has a maximum outer diameter of 1.35 mm (4.04 F), and can pass

through a 6 F guide catheter (lumen size 3 1.73 mm) easily. The 6 F gradiometer has

 

1 Cooner Wire, Chatsworth, CA 91311, www.coonerwire.com

2 As of Fall, 2005, the revised part number for CZ1215-2F is AS633-ZSSF

133

been fabricated by Xactex, Corporation and assigned the part number POT-9083 (Figure

7.12).

Figure 7.15 shows an enhanced 6 F gradiometer sensor head with a tapered anchor to the
lead cable. The taper provides a smooth transition from the lead cable to the sensor head

to facilitate the retraction of the gradiometer into the guide catheter should it becomes

necessary (e.g., a gradiometer replacement during the examination).

. . . ,.
534 “if” -
. ‘ - ___r'

w. I .1 :5.
6’ “’0‘; "(MT/ll ii H 7v.‘

vi
8 ”+77" /

      

t ~ ~—..

Figure 7.12 6 F gradiometer, POT-9083

134

¢=1.7—1.73mm (Ultem 1000 Polymer)

   
  

0.5mm

Gold Shielded Cable (107 turns.
(CZ1215-2F) 46 awe,
0v5mm Bobbin Cell)

(1 07 turns,
46 ,
Bobbin Coil)
Figure 7.13 Schematic diagram of the 6 F gradiometer sensor head (POT-9083)

X.“

  
 

. "anew“: :imt )5»

   

135

¢=1.7-1.73 mm (Ultem 1000 Polymer)

  

. V v 11.: fJ'T-iﬁf‘i’m I ’ ' ‘ .
j 0.5 mm
3.0 mm 2.7 mm 1.8 mm (107 turns.
46 AWG,

0.5 mm 0-5 mm Bobbin Coil)
Gold Shielded Cable (Shield) (107 turns,
(CZ1215—2F) 46 AWG,

Bobbin Coil)

Figure 7.15 Schematic diagram of a modiﬁed 6 F gradiometer sensor head with a tapered
cable anchor
7.4.2.1 Catheter Selection for the 6 F Gradiometer

Two catheter candidates have been identiﬁed for use with the 6 F gradiometer. The ﬁrst
is a Cardima, Incorporatedl, 9 F, 90 cm Naviport® deﬂectable guiding catheter (product

number: 08-003045). This catheter has an outer diameter of 9 F, inner diameter of 5.3 F
(1.7667 mm, 0.070”), a deﬂectable tip with an adjustable degree of curvature that is
controlled by a knob at its handle, and a curve reach of 10 mm (see Figure 7.16). The
catheter’s deﬂectable tip adds maneuverability and allows precise positioning of the
gradiometer in the heart. In cases where a conventional catheter is preferred (e.g., by the
cardiac surgeon), an alternative catheter may be used. The alternate catheter is a Boston
Scientiﬁc Corporationz, Scimed, Mach 1TM 7 F, CLS 4.5, guide catheter (catalog
number: H749 34357-576 0). This catheter is 100 mm long, has an outer diameter of 7.0
F, and a lumen size of 2.057 mm (0.081 in, 6.17 F). Table 7.4 lists the components and

catheter selections for the 6 F gradiometer assembly.

 

1 Cardima, Incorporated, Fremont, CA 94538—7330, www.cardima.com
2 Boston Scientiﬁc Corporation, Natick, MA 01760-1537, www.bostonscientiﬁc.com

136

In practice, a cardiologist may have his or her own cardiac catheter preferences. One
advantage of the 6 F gradiometer is that it can be easily adapted for use with any cardiac
guide catheter with a lumen of 6 F or larger. This allows the cardiologist to precisely
delivery and position the gradiometer with a familiar catheter rather than needing to

adjust to a new catheter system.

7.5 Gradiometer Selection
The 6 F and 9 F gradiometer assembly each has its advantages when used in the
examination system. The selection of a particular gradiometer is both case- and patient-
dependent. The following discussions present several considerations to facilitate the

selection process.

 

 

 

 

 

Handle Length ‘ Working Length
[ 21mn l 'Cuwe
l .Reach
ENAVIPORT® _J r)
l l \
Control Knob Deflectable Soft Tip

Shaft

Figure 7.16 Cardima, Incorporated, Naviport® guide catheter. After [154]

137

Table 7.4 Component list for the 6 F (POT-9083) gradiometer assembly

 

Component Maunfacturer

 

Gradiometer Xactex, Corp., POT-9083

 

 

 

D::S(::?C Cooner Wire, AS633-2SSF (formerly CZl215-2F)
Data Cable Cooner Wire, CW 3735 With a 92166 S/S braid and a white type 3520
, SPC shrink tube
(prox1mal)
Boston Scientiﬁc Corporation, Scimed, Mach lTM 7 F CLS4.5 guide
catheter (catalog no. H749 34357-576 0)
Catheter

Cardima, Inc., 9F, 90cm, Naviport® 10 mm deﬂectable guiding
catheter (product no. 08-003045)

Boston Scientiﬁc Corporation, Scimed, WiseguideTM 7 F CLS4.5
guide catheter (catalog no. H749 19500-576 0)

 

Boston Scientiﬁc Corporation, Scimed, Mach 1TM 8 F CLS4.5 guide
Alternative catheter (catalog no. H749 34358-576 0)

Catheters
Boston Scientiﬁc Corporation, Scimed, Mach 1TM 6 F CLS4.5 guide
catheter (catalog no. H749 34356-576 0)

 

 

 

(Any guide catheter with a 6 F or larger lumen)

 

7.5.1 Practical Considerations

The 6 F and 9 F gradiometers each has physical characteristics that may affect its
performance in a particular examination. Speciﬁcally, the 9 F gradiometer assembly has
a larger outer diameter that makes the unit more rigid and less affected by blood
turbulence, especially when the gradiometer assembly is positioned downstream from the
mitral annulus. The 9 F assembly is also suitable for use with a larger heart, in which the
catheter assembly can anchor more securely in the left ventricle (see Figure 7.4). One

drawback of the 9 F assembly is that it is a sealed unit. If the encasing catheter or the

138

sensor coil is damaged from either usage or sterilization, the entire assembly must be

replaced.

The 6 F gradiometer assembly is smaller and more pliant. The maneuverability and ease
of positioning of the assembly is determined by the choice of the guide catheter, which
may be selected on a case-by-case basis. The 6 F assembly is suitable for use with all
heart sizes as the guiding catheter may be chosen to match the particular patient.
Although the gradiometer is more pliant than the 9 F assembly and consequently more
susceptible to effects of hemodynamics, the guiding catheter can serve to stiffen the
assembly, especially if a deﬂectable catheter such as the Cardima Naviport® is used.
The 6 F gradiometer is generally more durable under repeated use as the gradiometer is
protected by the guide catheter, which is discarded after each examination. As the
gradiometer is advanced into position only after the catheter is positioned, the sensor
head assembly is not subjected to the same physical manipulation as experienced by the
catheter during its insertion. Even so, the gradiometer is re-sterilized and thoroughly

inspected for damages after each examination.

The 6 F and 9 F gradiometer assemblies have been tested and characterized in a clinical-

scale in vitro test system and the results are presented in the next chapter.

139

CHAPTER 8. SYSTEM CHARACTERIZATION AND TEST

RESULTS

A clinical-scale prototype has been constructed to study the performance of the
gradiometer-based BSCC heart valve examination system. The system components are
described in Chapter 6 while details of the gradiometer are presented in Chapter 7. The
prototype is used to characterize the system performance as different parameters
associated with the system are varied. These parameters include the strength and
frequency of the excitation ﬁeld, placement of the gradiometer in relation to the valve’s
outlet strut, and the output signal level as a function of the excitation frequency. Based
on these results, an in Vitro test system has been constructed. Double-blind tests of 34
BSCC heart valves with various outlet strut conditions using the system has achieved
100% classiﬁcation accuracy. These results demonstrate the feasibility and reliability of

the gradiometer-based examination system in detecting BSCC outlet strut fractures.

8.1 The Helmholtz Excitation Coil System
The ﬁnite element (FE) analysis of the Helmholtz excitation coil system (Chapter 5)
predicts a uniform magnetic ﬁeld at the center of the coils with a cylindrical volume of
0.5 meters tall and a radius of 0.25 m as shown in Figure 8.1. To verify this result,
physical measurements of the magnetic ﬁeld have been taken between the excitation coils
with an input of a 6 A peak-to-peak, 1 kHz sinusoidal excitation signal. The measured
ﬁeld distribution, shown in Figure 8.2, conﬁrms the FE analysis results. In this ﬁgure,

the average ﬂux density is 1.28><10'7 T (1.28x10'3 gauss) with a variation of 0.74 %.

140

With an input signal of 10.62 A (peak-to-peak) and 62.77 kHz, the magnetic ﬂux density

at the center of the coils is approximately 1.5x10-4 T (1.5 gauss). This result agrees

reasonably well with the FE model prediction of 1.47x10-4 T (1.47 gauss) shown in

Figure 5.6 and Figure 5.7. Table 8.1 summaries the parameters and results for the FE
analysis and physical measurement. The physical measurement results agree favorably

with those from the FE analysis.

  

Excitation coil

  
    

Measuring range \

25 cmx50 cm

  

Figure 8.1 The Helmholtz excitation coil system has a uniform magnetic at its center
within a cylindrical volume of radius 0.25 m and 0.5 m in height

141

Magnetic Field Distribution At Center of Helmholtz Coil x 10:7
. 1.305

x10—7
A2.5,w
b l
a
C
CD .
D 1.5.
X
2
LL 1 .
.2 ’
he? 05
O) ' 7 ‘
(5 >1.
2 ol
4o .\
20 ,
o .
-20

Axial Distance, 2 (cm)

 

 

Radial Distance, X (cm)

Figure 8.2 Magnetic ﬁeld distribution at the center of the Helmholtz coil. The results
show a volume of uniform ﬁeld with approximately 0.74 % variation

Table 8.1 Summary of the ﬁnite element analysis and physical measurement parameters
and results for the magnetic ﬂux density at the center of the Helmholtz excitation coil

 

 

 

 

FE Analysis Results Physical Measurements
Excitation Frequency 62.5 kHz 62.77 kHz
Input Current (peak-to-peak) 11 A 10.62 A
Magnetic Flux Density 1.47x10'4 T 15 x104 T

 

Uniform Volume

= 0.5 m tall, 0.25 m

radius

2 0.5 m tall, 0.25 in

radius

 

 

Unifomiity

 

< 0.5 % variation

 

z 0.74 % variation

 

 

142

8.2 Lock-In Measurement of the Gradiometer Signal
The gradiometer signal is processed by a lock-in ampliﬁer (LIA)(Chapter 6) to increase
the signal-to—noise ratio. The LIA performs a phase-sensitive demodulation of the
gradiometer output with reference to the excitation signal. Equations (6.6), (6.7), (6.11),
and (6.12) describe the relations between the LIA outputs, X and Y, and the gradiometer
signal amplitude and phase angle, P and 6, respectively. The in-phase component, X,
corresponds to P cos(6) while the quadrature component, Y, is P sin(6). To remove the
phase-dependency of the gradiometer output, the magnitude, R, of the signal pair (X, Y)
may be used (Equation (6.8)). If the reference signal is rotated such that its phase
matches that of the gradiometer signal, 6 becomes zero. In this case, R = X = P and 6 =
Y = 0. By zeroing the phase angle prior to taking measurements, a smaller dynamic
range may be used for measuring the quadrature signal Y, effectively increasing the
system’s sensitivity to small changes in Y and the phase angle, 6. The outputs of the LIA
are digitized, recorded and analyzed to determine the condition of the outlet strut. The
signal pairs (R, 6) and (X, Y) are used interchangeably in the following sections when the

LIA outputs are referenced.

8.3 Clinical-Scale Prototype System
Figure 8.3 shows the clinical-scale prototype of the gradiometer-base examination
system. In this ﬁgure, the heart valve is mounted at the centre of the diagonal crossbeam
while the gradiometer is mounted at the end of the radial beam. The radial beam is
connected to a two-dimensional programmable motorized stage for positional control.

Figure 8.4 shows the test ﬁxture. In this setup, the heart valve is mounted in an

143

elastromer ring with a compliance comparable to the myocardium. The gradiometer is
encased in, and protrudes from, a guide catheter. The relative position between the
gradiometer and the outlet strut is preciously controlled by the motorized stage. The
clinical-scale prototype has been used for system veriﬁcation and heart valves

examinations. The test results are presented in the following sections.

 

Figure 8.3 Clinical-scale prototype for the gradiometer-based examination system

 

valve is mounted in an elastromer holder and the gradiometer is encased in a guide
catheter

8.4 Output Signal as a Function of Distance
Figure 8.5-Figure 8.8 show the X, Y, R, and 6 signals, respectively, from the LIA as a
function of the distance between the gradiometer and the outlet strut. In each ﬁgure, the
results for three cases are shown: a XSLS (explanted SLS), IOS, and no heart valve. The
ﬂange number1 of each valve specimen is provided in the ﬁgure legends. The
measurements were obtained with an excitation signal of 60.21 kHz, applied magnetic
ﬁeld of 1.304 gauss, and the 9 F gradiometer. These ﬁgures show that, if the gradiometer

is initially positioned on the outlet strut plane, the relative change in amplitude for the

 

l . . . .
The ﬂange number is an unique identiﬁer number edged onto the ﬂange of each BSCC
heart valve.

145

LIA outputs as the gradiometer is moved away from the strut is larger for a heart valve
with an IOS than one with a SLS. The distinction between the two signals diminishes if
the gradiometer was initially placed away from the outlet strut. Therefore, for optimal
sensitivity, the gradiometer should be placed within 5 mm from the outlet strut. This
result agrees with that obtained from the mathematical model presented in Section 4.10,

as well as the ﬁnite element analysis of the system presented in Section 5.2.2.

 

LIA ln-Phase Component Output (X)
20"_ T i ”" i r i i r T I_Ll
3 3 3 ; iii-3133.... :
. . l l l l -— IOS (A2517)
151 j .‘ , ‘ ‘, ' I .......... XSLS (A5023) _

 

l

. l l
l l l l l l l l

i | . l l l i l 7

, i l ' I I i I

I | I l l | l

I l l ‘. l I . l

Rcos(9) (uV)

X:

 

 

 

 

 

0' 2 4 ‘ s we 10 12 '1‘4” 16 18 20
Distance(mm)

Figure 8.5 In-phase component output, X, from the LIA as a function from the outlet strut
for the IOS, SLS, and no valve cases

146

LIA Quadrature Component Output (Y)

 

 

407 ———«— r‘ W i _ ___—L_-_:l
; . : : : : F-nNoValve ‘
35 ..... -.-—ini—Iosmzsm l
l f ‘ 1 , l l .......... LS(’_\5023),

,_ ___, 2-...7 l.-_,.‘2, ,Li_,_,uL _;_____l _L- L
i - I

 

 

Rsine) (W)

Y:

 

 

 

 

 

a
2ﬁ—T—8 8 10 12 14 16 18 2o
Distance(mm)

Figure 8.6 Quadrature output, Y, from the LIA as a function of the distance between the
gradiometer and the outlet strut for the IOS, SLS, and no valve cases

 

 

 

 

 

LIA Signal Amplitude, R
351 l 47 i r I I T I I
i 5 i ----- NoValve
30 _____ : _____ '_,__;____jg__—iOS(A2517)
I I ' i I . .......... BLS (A5023) .

I l l l I

 

 

 

 

 

 

 

o 2 4 6 8 10 12 14 16 18 20
Distance(mm)

Figure 8.7 Magnitude, R, of the gradiometer output as a function of distance from the
outlet strut for the IOS, SLS, and no valve cases

147

LIA Signal Phase Angle (0)

 

 

 

 

0.2 l- V ”(‘7’ i_ "T l l 7 l
' ' ‘ ' f 1 I l:--- No Valve
' l i — IOS (A2517) .
3 _ 1 i 7 ' I .......... )GLS (A5023) ‘
0.15‘ "“' —“' "“‘ ““““ "’“‘ Lg—‘ﬂ'Tm—T—m'j'w“

6(degree)

 

 

 

l

4105—- 7 1 4 1 ~ - :
o 4 8 81012141618 20

Distance (mm)

Figure 8.8 Phase angle, 6, output as a function of distance from the outlet strut for the
IOS, SLS, and no valve cases

8.4.1 Movement Model

During the heart valve examination, the gradiometer is encased in a guide catheter and
the assembly is inserted into the heart and positioned close to the outlet strut. The
position of the catheter is affected by the hemodynamics inside the heart, and is

continually oscillating about a mean position during the ventricular systole and diastole.

The movement of the gradiometer may be modelled by ﬁtting polynomials, f X (z) and
fy (z) , to the X and Y values in Figure 8.9 and Figure 8.10, respectively. The resulting

polynomials express the X and Y outputs as functions of the distance, 2, from the outlet

strut. For the data from Figure 8.9 and Figure 8.10, the resulting polynomials are:

148

fx (2) =0.0001z5 —0.0041z4 + 0.1031z3 -1.248z2 + 7.44792: — 145.8305 (8.1)

and

fy (z)=—0.0001z5 +0.0069z4 —0.1751z3 + 2.1706z2 — 13.58152 +79.1808 (8.2)

The output of the gradiometer can then be expressed as a function of time by assuming an
oscillatory gradiometer movement with a frequency of f = 1.2 Hz (corresponding to a
heart rate of 72 beats/min). The maximum swing of the gradiometer (i.e., the largest
amplitude of the oscillation), L, is a function of the turbulence in the ventricle and the

rigidity of the catheter assembly. The oscillatory motion can be modelled as
z(t) = :— L[sin(27z ft) + 1] mm (8.3)

In (8.3), the neutral, or resting, position of the gradiometer is assumed to be U2 m away
from the outlet strut. Substituting (8.3) into (8.1) and (8.2) and removing the mean value
from each result yields the X and Y outputs as a function of the gradiometer position over
time. Figure 8.9 and Figure 8.10 show the estimated relative changes in X and Y as
functions of time for the IOS, SLS and no valve cases with L = 10 mm. The same
procedure has been applied for R and 6 and the results are shown in Figure 8.11 and
Figure 8.12. These ﬁgures suggest that large X and Y (or, R and 6) signal variations over
time correspond to an IOS valve whereas small signal variations are related to a valve

with SLS,

149

Model-Generated Signal Based on “Output vs. Distance” Data

 

 

 

 

 

 

 

 

 

0' cl

0
O. '0
I

 

 

 

 

lllllllll

l
I
l
I
l
H
F"_'
l
l
l
T
l
I
I

-- >sLs (A5023) :-

----- No Valve

" XSLS (A5023) a

----- No Valve

 

--5-_--_l.-.-

 

— lOS (A2517) 1
— IOS (A2517)

 

 

 

 

 

 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Model-Generated Signal Based on "Output vs. Distance" Data

 

 

 

 

3s E 8920 952mm 95 >5 885 23mm

T1me(s)
Figure 8.9 Model-generated X signal with respect to the changing distance between the

gradiometer and outlet strut for the IOS, XSLS and no heart valve cases. L = 10 mm.

)
Figure 8.10 Estimated Y signal over time with respect to the changing distance between

the gradiometer and outlet strut for the IOS, XSLS and no heart valve cases. L = 10 mm.
150

Time (s

Model-Generated Signal Based on "Output vs. Distance" Data

 

 

     
 

 

 

 

 

 

 

 

 

20 r -—1 1 1
l ‘ ‘ l i 1 —IOS(A2517)
15, ,,__________-_._,, _____:-___ l """"""""""" )GLS(A5023);
:n 1 n ' n - ..... NOVBIW
9 ~ 1:» ————— ----- wallw — 4
S ' ' i i I i
.5 i ‘ I I l
C "in l |
U) 1
co 7 j. 1
.C 1 "1 l
O 1 1
2 1-- ----- 1
'5
L“
d)
(I.

 

 

 

 

 

 

Figure 8.11 Estimated R signal over time with respect to the changing distance between
the gradiometer and outlet strut for the IOS, XSLS and no heart valve cases. L = 10 mm.

Model-Generated Signal Based on "Output vs. Distance" Data

—— 103012517) ll
.......... )GLS (A5023) 91
----- No Valve

0.1

 

 

 

l
l
I

 

 

 

 

 

Relative Chagne ind (degree)

 

 

 

 

 

 

 

 

 

 

 

' o 0.5 1 7 1.5” 2 25 5 3.5 4" 4.5
T1me(s)

Figure 8.12 Estimated 6 signal over time with respect to the changing distance between
the gradiometer and outlet strut for the IOS, XSLS and no heart valve cases. L = 10 mm.

151

8.5 Gradiometer Output as a Function of Excitation Frequency
The gradiometer output at different excitation frequencies has been examined. Prior to
each test, the examination system was calibrated by submerging the gradiometer in the
uniform excitation ﬁeld but away from the heart valve. The phase angle, 6, was set to
zero to remove the phase difference between the reference and baseline gradiometer
signals. The gradiometer was then positioned on the strut loop plane and LIA outputs
were recorded as the gradiometer was moved away from the strut loop. Polynomials
were ﬁtted to the LIA data to form mathematical models according to the procedure
described in the previous section, and the X, Y, R, and 6 outputs were generated from the
models for an oscillatory motion with L=10 mm. Figure 8.13-Figure 8.16 show the
model-generated outputs, X, Y, R, and 6, at various excitation frequencies for an IOS
(A2517) and SLS (A5023) valve. The excitation frequencies were 41.15, 45.58, 51.57,

54.85, and 60.45 kHz and the data were collected using a 9 F gradiometer. The magnetic

flux density was 1.03><10-4 T (1.03 gauss).

8.5.1 In-Phase Signal, X

Figure 8.13 shows the in-phase component, X, of the gradiometer output. In this ﬁgure,
the signal amplitude for the IOS valve increases with increasing excitation frequency
while the signal amplitude for the SLS case remains relatively uniform across the test
frequencies. Table 8.2 lists the peak-to-peak X signal for both the IDS and SLS case at
the excitation frequencies. For the IOS case, the X signal at 60.45 kHz is 34.169 uV.
This is nearly twice the value of that at 41.15 kHz (17.873 uV). Meanwhile, the X signal

for the SLS case remains relatively small and does not exhibit any remarkable changes

152

over frequency. Table 8.2 also lists the ratios of the X signals between the IOS and SLS

cases at each excitation frequency. For example, at 41.15 kHz, the ratio between the IOS

and SLS signal is 6.6233, while at 60.45 kHz, the ratio is 10.2669. These results suggest

that a higher excitation frequency should be used, and that the LIA in-phase output, X,

may be used to estimate the condition of the outlet strut.

 

 

 

 

 

 

 

 

 

 

 

2llélodel-Generated Output vs Excitation Frequency (B = 1 .03x10'4 T)
F) 1 E ' ' 1 ——1—105 41.15kHz I
15 _ - 5-1 SLS 41.15kHz ‘1'
1 , , 1’1" -—-— — IOS 45.48 kHz
5? 1oz”- . .. - . . ““ SLS 4548"” -
i .1631 , 1 . ; - 1~.¥*“"5‘1 — ~ — IOS 51.57 kHz
; _g' 111. L ‘1. 1 '1 1 117+" ’51. SLS 51.57 kHz
.5 5'» l 1; l i -- - ‘l l:— ----- l —— — —Ios 54.85kHzJ
81 1 . ' ,1" SLS 54.85 kHz
g 1.: ' , . """ .. IOS 60.45kHz
5 . . '. I 1 1 "i SLS 8p.45 kHz,é
% 11‘ 11... E i
a; -10 ——.-1 -_ 1 ....... 11. ,1, . 1,.
i 1 i if “ 111' ll 1 l ' i,‘ I“ it Ill
“(l '‘1‘ 1 1 U.’ 1'
-201- L2 AL- _. 1 1. _
o 0.5 1 15 2 2.5 3 3.5 4 4.5

Figure 8.13 Estimated relative changes in X for the IOS and SLS cases at 41.15, 45.58,
51.57, 54.85, and 60.45 kHz. L = 10 m

Table 8.2 Comparison of the peak-to-peak X signal voltages between the IOS and SLS
cases at various excitation frequencies

 

 

 

 

 

 

 

 

 

 

Excitation Frequency IOS Case Peak-to- SLS Case Peak-to-Peak Ratio
(kHz) Peak Signal (uV) Signal (11V)
41.15 17.8728 2.6985 6.6233
45.58 19.9623 2.9964 6.6621
51.57 27.5486 2.7037 10.1893
54.85 28.3733 3.2161 8.8222
60.45 34.1692 3.3281 10.2669

 

153

 

8.5.2 Quadrature, Y, and Amplitude, R, Signals

Figure 8.14 and Figure 8.15 show the model—generated quadrature, Y, and amplitude, R,
signals and Table 8.4 and Table 8.5 summarize the peak-to-peak output voltages for these
two ﬁgures at the excitation frequencies listed in the previous section. From these
results, no deﬁnitive frequency-dependent behavior is observed for the IOS case.
Meanwhile, signal from the SL8 valve show only a moderate increase with increasing
frequency. The Y and R signals are, therefore, not strong indicators of the heart valve

outlet strut condition in this case.

Model-Generated Output vs Excitation Frequency (B = 1 .03x10‘4 T)

 

 

   
 
  
  

 

 

  

 

8 1 . -
l ' ——+——~ IOS 41 .15kHz
a SLS 41.15kHz -
—— — IOS 45.48 kHz
,\ SLS 45.48 kHz
> -
3. .--<--*|0851.57kHz
;: 1= SLS 51.57 kHz
_5 * IOS 54.85 kHz “
g, - M SLS 54.85 kHz
5, __ IOS 60.45 kHz 1
.c 1:11: -~ SLS 60.45 kHz
0 l— 1 ' , “—
a: - , l _f_
CE 1 1,],
' 111“ 1
. __ ._ _ H_- _ _ z_ ___.1 w ‘1___ L____ ______ _1
o o 5 1 1 5 2 2.5 3 3 5 4 4.5

Figure 8.14 Estimated relative changes in Y for the IOS and SLS cases at 41.15, 45.58,
51.57, 54.85, and 60.45 kHz. L = 10 mm

154

 

Table 8.3 Comparison of the peak-to-peak Y signal voltages between the IOS and SLS
cases at various excitation frequencies

 

 

 

 

 

 

 

Excitation Frequency IOS Case Peak-to- SLS Case Peak-to-Peak Ratio
(kHz) Peak Angle (°) Angle (°)
41.15 10.7875 1.6192 6.6622
45.58 1 12.2927 1.6184 7.5956
51.57 5.0313 3.2072 1.5688
54.85 12.4253 2.9070 4.2743
60.45 6.0914 4.4050 1.3828

 

 

 

 

Model-Generated Output vs Excitation Frequency (B = 1.03x10"1 T)
8 ' 1 1 7 r 1 r

 

 

 

 

 

 
 

 

 

' 3 i 1 i L ——1—— IOS 41.15kHz
5. , _ . ; . j MEWSLS 41.15kHz _
111121“ : 1.1.11.1x .5113? M’WL IOS 45.48 kHz
i 41 1‘1 1.1 ’ i ; Li" ‘ : '11," ' 7 Li ,, 11 ~ — nos 51.57 kHz
5; 1 L I L g ' L. L L SLS 51.57 kHz
5 2L 111‘} L L ——i——1‘;—~L:—----—~——-Ioss4.85k1-Iz~
81 L L 3 L L E Lgﬂ L ~ .._ SLS 54.85 kHz
5 0 —- — " 3'1, - - — - “H 108 6045 kHz -
5 ,1. 1.1.1“ 1 1... - 111.1 ' SLS 60-45 kHz“
{PM-18+? 1L 1" W I 1L” ' W 1. “L L: ' w.” .' " ' 1
9 '2 .‘VLL", 11"! TL L “““ ‘L “““““ :1 “““
Its“ ILL "' | ‘ 7L? : LL LIL ' 11. ’ Ll 1'
— "L L 1 L 11 ' “1 - L‘ LL
0’ 1 l 11 '_ 11
m '4'“ I, " ' '11 Y 'LsrL‘L-s ——*_rE1L'—'["6 -------- LIL—I‘LL- -----
LIL L‘,‘ 1' “LA! 1 LL 11' 1 I L "1' L ‘1-
3 [LL 1 I. 1 LL 1V . .
-6L— ----- ‘—‘ —————————— 7””1 T' LV‘ \L
i . L 3 . 7
.81. 1 .__ _ . __ . l 1 l _ _ l _...__%_
o 0.5 1 1.5 2 2.5 3 3.5 4 4.5
T1me(s)

Figure 8.15 Estimated relative changes in R for the IOS and SLS cases at 41.15, 45 .58,
51.57, 54.85, and 60.45 kHz. L = 10 m

Table 8.4 Comparison of the peak-to-peak R signal voltages between the IOS and SLS
cases at various excitation frequencies

 

 

 

 

 

 

Excitation IOS Case Peak-to- SLS Case Peak-to-Peak Ratio
Frequency (kHz) Peak Angle (°) Angle (°)

41.15 10.5265 1.6351 6.4379

45.58 11.3158 1.5341 7.3762

51 .57 11.1909 2.0135 5.5578

54.85 12.1431 2.5230 4.8130

60.45 10.7192 3.0259 3.5424

 

 

 

 

 

155

 

8.5.3 Output Signal Phase Angle, 0

Figure 8.16 shows the model-generated output phase angle, 6, for the IDS and SLS cases
at the test frequencies mentioned in the last sections. Table 8.5 summarizes the phase
angles and their ratios as the excitation frequency is increased. The results in this table
show that the phase angle increases with frequency. The ratio of the phase angles for the
IOS and SLS cases increases from 6.822 to 9.741 as the excitation frequency is increased
from 41.15 kHz to 60.45 kHz. These results suggest that the phase angle, 6, may be used

to indicate the condition of the outlet strut.

Model-Generated Output vs Excitation Frequency (B = 1 .03x‘10'4 T)

 

 

 

 

 

 

  

 

 

20TH”? ———7- M —————
; L . ; ' L L - 2~- IOS 41.15kHz L
15'. _ , L _____ -::- SLS 41.15kHz ..
L .9 E 1 1i 1 .I § _— — IOS 45.48 kHz
3 L "L L LL L L L ' , ~~ SLS 45 48 kHz
0: 104— ~~ - —* = :49“ - *vv ' A ------ r - - M ' ._ - - ' _
3, I . I L I . ILL - IOS 51.57 kHz
ﬂ L w‘ : L431 : I21. : L = -— SLS 51 .57 kHz
CD 5; - —~;=,+ 7— — — -— - 2; ~ *; ————— ~- IOS 54.85 kHz ~
.5 = 1’4; I In. 5 f : — SLS 54.35 kHz
5 L - , i "I : ~ ~ SL860.45kHzL
_c . I . I i | L I [a T v is} '
LI I I I i l ‘\ . ,1.
m m...» uI the” “w“ *1 Io . '
.> \ ,1/1 LI Il \ 4L .. I, Li I L L H L ll LL11 \\.\ 1 Li
g 10 , Ll 1 LI L L LL: '2 L LIL“ L L'
E - 1:” L? ‘ .' L 'Lh-N ,~;'-4- —-~+L:\L — 2‘. =L~ H
CE LL I I I’ L ‘ 1
'15 5““ r ‘— “' XL ’ L “ ‘1‘; ‘ lb‘\ 1 / ““““ 'r ‘ , "
I w . t L
-201 i -v‘I _ z_ _ i _ i _ . 7 L
o 05 1 15 2 a5 3 35 4 45

Figure 8.16 Estimated relative changes in 49 for the IOS and SLS cases at 41.15, 45.5 8,
51.57, 54.85, and 60.45 kHz. L = 10 mm

156

Table 8.5 Comparison of the peak-to-peak 6 value between the IOS and SLS cases at
various excitation frequencies

 

 

 

 

 

 

Excitation Frequency IOS Case Peak-to- SLS Case Peak—to-Peak Ratio
(kHz) Peak Angle (°) Angle (°)
41.15 12.7342 1.8665 6.8223
45.58 . 16.2826 2.1362 7.6222
51.57 21.3592 2.5015 8.5386
54.85 26.3412 2.6941 9.7775
60.45 32.3120 3.3171 9.7411

 

 

 

 

 

 

8.5.4 Outlet Strut Condition Indicator and Excitation Frequency Specification
The results from Figure 8.13, Figure 8.16, Table 8.2, and Table 8.5 suggest that
1. Either the X or 6 signal may be examined to provide an indication for the outlet
strut condition.
2. The excitation frequency should be set as high as possible within the limits of the
power ampliﬁer and the bandwidth of the impedance matching circuit.
The excitation frequency for the gradiometer-based examination system is approximately

62.5 kHz.

8.6 In Vitro Test System
As discussed earlier, the position of the gradiometer inside the heart is affected by the
hemodynamics and movement of the heart due to the cardiac cycle. In order to
investigate the effect of this movement on the performance of the gradiometer-based
examination approach, an in vitro test system has been constructed. The test system uses
the clinical-scale prototype in Figure 8.3 to simulate the gradiometer movement in
relation to the heart valve. A cinegram of an unidentiﬁed BSCC heart valve patient is
obtained [194] and the in situ movement of the heart valve is traced from a series of

ﬂuoroscopic images (Figure 8.17). The trace result describes the movement of the heart

157

valve relative to a stationary observation point. Figure 8.18 shows the heart valve travel
path. This travel path is programmed into motorized stages connected to the gradiometer.
The result is a motorized simulation of the relative movement between the gradiometer
and the outlet strut as shown in Figure 8.19. The motorized test system transverses the
travel path every 0.833 seconds (frequency = 1.2 Hz), corresponding to a heart rate of 72

beats per minutes.

 

50 1 00 1 50 200 250 300

Figure 8.17 In situ movement of a BSCC heart valve obtained from a series of
ﬂuoroscopic images

158

Relative Movement of the Heart Valve

 

 

 

 

l | l l T l
3f!———-——-“--——--—jV—~——---f ———————————————— j ————————— —
E2 ———————— -------- ——

5, 1 : : : l .
>1. ................. * : ..... J. .
0 . ‘ . :

0 2 4 6 8 1O 12

Figure 8.18 Movement of the heart valve in relation to a stationary observation point

50
1 00
1 50
200
250
300
350
400

450

 

1 00 200 300 400 500 600 700

Figure 8.19 Motorized stages programmed to simulate the relative movement between the
gradiometer and the outlet strut

159

8.6.1 In Vitro Test System Outputs and Classifications

Figure 8.20-Figure 8.23 show the LIA outputs, X, Y, R and 6, respectively, for an
explanted BSCC heart valve with SLS (XSLS, A5023). Figures 8.24 through 8.27 show
the same set of signals for an IOS valve (87384). The top graph in each ﬁgure shows the
raw signal and the bottom graph shows the results after the signal is ﬁltered by a 5th-
order Butterworth low-pass ﬁlter. The ﬁlter cutoff frequency was 5 Hz and the sampling

rate was 250 samples/seconds. The system excitation frequency was 60.27 kHz,

magnetic ﬁeld was 1.173x10'4 T (1.17 gauss), and a 9 F gradiometer (POT-9075) was

used to measure the response. Figure 8.28 compares the X and Y signals for the XSLS
and IOS cases, respectively, and Figure 8.29 shows the comparison of the R and 6 signals
for the two valve cases. These ﬁgures show that the amplitudes of the X, Y, R and, 6
signals from an IOS valve are larger than those from a SLS valve. These ﬁgures also
show that the phase angle of the signal provides the best distinction between the two
outlet strut conditions. In this case, a simple thresholding of the signal is sufﬁcient to

classify the condition of the two outlet struts.

8.7 In Vitro Study Results

. . . 1 .
In vztro double-blind studies were performed on BSCC 60° 29 mm heart valves With
various outlet strut conditions using the gradiometer-based examination system

prototype. The test samples included valves with IOS as well as SLS cases that were

 

1 The three largest size (29, 31 and 33 mm) BSCC heart valves all had the same ﬂange
ring (29 mm) and disc sizes. The different sizing referred to the outside diameter of the
fabric sewing rings.

160

either manufactured1 (MSLS) or naturally occurring (explanted, XSLS). During each
study, heart valves were selected randomly from a pool of samples and tested in the
gradiometer-based system. The output signals from each valve were interpreted by two
operators, one experienced and the other with minimal training. Each operator
independently assessed the outlet strut condition of the valve based on the output signals.
The signals were also analyzed using a computer software. At the conclusion of the test,
the decisions from the operations and the computer system were compared against the
true conditions of the valves. The following sections describe the test system parameters,

gradiometer selection, and the test results.

 

 

 

 

 

 

 

 

 

XSLS valve (A5023)
0.04 w . T e 44
9 . I i .
35 0.02 ----- , —————— , ———————————— 4
—\
so . .
E 0 ‘ ‘ ‘l ‘ ‘ - - 'I- . --
l . . I I
{Sr-0.02 ------------- — q ————————
l l l 1 l |
>< 0,04 __A on __ no: : l : ___
0 1 2 3 4 5 6
0.04 o
S .
S 0_02_____..___(__.,_
‘8 I
*0 ,
q. 0.02
I . .
>< 4,04% - 0;- __ _'._ 1 _a“ . _ﬁ
0 1 2 3 4 5 6
Tlme(s)

Figure 8.20 Output signal X for a BSCC heart valve with SLS (A5023)

 

1 A manufactured SLS (MSLS) was typically created by using either a laser or diamond
cutter to sever one of outlet strut legs

161

Y=P*sin(0 ) (,1 V)

P*sin(6 ) (g V)

Y:

H (#V)

H (ﬂV)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I l |
0 P _ _ _ _ _ _ 1_ l r _ _ _______
l I I
4102 ----- , ----- 1 ------- 1 ----- .L ------- : ---------
.004 __CW -l 1 i l l . _
0 1 2 3 4 5 6
Low-Pass Filtered Signal (fC = 5 Hz)
0.04, T 1

 

Time (s)
Figure 8.22 Output signal R for a BSCC heart valve with SLS (A5023)

162

 

 

XSLS valve (A5023)

 

 

0.4; ___W ___—4.? , , #2, j

a (degree)

 

 

6 (degree)

 

 

 

 

 

 

0.1-v‘w —— , 7
S ' 1
3; 0.05; ________________________________
8
$005f “““ i ————————————————————— It ___________ i
X -O.1'- ;— ‘L _i _ j

 

 

 

S
3
—\
8i
.8 :
Q l
>|é ‘ E
01- - 4— 4 .- :
1 2 3 4 5 6
Time (s)

Figure 8.24 Output signal X for a BSCC heart valve with an IOS (87384)

163

Y=P*sin(0 ) ()1 V)

P*sin(0 ) (ﬂ V)

Y:

 

 

nos valve (87384)

I _n__1___n .
1 2 3 4 5 6

 

 

 

i 1 .l
1 2 3 4 5 6

Time (s)

Figure 8.25 Output signal Y for a BSCC heart valve with an IOS (87384)

F? (#V)

R (#V)

0.1 mm.“

0.051e-44 ~

-0.05 —

-0.1 -—-—

0.1 —

0.051-”-- n -

0 -2

0.1 1
0

 

IOS valve (87384)

 

 

 

 

 

 

 

. s T -f 2

__ a 1 H _l_ __ 1

1 2 3 4 5 6

Low-Pass Filtered Signal (fC = 5 Hz)
T i l "T-

'- .:, - 1 ' 1
-------- ‘—-—-——-— --~-— :L _ 1|
’7' “if 3 ‘_ "'4 5 3

Figure 8.26 Output signal R for a BSCC heart valve with an IOS (87384)

164

 

IOS valve (87384)

2;— —~» ,

 

a (degree)
' o

 

 

 

 

 

 

1 r “i - .L- - -
.2 ‘Mi _ i __
0 1 2 3 4 5 6
Low-Pass Filtered Signal (fC = 5 Hz)
2 l T l _
a 1i. -- i ...... I _______________ J
8 = 1
<5 05 - -- ' - __ 'l- 1
19. , :
q; -1»- - — — — L ------
-2 i 1 2L _J
0 1 2 3 4 5 6

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.15' if i T l l -. I
9 ; . 1 , 87384008) !
‘3‘ 0"" ' " 'i """""" ‘ ---------- A5023()6LS)fi
a 1 1 -— T :
E 0.05»----

*0
0IT o
><

-o.05

0-4 ﬂ“ “7 “n l 1 1
S‘ l . : : 87384(IOS)1
5* 0-2‘"" ‘ 1‘ - L—----A5023(x5LS)j1
‘\ l I l l l r '
3i. :_ J -

.S. 4
to) f
E}? 0.
>.

0

 

Figure 8.28 Comparison of the LIA X and Y signals for the IOS (solid line) and XSLS
(dashed line) cases.

165

Low-Pass Filtered Signal (fC = 5 Hz)

 

 

 

 

 

 

 

 

 

 

 

0% em, 1 1 1
l i i : 87384006)
005* “““ J'“ _ ""i' 7 ----- A5023(XSLS)1
§ ’- ' 1‘
I Q
3‘ ° 1" is; ‘ ‘1“ _ ‘ "i ““““
0: : . :
-005 -— J - - ------ ; ------ IL ------- 4
-0.1—— in i L i
0 1 2 3 4 5 6
T1me(s)
27"" 1 r :i' i” T “#—7
‘ 1 1 l 87384006) I
l

l

 

15 - 5 - “1 - --------- A5023(XSLS)

 

a (degree)
_.:7: <3
i
g \

 

 

Time (s)

Figure 8.29 Comparison of the LIA R and 0 signals for an IOS (solid line) and XSLS
(dashed line) cases.
8.7.1 In Vitro Study: Trial A, 19 BSCC Heart Valves, 9 F Gradiometer
In a double-blind in Vitro study with 2 operators, 19 BSCC heart valves were tested in the
gradiometer-based examination system with a 9 F (POT-9075) gradiometer. The test

samples included 7 IOS valves, 6 with XSLS and 6 with MSLS. The excitation

frequency was 60.27 kHz and the applied magnetic ﬁeld was 1.1726x10.4 T (1.173

gauss). Figure 8.30-Figure 8.33 show the phase angle, 6, signals from the tests. The
ﬂange number for the test samples are provided in the legends of the respective ﬁgures.
The computer classiﬁcation results and the true condition of the heart valves are listed in
the format: (computer classiﬁcation result / true valve condition). For each valve, the
computer software processed the LIA output signals by low-pass ﬁltering and removing

their DC offsets. The software then applied a predetermined threshold of 0.85 to the

166

 

processed signal. (The selection of the software decision threshold is discussed later in
this chapter.) An outlet strut with a signal amplitude above the threshold was classiﬁed
as intact. Otherwise, the strut was classiﬁed as fractured. Table 8.6 summarizes the test
results. The indications from the operators and the computer software all achieved 100%
accuracy in determining the outlet strut condition of the BSCC heart valves. Note that
two valves, ﬂange numbers 73234 and 83740, were originally labeled as IOS valves on
their shipping containers. Test results from the gradiometer-based examination system

indicated a SLS condition for both valves. Visual inspections veriﬁed a MSLS on each of

these valves.

Heart Valve Tests (9 F Gradiometer POT-9075) 6

 

 

 

1' —“"”i l l i l 1 j "_1
l ' 1 ' I : :1 ---------- 87161lOS/IOS .,
1.51 —----11——A5023SLS/XSLS11
,, : ...: ; .1 I ; ., :1 ----- 85083 SLS/XSLS *1

11 :31 . f. . If; .
1r‘$:__ _1___.-'__': ____-1-:._.':__:____.152_::2_.1 - ''''' 943828LS/XSLS j
‘: ;1 - 732346LS/MSLs11
o. ' . :1 : . L 7;: '. I : E J:

6(degrees)

 

 

Time (s)
Figure 8.30 Output 6 signals from the BSCC heart valve study (Trial A), part I

167

 

 

 

 

 

Heart Valve Tests (9 F Gradiometer,POT—9075) 6

 

 

 

 

 

 

 

 

 

 

 

 

 

306.33% 600533

I
I
5 p
_ﬂllllull .I..I.J.IJIIII1III.I. .I I l'lalllll , all“ le'lﬂll]...
.. 2." _ .. 4 m 1 _
V .0
..("III 0000-001. 2 S _
S S w m _ ......I. _ 2 w E S L ...2. .......
8 £51.... ...2. _ 2 ... a 321%.:
l lcl 1.lIl llll_l llrllll . . .- 2IIL llllllllllllllllllll
I l I 2 \o. _ 4 n 0 D M M w I ... at.
S S .2 T I I S ......
D D w w 0 Cocooouooo -W 0‘.“ r u ( \PW / S S S L
S S \lsL“o\ u. 00.00-0-0040000 y 7 m L L L S
8 7 w 1.. ._ ....... _ 1 ...s. ....... 4 d 0 1 S S S .....
1 .1," 0000010. U 3 1 5 ﬂ
m m 21-..- .2. _ .5. 6.. m m 2%
A 7 7 v. o a 00 00-. [law H e T 2 3 1 voca-
u _u 2 c V _ 5 .W. m 7744.15.18:
IIIIIIIIIIIIIIIII 2111112111-.. In - ill 0.0. I
n m . 0‘00. .000. . onI\ _ 3 a I - - m.“ 00000
_ - ml Iolrzootoou v r - I .
I I‘f“l‘ _ 0'0... 2 e -
_ - 1‘ . u 0L _ t — a
. 1111...... ........ 2 m e _ .
1211-11111111011111 ...... ...... -113 e m 1 .1 1
2 CID-EOIOOII. . _ I _ _ IIIIIIII
_ 0‘. IIIIII 4. IIIII _ ‘ J, _ ) h om ”‘00 _ ‘ _ 2
2 ...-... _ 2 ~ S C d ccccc fun!— I 2
_ :J \S\ _ 5 “v‘ C m ..I ..... r so _ 2 5
,2 1 1 r I 1 1 1 508390.» 1111111111111 12 . S 11111 2 1111111111 12 1 1 1 1.1 131......35 1 1 12 11111111111
_ 2 _ 21...... o. o. 2 2 _ 2 :00. 2 2
_ _ \I-\l|\_\..ll. 2r 06‘ 2 m B m _ 2 _ ”I _ love... 2
2 _ I)! _ oo o s 2 can 0.. 2
_ 2 12")!” , I... Qlwo m 9 n H H 0000000 I ..... _I 2 H
L f1_1.u..1.:..t:.u.u1hvv 11 111111 r11 2 ( r11--. 11111 1.3-0.1111! ”lle 11111111111 2 11111 2
000000 ‘ _ ‘ .7 . . 2 M _ . _
.00... _ 2V _ m $ oooooooo _ ‘o _
000000 1o . ‘1‘ 2 _ o S 'J-DOOOIOI ho _
_ .1 in“... 2 _ 2 tr“ e 2 mo’ooool ooooo 2
.11.Illliw+uu\u1H-I l . non-Wunooowr5’012 1115 T II 111111111111121119 lIIlIHIl-H-ouu 1111111111 5
U . “I", _ _rlo ~ 1 m m h H u ' ’/ “GUI-I‘D _ 2 _ 1|
2 1.35.... It... :2 2 2 2 1 001:... . . .
I}! n I...
_ IoOoOoIIOoLOoouo‘oo I. _ 2 g a _ 00 ccccccc .00. I _ _ 2
...........-..;- _ . 1 1.,2- - 1.1 ...... v - 14.1.3... ..... \v _1 :_. 1 .1
sane-on- ¢I\\. 6 n 2 2 :40. oooooo . a o\ _ 2 2
_ . ti...“ _ . a 2 _ 2 .1 00000000 rut. . 2
U _ 2“- I 004-003000000030 I. m e . 2 2 ﬂ . not!“ 2 2.

; 121 Audi: -113... 11111 5 D. H 22--.-” .......... .21--..o/111-»1.1w. 1111111 211-15.
_ﬁ 2 1:71)]... \r coco-.12 O m 2 2 _ on... .n 2 2 0
_ snoop-ooogooouo” oooooo 00‘ ...-11., n O n cccccccccc e..- I . H H H

o- uuuuu . .142. _ 1 _ 2‘ 00 0| . .0 _ 2 2
i ...-.1. 12 K. 11 2 *1) 1 O 3 .1 .111.- .. . e .111 2 2 1'11 0

1 5. 0 5 ..1 5 8 5 1 5 0 5 ..1 5
0 0 ..1 e 1 0 0 ..I

r

u

g

..I.

F

Tune (3)
168

Figure 8.32 Output 19 signals from the BSCC heart valve study (Trial A), part III

Heart Valve Tests (9 F Gradiometer,POT-9075) 19
[222 A . l f l l ‘ 2 I l I __
V ' -"- t .13": ' !"-. i ---------- 87384 IOS/IOS l
‘ ’ 83740 SLS/MSLS 1
I ————— 110830 IOS/IOS
. - ------ A4567 SLS/XSLS

 

 

 

 

 

1
"l

 

"51:
I1

0 (degrees)

 

 

 

 

Figure 8.33 Output 6 signals from the BSCC heart valve study (Trial A), part IV

Table 8.6 Summary of In Vitro Double-Blinded Test Results of 19 BSCC Heart Valves

with a 9 F gradiometer (Trial A)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8:13:18 Serial No Flange Operator 1 Operator 2 Computer Known
No Indication Indication Indication Condition

1 29MBRC14924 87161 IOS IOS IOS IOS
2 31MBRC61089 A5023 SLS SLS SLS XSLS
3 3 1MBRC 10029 85083 SLS SLS SLS XSLS
4 3 1MBRC 10177 94382 SLS SLS SLS XSLS
5 29ABC 10937 73234 SLS SLS SLS MSLS*
6 29MBC60029 A6628 IOS IOS IOS IOS
7 29ABC1 1608 83986 IOS IOS IOS IOS
8 3 1 MB RC601 12 A2517 IOS IOS IOS IOS
9 3 1 MB RC 1. 2926 76225 SLS SLS SLS XSLS
10 29ABC 10925 73159 SLS SLS SLS MSLS
l l 29ABC40035 72393 IOS IOS IOS IOS
12 31ABC10467 73021 SLS SLS SLS MSLS
13 31MBRC12282 73225 SLS SLS SLS MSLS
14 29M BRC 10776 61046 SLS SLS SLS XSLS
15 29MBRC 16797 102268 SLS SLS SLS MSLS
16 29MBRC151 10 87384 IOS IOS IOS IOS
17 29ABC1 1607 83740 SLS SLS SLS MSLS*
18 29ABC 12556 1 10830 IOS IOS IOS IOS
l9 3 1MBRC60939 A4567 SLS SLS SLS XSLS

 

 

* - MSLS valves originally mislabeled as IOS valves on their shipping containers

169

 

8.7.2 In Vitro Study: Trial B, 19 BSCC Heart Valves, 9 F Gradiometer
In a second double-blind in vitro study with 2 operators, the 19 BSCC heart valves from
Trial A were tested again in the gradiometer-based examination system with a 9 F (POT-

9075) gradiometer. The excitation frequency was 60.27 kHz and the applied magnetic

ﬁeld was 1.123x10.4 T (1.12 gauss). Figure 8.34-Figure 8.37 show the 6 output signals

from the tests. The software classiﬁcation threshold was again set at 0.85. Table 8.7
summarizes the test results. As before, the indications from the operators and the

computer software all achieved 100% accuracy in determining the outlet strut condition

of the BSCC heart valves.

 

Heart Valve Tests (9 F Gradiometer,POT—9075) 19

 

1.5__ _____ I _____ ............... 73021SLS/MSLS
' ‘ I A6628|OSIIOS

' ----- 110830lOS/IOS

‘”-.- ‘ - ------ 87384lOS/IOS

 

 
    

     
 

 
 

 

 

‘35.; .
.

 

bl“-..
l

    

1 "Iﬁ
D

19(degrees)

—————————————————————————

 

Figure 8.34 Output 0 signals from the BSCC heart valve study (Trial B), part I

170

Heart Valve Tests (9 F Gradiometer,POT-9075) 6

" 87161 IOS/IOS

 

 

‘1

A2517 IOS/IOS

—---- 85083 SLS/)GLS
------- A5023 SLS/)6

72393 IOS/IOS

 

 

36656325

  

 

Figure 8.35 Output 19 signals from the BSCC heart valve study (Trial B), part 11

Heart Valve Tests (9 F Gradiometer,POT-9075) 19

-- 61046 SLS/)GLS

 

 

1.51

mm .18.
SSS
mmmm.
mama
SSIS
7mm
_-
2..
.. .
__

 

 

 

 

 

36666an

Time (s)
Figure 8.36 Output 6 signals from the BSCC heart valve study (Trial B), part III

171

Heart Valve Tests (9 F Gradiometer,POT-9075) 6

 

 

 

. 7;.— ___,7,,,r ,, ,
“___——

1.59 ._:___-,w ---------- A4567SLS/XSLs-i
‘ 1 ‘ 83740 SLS/MSLS ‘ ’
l l . , } --—-— 73234 SLS/MSLS{
‘ ‘ ' *‘- “““““ - ------ 76225 SLS/XSLS 1

I

 

 

 

6 (degrees)

 

Time (s)
Figure 8.37 Output 6 signals from the BSCC heart valve study (Trial B), part IV

Table 8.7 Summary of In Vitro Double-Blinded Test Results of 19 BSCC Heart Valves
with the 9 F gradiometer (Trial B)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Serial No Flange Operator 1 Operator 2 Computer Known
No No Indication Indication Indication Condition
1 3 1 ABC 10467 73021 SLS SLS SLS MSLS

2 29MBC60029 A6628 IOS IOS IOS IOS

3 29ABC 12556 1 10830 IOS IOS IOS IOS

4 29MBRC15110 87384 108 IOS IOS IOS

5 29MB RC 16797 102268 SLS SLS SLS MSLS
6 29MBRC 14924 87161 IOS IOS 108 108

7 31MBRC601 12 A2517 IOS IOS IOS IOS

8 3 1MBRC 10029 85083 SLS SLS SLS XSLS
9 31MBRC61089 A5023 SLS SLS SLS XSLS
10 29ABC40035 72393 IOS IOS IOS IOS

1 1 29MBRC10776 61046 SLS SLS SLS XSLS
12 31MBRC12282 73225 SLS SLS SLS MSLS
13 29ABC 10925 73159 SLS SLS SLS MSLS
l4 29ABC1 1608 83986 IOS 108 108 IOS
15 31MBRC10177 94382 SLS SLS SLS XSLS
16 31MBRC60939 A4567 SLS SLS SLS XSLS
17 29ABC1 1607 83740 SLS SLS SLS MSLS*
18 29ABC10937 73234 SLS SLS SLS MSLS*
19 3 1MBRC 12926 76225 SLS SLS SLS XSLS

 

*-Valves with MSLS mislabeled as IOS valves on shipping containers

172

 

8.7.3 In Vitro Study: Trial C, 34 BSCC Heart Valves, 6 F Gradiometer
A double-blind in vitro study was conducted on 34 BSCC heart valves (10 XSLS, 10
MSLS, and 14 IOS cases). Additionally, the control case with no heart valve was also

examined. In this study, a 6 F gradiometer (POT-9085, #2) was used. The excitation

frequency was 63.32 kHz and the applied magnetic ﬁeld was 1.16l><10.4 T (1.161 gauss).

The software decision threshold was at 6.0810. (The selection of this threshold is

discussed in the following section.) All other test conditions were similar to those in the
previous trials. Figure 8.38-Figure 8.44 show the output 6 signals from the study. For
each valve, the ﬂange number, computer decision, and the true valve condition are
displayed in the respective graph legend. Table 8.8 summarizes the study results. The
assessments from the operators and the computer software all achieved 100% accuracy in
determining the outlet strut condition of the 34 BSCC heart valves. Because of the low
signal output level, the “no valve” control case was classiﬁed as a SLS condition.

Heart Valve Tests (6 F Gradiometer,POT-9085): 6

 

‘1
l

15_.. I H __; ---------- 72316IOS/IOS lg
‘ ‘ ' ‘ ’ —1108301081108

7 _ , _ v i, -__—1 0‘? __ ___.

 

. ----- 72243 IOS/IOS
I ' - ------ 72315 IOS/IOS 1
' 71981 IOS/IOS ’

 

10'

 

     

O
r

  

6(degrees)

 

 

Figure 8.38 Output 6 signals from the BSCC heart valve study (Trial C), part I

. 173

 

 

 

 

 

 

 

 

 

   

 

 

 

 

Heart Valve Tests (6 F Gradiometer,POT-QOBS): 6

 

 

 

 

H
m.
. ‘\ 5 \I’ III. Iv 1431111111111: 1111
Sw ...... u H _ C 8883 _ 7 n u
L _ canon- _ . . .
ssssm A» ... ....... ......_ ... .... ammas ,. _ _
-mmmm . ..... _ ...... .....-......4 .n a-m11mm:_r: . ........ -
swam-......z. _ a ) wmwws_ OF! _
DDSSS " H Y 5 $8880 _ . _ n H
n mtg-...... .. _ .m m-m w ......... _ ...... _ ..........
mmamw _ .. a e 7mmmy _ «V n “
77A " e T &9 78 _ -. _ u .
u u m .m m -u u “1..-? ........
_ u m . .... v r. u m . n _
_ . m w _ . - _ u .
- 1 6 11I: _ _ _
e m 111411 111 IIIIVIJ llllll _I|-.-4ll
s C .m _ _ . .
1\ C _ -
e S m Illrl IIT lllll rill! , lllllllll
m B G _ _ . n
T e F _ _ _ . .
.m 1611?}:- ..... -, .... ...... H ..... .... -
m _ u . _ _
m m _ _ oil!
f _
s _I ......... _- ----- _ ..... _ ...... “-1- l--
.. m _ ‘v _ u _
g a r _ _ a
.9 W - - ..... .1! a - _ --.“ ..... ...- -
u _ . . _
m. H - 1e. ...... h -------
u _ _ . . .
O m u . . . H
mu 7. - M ...-cl . _ l0
8. ..m m 5 0 6 m w
m g .
360.83% cue 36656an
H

. 3.5 4.5
Time (s)

Figure 8.40 Outpth signals from the BSCC heart valve study (Trial C), part III
174

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Heart Valve Tests (6 F Gradiometer,POT—9085): 6

 

 

 

 

 

 

 

 

V
I
- 1} .11 1, P .1 1-
1 «.1 .. 13... Q 111-$114 ----- .-.-1.11- .5
ww as a wssmm \JI‘ ( _ _ .
.mmmwp .n a,mmmwm.-1.-:._ ..... _ ..... .. 1.“.
SSWLW a \11 $03 mow 11111111 . _
LL SD 5 DS _ _ III. .
SSnlV v1 8 S7 38 _ ... _ .1} . m
1 3 8 d O 1 075 l. ..... hhlill..r\-\..u--_-1--.-.-4
7 nkHA e T H7H7r” . n " 111V
. _ . V O In . u r-J-INII 1 . _ 5
m _ m a P- . _ . _ ., --...-H ..... _:-::_:3.
m u m . v m u _ . . H 1...». H
m _ m mm .mm _ . _ nit-.1111... .
F a 111 . _ e m IttallllhTIllL-M‘It‘ltl .1 _ lllll J. lllll 4_I113
" U Q N u ) .n m _ A," " ”U
M H _ ._ H n 16 m w . _ 11.1- _ _
fl » pl-tIITll! . llllll .ll- :15 II...1._II _ ll.ll.ltrultii_l .-ln [[[[[ “I115
_ H u 4" H u 2m m. a H A. YMMV” 2
. H n . _ _u e F . . - . ------- _ .
- 2 1 1 ‘1--. ------ ------ .---2 .m. 16‘ . .....- ., .
n . H A m __ m m _ .
----- .- ----- _-1-4.' 5 ....u _mw . ..
_ ﬂ . _ _ 1 S a
_ . _ _ n H a m H .
. . n a _
....... 1‘ _ m. v -
a . .- . . -------- 41 s n
N _ I _ 6 a .
. H .‘v 5 m ”m _ ..
..... $1.111. -o m.- . -
. _ u .
.ﬁ “ h . 0 .
5 01 5 . 0 ::w 0 5 0 A” - .. --
1 1 . ..- ..1 8
e
AmmeomEm m Emma—weave
g
H

Time (s)
175

Figure 8.42 Output 6 signals from the BSCC heart valve study (Trial C), part V

6(degrees)

 

Heart Valve Tests (6 F Gradiometer,POT—9085): 6

 

 

 

 

151“ V --------- 71903|osnos _1
. 1 ' ‘ 1 1 1; 73130 SLS/MSLS :!

5 1 ‘ '1 ----- 87161IOS/IOS -

1° - - ------ 7622SSLS/XSLS 1'

 

Figure 8.43 Output 6 signals from the BSCC heart valve study (Trial C), part VI

6 (degrees)

Heart Valve Tests (6 F Gradiometer,POT—9085): 6

 

 

 

 

 

151 .-.- 1-.- ‘- --.1 --- -2- .......... 84093SLSDGLS l“
1 ' i : 1 i 1 102268 SLS/MSLS ‘

g - . i 1 1 ————— A4567 SLS/XSLS H
1%; - ------ 7315QSLS/MSLS I1
; ; 1 - L__' ~A0NV/NVM 1

. __—r_.___

 

 

l - 1__.L- 1._ -.___!.__-'- - 1 1.__ .1
o 05 1 15 2 2.5 3 35 4 4.5 5
T1me(s)

Figure 8.44 Output 6 signals from the BSCC heart valve study (Trial C), part VII

176

Table 8.8 Summary of In Vitro Double-Blinded Test Results of 34 BSCC Heart Valves
with the 6 F gradiometer (Trial C)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Serial No Flange Operatorl Operator2 Computer Known

No No Indication Indication Indication Condition

1 29MBRC 12432 72316 IOS IOS IOS IOS

2 29ABC12556 1 10830 IOS IOS IOS IOS

3 29MBRC 12423 72243 IOS IOS IOS IOS

4 29MBRC12431 72315 IOS IOS IOS IOS

5 29MBRC12407 71981 IOS IOS IOS IOS
6 31MBRC601 12 A2517 IOS IOS IOS IOS

7 29ABC1 1608 83986 IOS IOS IOS IOS
8 29ABC 10934 73230 SLS SLS SLS MSLS
9 3 lABC 10467 73021 SLS SLS SLS SLS
10 31MBRC61089 A5023 SLS SLS SLS XSLS
1 1 29ABC1 1607 83740 SLS SLS SLS MSLS*
12 3 lABCB 10022 93894 SLS SLS SLS XSLS
l3 3 1MBRC 10177 94382 SLS SLS SLS XSLS
14 29MBRC 12477 72330 SLS SLS SLS MSLS
15 29MB RC 15 1 10 87384 IOS IOS IOS IOS
l6 29ABC10937 73234 SLS SLS SLS MSLS *
17 29MB RC 1 5994 90063 SLS SLS SLS XSLS
18 29ABC40035 72393 IOS IOS IOS IOS
19 29EMBRC 10332 1 17432 SLS SLS SLS MSLS
20 29MBC60029 A6628 IOS IOS IOS IOS
21 31MBRC10029 85083 SLS SLS SLS XSLS
22 31MBRC16369 1 13227 IOS IOS IOS IOS
23 29MBRC 12430 72310 IOS IOS IOS IOS
24 S3 1MBRC 10032 115573 SLS SLS SLS MSLS
25 31MBRC12282 73225 SLS SLS SLS MSLS
26 29MBRC 12406 71903 IOS IOS IOS IOS
27 29ABC 10932 73180 SLS SLS SLS MSLS
28 29MB RC 14924 87161 IOS IOS IOS IOS
29 31MBRC12926 76225 SLS SLS SLS XSLS
30 29MBRC61636 A4865 SLS SLS SLS XSLS
31 31MBRCB 10006 84093 SLS SLS SLS XSLS
32 29MB RC 1 6797 102268 SLS SLS SLS MSLS
33 31MBRC60939 A4567 SLS SLS SLS XSLS
34 29ABC 10925 73159 SLS SLS SLS MSLS

 

 

 

 

 

 

 

*-Valves with MSLS mislabeled as IOS valves on shipping containers

177

 

8.7.4 Selection of the Computer Decision Threshold

The initial decision thresholds for the computer inspection software package were
established experimentally during system development. A threshold was selected for
each of the two gradiometers from measurements using the system prototype. A more
rigorous approach for selecting the threshold involves examining the mean and standard
deviation of the peak-to-peak 6 variation for the valves with SLSs. Figure 8.45 shows the

output 6 variation for the test cases in Trial C. From these results (excluding the “no

valve” case), the mean of the 6 variation for all the SLS valves is [.t = 1.51510, and the

standard deviation is a: 1.1320. The decision threshold may be selected as

Threshold = ,u + no n =1, 2, 3... (8,4)

Thresholds for n = 3 to 6 are plotted in Figure 8.45. Comparison with the output 6 values

show that only the thresholds at 4.9110 (n=3) and 6.0420 (n=4) are able to separate the
two valve classes. Note that the threshold 6.0420 matches loosely the experimentally

established threshold 6.0810.

The data in Figure 8.45 show that the mean and standard deviation of 6 have been
inﬂuenced by two outlier SLS data points corresponding to Sample ID #12 and #17. If

these data points are excluded, the adjusted mean and standard deviation become 11

(adjusted) = 1.3970 and 0(adjusted) = 0.8150. Applying these values to (8.4) yields the

adjusted thresholds (n=3 to 6) in Figure 8.46. These results suggest that the threshold

178

7.175o (n=5) or 8.3070 (n=6) may be used to distinguish between the two outlet strut

Classes. The mean, standard deviation and thresholds for the cases when all the SLS data
are used and when the outliers (#12 and #17) are excluded are summarized in Table 8.9.

The gradiometer-based examination system software employs a decision threshold of

8.307001 + 60', adjusted) when used with the 6 F gradiometer, and 1.200 when used

with the 9 F gradiometer.

Table 8.9 Summary of the mean, standard deviation, and decision thresholds for the cases
when all SLS data are used and when the outlier data points are excluded.

 

 

 

 

 

 

 

All SLS cases Excluded Outliers
(°) #12&#17(°)

a 1.5151 1.3965

0' 1.132 0.81451
Threshold T3: 11 + 30' 4.911 3.84
Threshold T4: 11 + 40' 6.043 4.6545
Threshold T5: 6 + 50 7.175 5.4691
Threshold T6: [1 + 60' 8.307 6.2836

 

 

 

 

 

179

A9 (degrees)

Figure 8.45 Peak-to-peak 6 variation for the samples examined in Trial C. The mean,
standard deviation, and various thresholds for 6 are as indicated.

 

 

 

Peak-to-Peak Phase Angle (A0)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

30 4 I—_ - 'I‘T_T"-I I I I I I I T I I I I I I I I l I lfI T 1
1 -—--—p,(AllSLSs): 1.5151 1
1 -4— o (All SLSs): 1.132 -
- I—o—Thres (Applied): 6.0805 f
p ‘ “+60 (All SLSs): 8.307 11
25" - ------ 1141506113153): 7.175 1
g ————— “+40 (All SLSs): 6.043 2
1 ---------- “+30 (All SLSs): 4.9111 1
o
20 a
1 i 9
1
15 a. _-
’ 1 o
2 l ’ l i
10.: ' 4
o
1% ¢¢¢¢¢¢¢ e: c e c: : ¢¢e¢¢¢¢¢
5,1- : -
i ' ,
1 i - 1
1 1 I -
T323132] *- 1? i i ‘ “
‘1 1 1. ‘ - 1‘ T . T 1 i a i - ! .. ’1
1 2 3 4 5 6 7 8 91011121314151617181920212223242 26272829303132333435

Test Sample ID Number

180

Peak-to-Peak Phase Angle (A6)
30' ' ' 7 "'"HT— "—I——I—TTHI" ' I I I I I I I I I I I I I I I I I‘VTWI _T’AT—‘L f
’ —-— p, (adjusted): 1.3965
-4--a(hdnsuxn201h451
—0— Three (adjusted): 6.0%
“+60 (adjusted): 6.2836

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

' """ 111150 (adjusted): 5.4691 *
I ---- “+40 (adjusted). 4.6545
.......... “+30 (adjusted). 3.84
O .
. I
l
Cl
3 . e
95’. * ‘ ‘
a) 15” 3 l I
2 * 1
CD . 1
<1 0 . , O .
! Q Q l . l ! l
10'" Z l : l 4
o ' ‘
1 : l
I l
l i 1
5_ l I r
_”‘l—_‘#4_A—H-7_j ————————— ! ————— 1»—~-.——1——-—-———1--1 ————— -4———-l —————————————————
I I I T '
:“f l f: ? ::
V Y I H_‘ H H

 

 

 

 

 

 

0'. ‘ I 1 g ; i
1 2 3 4 5 6 7 8 91011121314151617181920212223242526272829303132333435
Test Sample ID Number

Figure 8.46 Peak-to-peak 0 variation for the samples examined in Trial C. The adjusted
mean, standard deviation, and thresholds for 6 are as indicated.

181

8.7.5 Trial D: Threshold Veriﬁcation, ll BSCC Heart Valves, 6 F Gradiometer
A double-blind study with 11 randomly selected BSCC heart valves of various outlet
strut conditions (2 XSLS, 4 MSLS, and 5 IOS) was conducted to verify the decision

threshold established in the previous section. In this study, the excitation frequency was

63.02 kHz, applied magnetic ﬁeld was 1.55x10'4 T (1.55 gauss), and a 6 F gradiometer

(POT—9085, #2) was used. All other test conditions were identical to those employed in
the previous trials. Figure 8.47-Figure 8.49 show the 6 signals from the study, and Table
8.10 summarizes the decisions from the operators and computer software. These results
demonstrate 100 % accuracy in indicating the valve conditions. Figure 8.50 shows the
peak-to-peak 6 values for the heart valves examined. The adjusted thresholds from Table

8.9 are also plotted in this ﬁgure for comparison. The results from this ﬁgure verify that

the 6-sigma threshold 6.28360 (the value employed by the gradiometer-based

examination system) may be used to classify the outlet strut condition of the BSCC heart

valves.

182

Heart Valve Tests (6 F Gradiometer,POT-9085): 6

 

SLS ,1
11

72330 SLS/MSLS

----- 71903 IOS/IOS
-'-°-'- 93894 SLS/XSLS

 

-- 102268 SLS/M
73159 SLS/MSLS

 

 

f! fill

 

 

4'."
(.
‘u’ld’l -
I fuels-Ill

T
|
1"" ‘

 

$35233

 

 

 

Time (8)

Heart Valve Tests (6 F Gradiometer,POT-9085): 6

 

_ ‘.‘

 

h it-.-

86

104

A251 7 IOS/IOS

, _ _
0‘
.. 1 - - detailsi- - -
r ‘

 

U‘r‘.‘ _

ILO‘I‘I

11-..:11._111

4.2 02

Ammmaouvm

  

111-11.1111111

_ _ . | I... _

. _ . o‘t“n‘ Ital. _ _ _

_ .I' I...‘ P _ . _ _

\o‘. I

_9'0'_ _ . \ltll _ 5
rlrllrwﬂ. .wn‘X‘“lll_lllLlilrll I. llllll 2

_ ‘0 J'u'u’u' r _ a

_ A _ _ _ a _ I l

_

,

'13
'l
‘f
3
!_
l
.i
I
I
f
i
‘—
1 5

 

 

Time (s)

Figure 8.48 Output 6 signals from the BSCC heart valve study (Trial D), part II

183

19 (degrees)

 

Heart Valve Tests (6 F Gradiometer,POT-9085): 6

F' I ' I 1 I l T r___ _I l I ' T1
101 — - '- 1, A5023 SLS/XSLS 1:7

81 t ----- ..----i .............
61- - . - - A :— eeeeeeeeeeeeee ',
4..- ...___~3__._LCA, 2L___4L .1} ....... -4:
1 1 1 1
2 -------------------------------- '1— « - 1 -------------- l
1 1 : : 1
0 }["°- rt ~W~~9M'."“— ----~u-*'°."- ~~~~~ 0" 3'1»......:...“.o2' ~ “MO." vr' ”~44
-2 ; ............................................. .I
l
«1‘» ———————————————— 1‘
1 1
'5‘ t I . *1
-8 .............. ‘1 .................... L ............. 4
l l l - - , l _ l _L.. __ ..- ---... l_ __ _I__ -
o 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Time (s)

 

 

 

Figure 8.49 Output 6 signals from the BSCC heart valve study (Trial D), part III

Table 8.10 Summary of In Vitro Double-Blinded Test Results of 11 BSCC Heart Valves

with the 6 F gradiometer (Trial D)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Serial No Flange Operatorl Operator2 Computer Known
No No Indication Indication Indication Condition
1 29MBRC16797 102268 SLS SLS SLS MSLS
2 29MBRC 12477 72330 SLS SLS SLS MSLS
3 29MBRC12406 71903 IOS IOS IOS IOS
4 31ABCB10022 93894 SLS SLS SLS XSLS
5 29ABC 10925 73159 SLS SLS SLS MSLS
6 29ABC 10934 73230 SLS SLS SLS MSLS
7 29MBRC12430 72310 IOS IOS IOS IOS
8 29MBRC14924 87161 IOS IOS IOS IOS
9 29MBRC 15 1 10 87384 IOS IOS IOS IOS
10 31MBRC601 12 A2517 IOS IOS IOS IOS
1 1 3 1MBRC61089 A5023 SLS SLS SLS XSLS

 

184

 

Peak-to-Peak Phase Angle (A6)

 

 

 

 

 

20: e — W
l
19: e .
’ 1
16: e
j T
141 ‘7 1 9 j
8 12:1 —o—Thres (Applied): eoaosf ; i
9 3 Threshold T6: 6.2836 1 1
8 10’ ‘ - ------ Threshold T5: 5.469 ' .
'3 1' ----- Threshold T4: 4.6545 * 3
CD 311 ---------- Threshold T3:3.84 J , = 1
<1 WW-

   
 

 

 

’71 2'79 4_"5 6 7 8 9 1o 11
Test Sample ID Number

Figure 8.50 Peak-to-peak 6 variation for the heart valves examined in Trial D. The
adjusted thresholds from Table 8.9 are included for comparison.
The results presented in this chapter demonstrate the proof of concept of the gradiometer-
based examination system in determining the outlet strut condition of a BSCC prosthetic
heart valve. In the following chapter, a probability of detection model is developed to
quantify the system performance when the gradiometer is placed away from the outlet

strut plane.

185

CHAPTER 9. PROBABILITY OF DETECTION MODEL

The reliability of the gradiometer-based prosthetic heart valve evaluation system is
inﬂuenced by several variables such as the alignment and distance of the gradiometer
with respect to the heart valve outlet strut. To quantify the system performance, a
probability of detection (POD) model that can predict the output signal in the presence of
the measurement variabilities may be used. A POD model for the gradiometer-based
system accounts for the uncertainties that affect the output signal by generating a
distribution of this signal around its mean value to predict the variations. The
development, analysis, and beneﬁts of applying POD models for nondestructive
evaluation techniques have been discussed extensively in [180]-[183]. This chapter
presents a POD model for the gradiometer-based examination system. In particular, a
POD model that takes variabilities in the gradiometer’s initial position from the heart
value outlet strut into account is presented. The initial position of the gradiometer is
deﬁned, in this context, as the closet distance of the gradiometer to the outlet strut during
the cyclical movement. The concept described herein may be adapted to develop POD

models accounting for other sources of variability in the examination system.

9.1 Sources of Variability
Figure 9.1 shows the optimal position of the gradiometer during a heart valve
examination. The primary source of variability for the examination system is the

deviation of the gradiometer from this optimal location. Figure 9.2 (a)—(c) show

186

examples of translational deviations and Figure 9.2 (d)-(e) show examples of alignment

angle variations of the gradiometer. Other examples of variability include:

0 Human factors (e. g., the skill set of the medical personnel);
0 Variations in patient physiology such as the heart rate and cardiac output volume;

0 Variations in the excitation source such as the input current value and the match
between the excitation frequency and the resonant frequency of the load;

0 Instrument noise.

The following sections present a POD model for the initial position of the gradiometer
with respect to the outlet strut as shown in Figure 9.2(a). The technique employed is

applicable for developing POD models for other parametric variabilities in the system.

9.2 POD Model
The probability of detecting a defect by an inspection system can be determined by
generating conditional probability density functions (pdfs) for the output parameter of

interest. Figure 9.3 shows examples of the output signal parameter distribution in the

presence of a defect, p(y|x1), and without a defect, p(y|x0). In this ﬁgure, y is the output

parameter and x0 and x1 represent the case where a defect is absent and present,

respectively.

187

10 mm

1
&
Direction of B Field

Figure 9.1 Optimal position and movement range for the gradiometer with respect to the
outlet strut

 

 

AX AY
‘—§
AZ
(a) (b) (c)
. 1A ,
I <13 /AB
(d) (c)

Figure 9.2 Examples of variations in the gradiometer position with respect to the heart
value outlet strut: (a)-(c) translational, (e)—(f) alignment angle.

188

1 Reject Accept

‘——__ _———>

   

 

 

 

y (signal parameter value)

Figure 9.3 Probability density functions of the output signal parameter when a defect is
present (p(y|x1)). and without a defect (p(y|x0)).

In a typical inspection system, a specimen is accepted when no defect is detected, and
rejected when a defect is found. The decision threshold, T, is chosen such that all
observed signal parameter values, y, that are below T are classiﬁed as defect signals and
those exceeding the threshold are interpreted as absence of defect. The probability of
detection is then given by
T

POD = jp(y|x1)dy. (9.1)

A false alarm occurs when a specimen without defect is rejected due to misinterpretation

of the signal. The probability of false alarm (PFA) is given by

189

T
PFA = Ip(y | x0)dy (9.2)

On the other hand, a false acceptance is the case when a specimen is accepted even when

a defect is present. The probability of false acceptance (POFA) is related to the POD as

POFA = 1- POD (9.3)

The POFA can also be obtained from

POFA = ojpjy I “‘1 )dy (9.4)
T

Figure 9.3 shows the regions under the two pdfs corresponding to the POD, PFA, and
POFA for a given threshold, T. The degree of overlap between the defect and no-defect
signal parameter distributions is typically a function of the defect characteristic such as
geometry and dimensions. If the characteristic of the defect is not distinctive under the
inspection modality, the overlap region between the two pdfs will be large, and the ability
to detect a critical defect is low. In this case, the detection ability of the system becomes
a function of T. The value of T may be selected based on one of the following three

criteria [180]-[183]:

1. Set the probability of false alarm (PFA) to a constant value.
2. Set the probability of false acceptance (POFA) to a constant value.

3. Minimize the total signal classiﬁcation error.

In the third criterion, the total signal classiﬁcation error, E( T), is a weighted sum of the

PFA and the POFA [180]-[183], [191] given by:

190

E(T)= a-POFA+ (i—a) PFA

°° T (9 5)
=ajp(y|x1)dy+(l—a) jP()’lx())dy '
T — 00
In (9.5), a is the prior probability of a defect being present in the specimen. The
threshold T is then obtained by solving
8E( T)
—— = 0 9.6
3T ( )

To calculate the POD, PFA, and POFA using (9.l)-(9.4), the signal pdf with and without
a defect are needed. Several probabilistic models for obtaining these pdfs are described
in details in [180]-[183] and [192]-[193]. These include the experimental, numerical, and

functional models.

In the experimental model, a large number of measurements are taken on the output to
generate the pdf for the signal. This approach is laborious, costly, and time consuming as
it requires the generation of a large observation set to produce statistically signiﬁcant
estimates. To overcome this limitation, a numerical model may be used to predict the
defect signal and effects of measurement variabilities. Although it is more cost effective
than the experimental model, the numerical model is often computationally intensive
since three dimensional numerical models must be simulated a large number of times to
produce the probabilities. The functional model, on the other hand, does not require

extensive human or computational resources. The following sections describe the

191

functional model and its application to POD evaluations for the gradiometer-based

examination system.

9.2.1 Functional Model

The functional model [180], [181] uses a small set of system output values to establish a
mathematical relationship between the output signal parameter, y, and a perturbed system
parameter, r. Examples of r include liftoff, probe alignment, and excitation signal level,
while examples of y include the maximum amplitude and largest phase angle of the
output signal. The output signal parameter values may be obtained from either physical
measurements or a numerical model. With a given perturbed system parameter value, r,
two sets of y data are collected—one with the inﬂuence of a defect, and the other without.
Two corresponding mathematical functions are then developed using simple curve ﬁtting

techniques. If the pdf of the system parameter perturbations is known, the conditional

pdfs of the output signal parameter with (p(y|x1)) and without (p(y|x0)) a defect can be

calculated using the proper functional model as shown in Figure 9.4. Once the
conditional pdfs are known, equations (9.1) and (9.2) are used to obtain the POD and

PFA, respectively.

9.3 Parametric Study
The outputs of the gradiometer-based prosthetic heart valve examination system represent
the scaled amplitude, R, and phase angle, 6, of the gradiometer signal. The results in
Chapter 8 show that the condition of the heart valve is related to the maximum value of 6.

Hence, in the POD formulation, the maximum value of 6 is the signal parameter of

192

interest, y. With a given threshold, T, the following sections describe a technique to
develop a functional model and calculate the POD for the examination system with
respect to the initial position between the gradiometer and the outlet strut. The technique

presented may be used to develop the POD with respect to other sources of variabilities.

Output Signal

System Parameter Parameter p df

Perturbations pdf _
Functional Model
'1-(fr)

System parameter value, r
Output Signal Parameter Value, y

Figure 9.4 Application of the functional model to obtain the output signal parameter

conditional pdf. Adapted from [180] and [181].

9.3.1 POD With Respect to the Gradiometer Distance from the Outlet Strut
Two BSCC heart valves, one with an IOS (ﬂange no. 719811) and the other an explanted

valve with a SLS (ﬂange no. A48652), have been chosen randomly for the POD study.
The objective of the study is to develop the POD for the examination system with respect
to the initial position, 2, between the gradiometer and the outlet strut. Figure 9.5 shows
the output phase angle, 6, for the two valves over ﬁve cardiac cycles obtained from the in
vitro test system (Chapter 8). Using the gradiometer movement data from Figure 8.16,

the output phase angle for the two valves as a function of distance are shown in Figure

 

1 Valve ﬂange number: 71981, S/N: 29MBRC 12407, condition: IOS
2 Valve ﬂange number: A4865, S/N: 29MBRC 61636, condition: explanted SLS

193

9.6. For comparison, offsets have been added to the results in this ﬁgure such that the
two curves intercept at z = 12 mm, corresponding to the maximum distance of the
gradiometer from the outlet strut during a cardiac cycle. A functional model, f, is
developed for each data set in Figure 9.6 to relate the input distance, 2, to the output

phase angle, 6. For the two outlet strut cases, the functional models are

9103 = 005(1) (97)

and
951.5 = f SLS (Z) (98)

t
In each case, a 7 order polynomial is ﬁtted to the data to produce a mathematical

model. Comparisons of the original and model generated outputs for the IOS and SLS

cases are shown in Figure 9.7 and Figure 9.8, respectively.

The perturbations of z is assumed to follow a Gaussian distribution with a selected mean

. 2 . . . . .
and ﬁxed variance of 0.4 mm . A ﬁxed variance is used Since, in practice, the

gradiometer oscillates about a mean position with the same amplitude regardless of the
distance between this position and the heart valve outlet strut. Figure 9.9 shows an input

pdf with a mean distance z = 2.5 mm from the outlet strut and a positional variance of 0.4

mmz. Applying this pdf to the functional models in (9.7) and (9.8) yields the pdf for the
output phase angles 6103 and QSLS in Figure 9.10 and Figure 9.11, respectively. Figure

9.12 shows the locations of these two pdfs. In this ﬁgure, the pdf of 6313 (dashed line)

194

corresponds the conditional probability p(y|x1), and the pdf for 6103 (solid line)

corresponds to the conditional pdf p(y|x0), of Figure 9.3.

Heart Valve Tests (6F Gradiometer,POT-9085,#2): e
10 1 1 1 1 1 1 ﬂ 1 1

 

‘— 71981 los
----- A4865 XSLS r

l
___...l _____ —1
l
l
l

l
l
l
|
_41

 

 

     

|
l
l
l
l
l

     

 

 

6 (degrees)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5
Tlme (s)
Figure 9.5 Output phase angle, 6, over time for a valve with an IOS (solid line), and one
with a SLS (dashed line).

195

Phase Angle 6 (Gradiometer: POT-9085,#2)

 

 

 

 

 

 

20 m - -s— ——1—- 1 1 1- _ _1
‘ ' ‘ * I 71961 105
' ----- A4865 XSLS 1
1 "' l
15. --—t——--——-.
' l
i l
a 10‘- 111111111
CD
9
U,
Q)
E
G 5. 7777777
0 -----~---‘~—----ﬁ——h;.}-;;;;--;;;lE—.-_--A-h-_—+---- , FFFFF 7i,
-5 ...EW- . __ . i .__--____ l
0 2 4 6 8 10 12 14

2 (mm)

Figure 9.6 Output phase angle, 6, as a function of z, the distance between the gradiometer
and outlet strut, for a valve with an 108 (solid line), and one with a SLS (dashed line).

Original and Model Generated

Data [71991(103)]: 9

 

T T; - 7 7 ‘
7 ' + Original 1
""" _ ' --+-- Model Gen. i '

 

 

‘ _ I

201

 

6 (degrees)

 

2 (mm)
Figure 9.7 Original and model generated outputs for the IOS case

196

6 (degrees)

Normalized Count

Original and Model Generated Data [A4865(XSLS)]: 6

 

 

I I I _*- ___ ,WWJ ,1 , ,_

' I l + Original I V

 

o 2 4 ' 6 9 160' ‘1? ' 134
2 (mm)
Figure 9.8 Original and model generated outputs for the SLS case

Gaussian Input PDF (“=25 mm, 0:0.625 mm)
0.07.. , _ _.

 

 

 

 

; I j j P l
0.06L 1 _i I _i-_ ,,,,,,, 1 _______
0.051
0.04 eeeeeeeee * .................... ’ ....... i ...... z

I .

0.03I ..... 1 ................
0.02; .......
0.01 eeeee ‘ ..... LII _______ I _____

0 _ I .lili iilll ' ___,

-1 o 1 2 3 4 5 6

z(mm)

Figure 9.9 Example of a Gaussian distribution for the input pdf (2)

197

 

(102L1

OIHv

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11

6 (degrees)
Figure 9.10 PDF of 6103 for heart valve 71981 with an IOS (z = 2.5 mm)

PDF for the Output Phase Angle (6): A4865
(106I““"' l l l ‘T’T l

 

I
l
|
l
1
l
l
1
l

\ ‘
WA!
ll
II

I

    

—T
l
1
l
I
l
l
T
l
I
1
I
4
|
I
I
l
l
i 1
_ ----___‘___.__-'__
l
I l
l l
T *1 l
I 1 l
_—.,&_.,11_.,.s—.L-_ ___—5....____
' 1
I

I l . E
l I I I

0.02 ------ 1 eeeeee ————— ~
l I l

 

0.01M”;

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0L 1 ., ---,_-
43.3 -o.25 -o.2 9.15 -o.1 -o.05 o 0.05 0.1
6 (degrees)

Figure 9.11 PDF of QSLS for heart valve A4865 with a SLS (z = 2.5 mm)

198

PDFs for the Output Phase Angle (6)

 

 

0.07 —— .~——r___ , , ..__.

 

 

 

 

 

 

 

 

 

 

 

1 71981IOS l f I : II
2 ----- A4865>GLS 1 1
0.06? u 1 -» ............................... I
I 1
0.051 AAAAAAAAAAA 1 ——————————————————————————————— — —- ~
0-045“ --------- 1 ——————————————————————————————————— _- I
l
1. I
ll
0-03 7777777777 H ““““““ T —————————————————————— ___ ..
: I 1 1 1
ll . , 1
0.021~- . -, eeeeeeee .......... -__---____ __
|
0.01 ﬁg ..................................
1 It
.1 I
u 1
.. 1 l 1
01 ll____ ___1_ 1. _ l
-5 0 5 10 20

 

6 (degrees)
Figure 9.12 PDF for 6103 (solid line) and QSLS (dashed line), (z = 2.5 mm)

9.3.1.1 POD Curve

With the conditional probability for 6103 and 6315, the corresponding POD and PFA are

calculated using the integrals in (9.1) and (9.2) for the current the speciﬁc 2. The
threshold, T, obtained by ﬁxing the PFA to 5 %, is 18.01 degrees. Note that since the two
pdfs in Figure 9.12 do not overlap, the POD is 1. If the shortest distance between the
gradiometer and the outlet strut plane is non-zero (i.e., an initial position of z > 0), the
separation between the two pdfs decreases. Figure 9.13 shows the pdfs overlapping when

2 = 12 mm. Figure 9.14 shows a magniﬁed view of the overlapped region between the

two pdfs. In this case, T is 0.0188 degrees and the corresponding POD is 1.66x10-5.

Figure 9.15 shows the POD curve with respect to the initial position (shortest distance)

between the gradiometer and the outlet strut. In this ﬁgure, the POD remains unity until

199

the gradiometer is at a distance from the outlet strut. At 2 > 10 mm, the phase angle of
the output signal between the IOS and SLS cases become indistinguishable and the POD
decreases rapidly towards zero. These results agree with the mathematical (Section
4.10), numerical (Section 5 .2), and experimental results (Section 8.4) in suggesting that
the gradiometer be positioned within 5 mm of the outlet strut for a heart valve

examination.

The POD curve for the gradiometer-based examination system with respect to the initial
position between the gradiometer and the outlet strut has been developed. The POD
curve shows that the examination system is reliable when the gradiometer is placed
within 5-10 mm of the outlet strut. The technique presented may be used to develop
POD curves for other sources of system parameter variabilities.

PDFs for the Output Phase Angle (6)

i 9 1 f 1 1 71981 IOS
‘ ‘ ' - f _ j ----- A4865 XSLS f

 

 

‘I 7 —ﬁ
:

 

 

 

 

________________________

 

 

1.5 2 2.5 3 3.5
0 (degrees)

Figure 9.13 PDF for 6103 (solid line) and QSLS (dashed line), (z = 12 mm)

200

 

 

 

PDFs forthe Output Phase Angle (6)

 

. ; ' 1‘ A _ 71981 IOS

0.4..1 ........ 7

 

0.35;.

0.25j

0.2L ........ *

 

 

 

l I
l ”’I

0 -a-e-J‘ff’. _ .I
1

,. ,,2..l_

or 0.05

 

9 (degrees)

Figure 9.14 Magniﬁed view of the overlapped region between the two output phase angle
pdfs (z = 12 mm)

POD vs Mean Distance From Outlet 8th

 

 

 

 

0.8 ------ is — r ................. .- ........ g
D 0.6f ~~~~~~~ — —- E————~-—i —————— ‘_ ....... D‘
O ' l l
“- l : : : . . i

l I I ' : : .
-------- . -~ ~: ~~~~~~~~ : ~~~~~ :—- —— 1 —— *1
l ' ' I I l :1

t | l l I I :

T I I I I l

0.2 -------------- .- ~ ~—i ———————— i~ ——————— i
i : I I l 1 1

ol__ ___ ~_ __g ___;_ _ _L ' __J
0 2 4 6 8 1o 12 14

z(mm)

Figure 9.15 POD with respect to the distance between the gradiometer and outlet strut

201

CHAPTER 10. SIMULATION OF INDUCTIVE HEATING EFFECTS

A radio frequency (RF) radiation source induces electric charges or currents in nearby
conducting objects. In the gradiometer-based method, the excitation ﬁeld induces
currents in the ﬂange and struts of the BSCC heart valve, resulting in resistive (joule)
heating of the metallic parts. This heating process is also known as inductive heating.
The rise in the heart valve temperature and the surrounding environment such as the
suture ring and the myocardium under RF tone-burst excitation was previously studied
using ﬁnite element analysis and reported in [155]. In this chapter, we present a similar
model to investigate the temperature rise under continuous wave excitation for the
gradiometer-based method under two conditions. In the ﬁrst case, the model takes into
consideration the temperature-regulating effect of blood circulation around the prosthetic
heart valve as well as the heat conduction from the heart valve into the suture ring and the
myocardium. In the second case, the hypothetical situation where no convective or
conductive heat loss occurs from the heart valve into its surrounding environment is
examined. This scenario provides an estimate on the worst-case temperature rise of the
prosthetic heart valve due to inductive heating. Effects of medium and long duration
excitation (3 and 15 minutes, respectively) are also investigated. The results of these
studies show that the temperature rise of the heart valve due to excitation by the
gradiometer-based examination system is negligible, and that the system is safe for

patient use.

202

10.1 Model Formulation
To determine the time-varying temperature, T, of the prosthetic heart valve, a variation of
the Pennes’ [156], [157] bioheat equation is used. In Pennes’ formulation, the thermal

energy balance for perfused tissue is expressed as:
8T
pCpE—kV2T=Q+qm+qb (10.1)

where
qb = VPbe(Tb — T) (10.2)
and
. 3
,0. denSIty (kg/m ),

C 1,: speciﬁc heat capacity (J/(kg-°C)),
k: thermal conductivity (W/(m-°C)),

Q: power density from inductive heating (W/m3),

qm: metabolic heat production rate per unit volume (W/m3),
qb: arterial-tissue heat transfer rate per unit volume (W/m3),
V: perfusion rate per unit volume (5-1),

pb: density of blood (kg/m3),

C b3 speciﬁc heat capacity of blood (J/(kg-°C)),

Tb: arterial blood temperature (°C), assumed as core temperature (37 °C).
The power density from inductive heating, Q, is given by
l 2
Q=§dﬂ no»

where O'is the electrical conductivity (S/m). The electric ﬁeld intensity, E, (V/m) is

governed by the time-harmonic equation

203

(jwa—w26)E+Vx[%VxE]=—ij (10.4)

In (10.4), w=27zf is the angular frequency (rad/s) associated with the excitation

frequency, f (Hz), Eis the permittivity (F/m) and p. is the permeability (N/AZ). The

external current density J (A/mz) is related to the externally applied magnetic ﬂux

density, B (T), as J = i(VXB).
,u

In the model formulation, the metabolic heat production process, qm, in (10.1) is

neglected as it is small compared to Q, the energy generated from inductive heating. The

arterial-tissue heat transfer process, qb, between the perfused blood and the myocardium

and that with the body muscle are also neglected as the temperature regulation of the

prosthetic heart valve and its surrounding tissues is dominated by the circulating blood in

the atrial-ventricular chamber. By ignoring the contributions from qm and qb, the

temperature effects due to Q are also more readily observed.

The convective heat exchange process with the circulating blood is described by:

k%£—=hb(T—Tb) (10.5)

where n is the unit vector normal to the boundary and hb (W/(m2-°C)) is the convective

heat transfer coefﬁcient, or ﬁlm coefﬁcient. The temperature, Tb, of the blood entering

the atrium is at the core temperature (37 °C). By substituting different values for the

204

thermal conductivity, k, and the convective ﬁlm coefﬁcient, hb, the heat ﬂux at the blood-

valve and blood-myocardium boundaries may be calculated. In the numerical model,

convection ﬁlm coefﬁcients listed in Table 10.1 from swine study data [158] are used.

Table 10.1 Selected convective ﬁlm coefﬁcients. After [158]

 

Boundary hb (W/(m2-°C))

 

Circulating Blood Stagnant Blood

 

 

 

Blood—Left Atrium 5700 2430
Blood—Flange Ring 6830 2355
Blood—Left Ventricle 4480 2280

 

 

 

 

 

With the omission of the metabolic heat generation rate, qm, and the tissue perfusion heat
exchange, qb, terms, equation (10.1) becomes the heat equation
8T
pCp—g-kvzhg (10.6)

The governing equations (10.5) and (10.6) are solved numerically to simulate the heat

exchange between the prosthetic heart valve and its surrounding environment.

10.2 Model Description

In the gradiometer-based examination system, the prosthetic heart valve is aligned such
that the ﬂange plane is parallel to the direction of the applied ﬁeld. This orientation
causes the currents induced in the outlet strut to dominate over those induced in the

ﬂange. In the worst-case scenario, the heart valve may be positioned such that the ﬂange

205

plane is orthogonal to the direction of the applied ﬁeld. In this orientation, the maximum
amount of eddy current is induced in the ﬂange, resulting in the largest temperature rise
of the valve due to inductive heating. This worst-case scenario is investigated to ensure

patient safety.

A two-dimensional axisymmetric ﬁnite element model of an implanted BSCC heart valve
is shown in Figure 10.1. In this ﬁgure, the vertical line at the center is the axis of
symmetry. The outer-most layer represents the (generic) body muscle. The middle layer
is cardiac muscle, and the inner-most atrial-ventricular chamber is ﬁlled with
“circulating” blood. Figure 10.2 shows the model of the heart and the BSCC heart valve
ﬂange ring. The inside surface of the ﬂange is in contact with the atrial and ventricular
blood in the respective chamber. The outside surface of the ﬂange is in contact with the
woven Teﬂon fabric suture ring, which in turn is in contact with the myocardium. The
dimensions and conﬁguration of the ﬂange and suture rings are given in Figure 10.3.
Note that the inlet and outlet struts, as well as the occluder disc, have been excluded from

the model to reduce computational complexity.

A uniform magnetic ﬁeld with peak-to-peak amplitude of Ba and frequency f is applied

along the vertical (2) direction. To apply this ﬁeld, a circumferential surface current, JScp,

is imposed along the outside boundary of the body muscle such that

-ns><H=Jsa (10.7)

where n3 is the unit vector normal to the surface of the body muscle.

206

The core temperature of the model is set at T0 = 310 K (37 °C). The outer boundary of

the body muscle is set at a constant temperature of T = To, as is the upper boundary of the

atrium (where the blood “enters”). The systole and diastole durations are set at 0.3 s and

0.5 s, respectively, corresponding to a heart rate of 75 beats per minute. The material

properties of the ﬂange ring (HAYNES® 25 alloy) are obtained from [159], the suture
ring (woven Teﬂon fabric) from [160], and those for the blood and cardiac and body
muscles are from [161] and [162]. The data is summarized in Table 10.2. The material

properties of interest are the density, speciﬁc heat coefﬁcient, permeability and the

electrical and thermal conductivities.

Table 10.2 Material properties used in the ﬁnite element model. After [159]-[162]

 

 

 

 

 

 

 

 

kg W J {5) (F)
— k C _ .9 _
ﬁlms) 1m) 1m) ,. m
Flange Ring 9130 9.4 385 1123600 1
Suture Ring 2160 0.25 1400 0 2.1
Blood 1000 0.627 4178 0.67 2740
Cardiac Muscle 1070 0.59 3500 0.5 50000
(myocardium)
Body Muscle 1050 0.5 3600 0.575 26050
(Generic)

 

 

 

 

 

 

 

 

207

i Myocardium
I

 

Atrium

0.03 m //
0.01 m I Sumfe

I i /” Ring

 

 

 

 

 

 

 

 

........ “i
0.06 m 0.04 m \
\Heart Valve
\ Flange Ring
\
0.03 m Body Muscle Ventricle

(Generic)

 

 

~— 0.06 m

 

Figure 10.1 Geometry of the axisymmetric ﬁnite element model. The vertical line at the
center is the axis of symmetry. (See Figure 10.2 and Figure 10.3 for more details.)

2 (m)

 

0'01 Suture

Atrium Ring

\

2.78x10'3
0

Heart Valve/

Flange Ring 0 01

Ventricle

-0.02

 

-0.03

 

 

 

-0.04 (m)

r

0 0.01 0.02 0.03

Figure 10.2 Conﬁguration and dimensions of the heart and the heart valve ﬂange ring at
the center of the ﬁnite element model.

208

 

 

 

 

 

 

 

 

 

BSCC

Heart Valve
1 I.

1.346‘mm

2.870 mm suiure 5.563 mm
1 Ring

0.826'mm ,a a = g... =

1.537 mm 2.811 mm

 

Figure 10.3 Cross-sectional dimensions of the ﬂange and suture rings used in the ﬁnite
element model.

10.3 Simulation Results

In the gradiometer-based examination system, a 1.5><10.4 T (1.5 gauss) excitation ﬁeld is

applied to the BSCC heart valve in intervals of 90 seconds, with a maximum of six
intervals per examination. Between the excitation intervals, the magnetic ﬁeld is
removed for a period of more than 5 minutes, during which time the heart valve and the
surrounding structures return to the core temperature. To determine the temperature

variations of the prosthetic heart valve with a conservative safety margin, a sinusoidal

magnetic ﬁeld of frequency f = 62.5 kHz and a 2x10.4 T (2 gauss) peak-to-peak

amplitude (BO) is applied as the excitation ﬁeld in the FE model simulation. The

maximum temperature elevation over 3 and 15 minutes of continuous exposure are

examined. Also, two special cases are investigated to study the heating effects of the

209

prosthetic heart valve and the surrounding structures: (1) Model with conduction and
radiation; (2) Model with radiation but no convection or conduction. In the ﬁrst case,
blood circulation is stopped in the model and the blood in the atrial-ventricular chamber
becomes stationary. This essentially removes the convection cooling of the prosthetic
heart valve and myocardium. This arrangement allows the study of conduction and
radiation heating of the tissue, as well as the heating of any stagnated blood within the
heart. In the second case, both the convection and conduction cooling of the prosthetic
heart valve and its neighboring structures are eliminated. This presents the worst case
scenario in which the temperature of the prosthetic heart valve continues to rise due to
inductive heating and the lack of heat dissipation except via thermal radiation to the
surrounding structures. The results of these two studies are presented later in this

chapter.

10.3.1 Maximum Temperature Increase After Three Minutes of Exposure

Figure 10.4 shows the spatial temperature elevation (AT = T— To) proﬁle of the model

after 180 seconds of continuous excitation. The time step interval is 0.01 seconds and the
measurement unit is Kelvin. The background “dark” areas of this ﬁgure represent
regions with AT = 0. The “highlighted” area around the heart valve shows the
temperature elevation due to the heating of the prosthetic heart valve. A magniﬁed view
of this region is shown in Figure 10.5. In this ﬁgure, localized temperature elevations are
visible around the outer “rims” of the ﬂange ring. Figure 10.6 shows a close-up view of

this heart valve region. In this ﬁgure, the location with the highest instantaneous

210

temperature (Tm = 310.002953 K) at t = 180 s is also indicated. The maximum

temperature rise at this point is AT: 2.953x10.3 degrees.

0.04

.0
o
o N

I
.o
o
N

Axial Distance, z (m)

-0.04

-0.06

 

Temp. Change (K), Magnetic Field (Arrows)(A/m),

11111111
11111111
11111111'
111111111
111111111
111111111

r(.111111

111111111

1111111111

1111111111
11111111111
111111111111111
111111111111111
111111111111111
111111111111111

-0.02 0 0.02 0.04 0.06

Radial Distance, r (m)

 

Max: 2.9829-3
x103

2.5

 

 

 

 

0.08
Min: 0

Figure 10.4 Finite element analysis results showing the spatial temperature elevation
proﬁle of the BSCC heart valve, suture ring and the surrounding biological materials after

180 seconds of continuous excitation.

211

Temperature Change, Delta_T = T-TO (K) Max: 2.982e-3

 

 

 

x10'3
0.01
2.5
0.005
E
N - ‘2
o“
E 0
3 - -1.5
3
{0005
~ 1
-0.01 05
0.005001 7 0.015 0.02 0.025 0
Radial Distance, r (m) Min: 0

Figure 10.5 Finite element analysis results at t = 180 5 showing localized temperature
elevations around the outer “rim” of the ﬂange ring.

212

Temperature Change, Delta_T = T-TO (K), at t = 180 s Max: 2.952e-3

 

 

x10‘3 x10‘3
3 .
, 2.5
2
1 - 2
- -1.5

Axial Distance, 2 (m)
o

 

 

 

 

o 1
_2 max: 310.002953 ‘3
-3 1' - 1. I" 0.5
0.009 0.01 0.011 0.012 0.013 0.014 0.015 0.016 0
Racial Distance, r (m) Min: 0

Figure 10.6 Magniﬁed view of the temperature elevation proﬁle on the BSCC heart valve
at t = 180 s. The location of the peak instantaneous temperature (Tmax = 310.002953 K)

is as indicated.
Figure 10.7 shows the temperature variations of the heart valve over 180 seconds of
exposure at point “A” of the peak instantaneous temperature as indicated in Figure 10.6.
The results in this ﬁgure show that the valve temperature varies cyclically over time due
to the effects of inductive heating and cooling from the perfusion of “cooler” blood into
the atrial-ventricular chamber. The upper envelope of the temperature function peaks at a
value just below 310.003 K. Figure 10.8 shows the temperature variations from t = 171—

180 s. From this ﬁgure, the last temperature peak for the 180-second simulation occurs

at t = 177 s, with a maximum temperature of Tmax = 310.002983 K (AT: 2.983x10_3 K).

213

This indicates that the maximum temperature elevation proﬁle does not occur at t = 180 s,
but rather, at t = 177 s. Figure 10.9 shows the temperature elevation proﬁle at t = 177 s.

A magniﬁed view of this proﬁle at the prosthetic heart valve and the point of highest

instantaneous temperature are shown in Figure 10.10.

Heart Valve Temperature Over Time

 

 

 

 

310.003 ; r
310.0025 _. ............. ................ ................ ................ ................ ................ ............ .......... .
310-002 ............... ................ ................ ................ ................ ................ ................ ............... .
g s s z : 2 = ' -
a i i i a 2 2 i 5
8310.0015 .............. ................ ................ .............. ................ ........... q
0 3 f 5 3 5 : : :
g- . .
'—
310.001 ................................................................................................................................................... 4
3100005 . ............... ................ ................ ................ ................ ................ ................ ........... .
0 20 40 60 80 100 120 140 160 180
Time (s)

Figure 10.7 Finite element analysis results showing the BSCC heart valve temperature
over 180 seconds of exposure to the gradiometer-based examination system (observed at
the peak temperature point as indicated in Figure 10.6).

214

 

 

 

 

Heart Valve Temperature Over Time
310.003 . ................................................................................................................................ 4
a": .
23100029 .. . , __________ _,
3
g E
0 I
§
..
310.0028 .. ......................................................................................................................................... ”-1
172 173 174 175 176 177

178 179 180
Time (s)

Figure 10.8 Finite element analysis results showing the last temperature peak in a 180-
second simulation at t = 177 s, with Tm = 310.002983 K.

Results from Figure 10.4-Figure 10.10 show that the maximum temperature rise is
limited to AT: 2.983x10-3 K (Tmax = 310.002983 K) after 3 minutes of exposure to the

excitation ﬁeld. Furthermore, there is no temperature rise in the tissue regions outside of

the heart (myocardium).

215

Temperature Change (K) and Magnetic Field (Arrows)(A/m) Max: 2.982e-3

    

 

 

 

 

   

0.04 1111111111
1111111111

m 1111111111
1111111111

:8: 1111111111

:1— 0 1(11111111 ‘2

3 1111111111

€0.02 1111111111 ~15

.3 1111111111
1111111111

004 1111111111
1111111111

_006 1111111111
1111111111

-0.02 0 0.02 0.04 0.06 0.08

Radial Distance, r (m) Min: 0

Figure 10.9 Spatial temperature elevation proﬁle of the BSCC heart valve, suture ring
and the surrounding biological materials after 177 seconds of continuous excitation.

216

Temperature Change, Delta_T = T-TO (K), at t = 177 5 Max: 2.982e-3
x103 7 x10‘3

 

   

 

3
2.5
2
£1 12
N
o“
g 0
§ 1-1.5
O
2.1

 

 

 

 

   

0.009 0.01 0.011 0.012 0.013 0.014 0.015 0.016
Racial Distance, r (m) Min: 0

Figure 10.10 Magniﬁed view of the spatial temperature elevation proﬁle on the BSCC
heart valve at t = 177 s. The point in the ﬁgure indicates the location of the peak

instantaneous temperature (Tmax = 310.002983 K).

10.3.2 Maximum Temperature Increase After Fifteen Minutes of Exposure
Figure 10.11 shows the temperature variation of the BSCC heart valve ﬂange ring

(observed at the peak temperature point indicated in Figure 10.6) over 15 minutes of

exposure to the excitation ﬁeld. In this case, the valve temperature reaches Tmax =

310.00296 K (AT = 2.959x10-3 K). The corresponding temperature elevation proﬁle at t

= 900 s is shown in Figure 10.12. A magniﬁed view of this proﬁle for the heart valve

217

and the neighboring structures is shown in Figure 10.13. The results from these ﬁgures
suggest:

1. The temperature of the prosthetic heart valve and the surrounding structures do
not exceed 310.003 K (37.003 °C) even after 15 minutes of continuous exposure
to the magnetic ﬁeld of the gradiometer-based examination system.

2. The RF heating on the patient body is localized to the regions immediately
surrounding the heart valve (Figure 10.12). Exposure to the RF energy from the
gradiometer-based examination system does not cause any temperature rise in
regions outside of the myocardium (e.g., the “body muscle” region in Figure

10.1).

Heart Valve Temperature Over Time

”11111111111111

310.003

l1 1" it "1")" " "’1"W‘1"1"' """"'1'"’1"'1'11””11’1'" 'l 1' "1111"! '1 "1'1" "1'1 " " "1""

1 1 111111 ll 1 1111 1111

”.vr
|.

 

310.0025

310.002 -

310.0015 1

Temperature (K)

310.001 1

310.0005 -

 

 

 

310 a L l A n n A a
0 1 00 200 300 400 500 600 700 800 900
Time (s)
Figure 10.11 Finite element analysis results showing the BSCC heart valve temperature
(observed at the peak temperature point indicated in Figure 10.6) over 15 minutes of
exposure to the excitation magnetic ﬁeld

218

 

 

 

 

 

 

 

   

Temperature Change (K) and Magnetic Field (Arrows)(A/m) Max: 2.9599-3
’1 I 1 ’I‘ 'l‘ ’I‘ 'l‘ '1 x10'3
0.04 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1
0.02 1 1 1 1 1 1 1 1 2.5
1 1 1 1 1 1 1 1
E 1 1 1 1 1 1 1 1 _2
2',— 0 1‘ 1 1 1 1 1 1
,‘g’ 1 1 1 1 1 1 1 1
50.02 11 1 11111 1.5
g 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 _1
-0.04 1 1 1 1 1 1 1 1 .
1 1 1 1 1 1 1
-0 06 1 1 1 1 1 1 1 0-5
' 1 1 1 1 1 1 1
-0.02 0 0.02 0.04 0.06 0.08 0
Radial Distance, r (m) Min: 0

Figure 10.12 Spatial temperature elevation proﬁle of the BSCC heart valve, suture ring
and the surrounding biological materials after 900 s of continuous excitation. T0 = 310 K.

219

Temperature Change, Delta_T = T-TO (K) Max: 2.959e—3
X103 X10-3

2.5

- -1.5

Axial Distance, 2 (m)

 

 

 

0.004 0.008 0.012 0.016 0
Racial Distance, r (m) Min: 0

Figure 10.13 Magniﬁed view of the spatial temperature elevation proﬁle of the BSCC
heart valve at t = 900 5. T0 = 310 K.

10.3.3 Special Case I: Excitation Without Blood Circulation

To investigate the inductive heating of the prosthetic heart valve and the temperature
increase in the neighboring regions in the absence of blood circulation, the convective
ﬁlm coefﬁcients are set to zero in the ﬁnite element model (see Equation (10.5). This
eliminates the convective cooling of the structures as well as the automatic blood
temperature regulation in the heart. This model may also be used to describe the
temperature rise of any stagnated blood pool in the heart when heated indirectly by the

thermal radiation from the prosthetic heart valve which is, in turn, being induction heated.

220

Figure 10.14 shows the temperature elevation proﬁle in the vicinity of the prosthetic heart
valve after 15 minutes of continuous excitation. A three-dimensional view of this proﬁle
is shown in Figure 10.15, and a magniﬁed view of the myocardium and the inner region
of the body tissue is shown in Figure 10.16 . From the results in Figure 10.14-Figure
10.16, the maximum temperature rise is limited to AT = 0.0652 K. The results also show
that the temperature elevation is spatially localized to within a 4 cm-radius from the
center of the ﬂange ring. Figure 10.17 shows the magniﬁed view of the prosthetic heart
valve spatial temperature proﬁle after 15 minutes of excitation without convection

cooling (no blood circulation).

221

Temperature Elevation Proﬁle (K)(t = 9005) Max: 0.0652

 

 

 

0.06
0.04
. 0.05
0.02 * '
E
N _ _
a; o 0.04
8
a
.2
E-o.02 - -o.03
.2
2
pm 0.02
-0.06 0.01

 

 

 

 

-0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1 0
Radial Distance, r (m) Min: 0

Figure 10.14 Spatial temperature elevation proﬁle of the BSCC heart valve, suture ring
and the surrounding biological materials after 15 minutes (t = 900 s) of continuous
excitation without blood circulation (convective cooling).

222

Temperature Elevation Proﬁle (K)(t = 900s) Max: 0.0652

 

 

.‘lo.os\(m) 0.06
.0

__ -p.05 5

0.06

‘ 0.04
T-TO

 

 

 

 

 

0

 

Min: 0

Figure 10.15 Three dimensional view of the proﬁle from Figure 10.14.

223

Axial Distance, 2 (m)

Temperature Elevation Proﬁle (K)(t = 900s) Max: 0.0652

 

 

0.02 0.06
0-01 0.05
0 ~ ~ 0.04
-o.o1
- ~ 0.03
-0.02
-o.03
-0.04

 

 

 

-0.01 0 0.01 0.02 0.03 0.04 0.05
Racial Distance, r (m) Min: 0

Figure 10.16 Temperature elevation from inductive heating (with no blood ﬂow) is
spatially localized to a region of 4 cm radius from the center of the ﬂange ring.

224

Temperature Elevation Proﬁle (K)(t = 9005) Max: 0.0652

 

 

 

 

 

 

0.01 0.06
0.05
,2 0.005 ‘
E
N_ - -0.04
O
E 0
E — - 0.03
.9;
2
-o.005 ‘0.02
0.01
-0.01
o 0.005 0.01 0.015 0.2 o
Racial Distance, r (m) Min: 0

Figure 10.17 Magniﬁed view of the spatial temperature elevation proﬁle of the BSCC
heart valve at t = 900 s with no convection cooling. The peak of the instantaneous
temperature is located at the center of the heart valve cross-section. The maximum

temperature rise is AT = 0.0652 K. (To = 310 K).

10.3.4 Special Case 11: Excitation without Convection and Conduction Cooling

In this case, a hypothetical scenario was considered and the temperature rise of the heart
valve and the body tissues from inductive heating in the absence of convective and
conductive cooling was studied. The results from this study may serve as the theoretical
upper bound for the temperature increase in the heart valve and the neighboring regions.
In this study, the convective ﬁlm coefﬁcients for the blood—ﬂange ring interface in
Table 10.1 are set to zero in the ﬁnite element model. The thermal conductivity, k, values

in Table 10.2 are also set to zero except that for the ﬂange ring. This essentially

225

eliminates the convective cooling by the circulating blood (Equation (10.5)), as well as
the conductive heat transfer from the heart valve to its surround environment. Figure
10.18 shows the temperature of the heart valve after 180 seconds of continuous excitation
under this special heating condition. The measurement point, P, is at the center of the

valve cross-section as shown in Figure 10.19.

Heart Valve Temperature Over Tlme
310.73 1 , ...m . , - .___w

l
I ‘ l
l

 

‘1' ' T _ ' ‘_ " T—'—
l

l l

I

Temperature (K)

 

,,,l .i,m ..__nm ..# l- __ _L_.

o 20 4o 60 80 100 120 140 160 180
Tlme(s)

 

Figure 10.18 Finite element analysis results showing the BSCC heart valve temperature
over 180 seconds without convective and conductive heat transfers.

226

BSCC
Heart Valve

\

3.

 

 

P o Suture
Ring

 

 

 

 

Figure 10.19 Measurement point, P, for the second special case (11) temperature study

In this ﬁgure, the valve temperature rises in an approximately linear fashion to T =
310.6682 K after 180 seconds of continuous excitation in the absence of convective and
conductive cooling. Figure 10.20 shows the temperature of the heart valve after 15
minutes of exposure. In this case, the valve temperature reaches T = 313.34 K, showing
an elevation of AT = 3.34 K. Figure 10.21 shows the temperature elevation proﬁle of the
heart valve and its surrounding structures at t = 900 s. This ﬁgure is then processed to
produce Figure 10.22, which shows the temperature elevation proﬁle for the heart valve

alone. (The temperature scale to the right of this ﬁgure shows only markings of 3.341

degrees since the range variation is too small (less than 10'4 degrees) for proper labeling.

Note that the “ripple” pattern around the ﬂange ring in Figure 10.21 is due to the
accumulation of numerical noise over 90000 calculation time steps (a 15-minute
simulation duration with a time step size of 0.01 seconds). To better observe the
numerical noise, the early time (t = 0.01 5) temperature proﬁle from the ﬁnite element

simulation is shown in Figure 10.23. In this ﬁgure, the temperature of the ﬂange ring has

227

just begun to rise (maximum temperature rise, AT, is 4.21x10-5 degrees) and the effect of

the numerical noise, though small in amplitude, is visible. Since there is no heat transfer
into this region due to the lack of convection and conduction from the heart valve, the
numerical noise becomes the dominant signal. The numerical noise is accumulated over
time, and eventually manifests into the “ripples” seen in Figure 10.21. This numerical
noise does not affect the accuracy of the simulation results for the most part, especially in

the cases where heat transfer exists between the heart valve and the neighboring regions.

Figure 10.21 and Figure 10.22 show that the temperature elevations are highly localized
to the prosthetic heart valve under no conduction or convection cooling. Regions outside
of the heart valve show no temperature rise after 15 minutes of continuous exposure to
the RF energy from the gradiometer-based examination system. This result is expected
as the generated heat is mostly retained and not dissipated by conduction or conduction

into the neighboring structure.
Figure 10.24 combines the results from Figure 10.11 and Figure 10.20 to show the heart

valve temperature elevation over 900 seconds of continuous excitation for the normal and

special (no convection and conduction) cooling cases.

228

 

Heart Valve Temperature Over Time

313 __ ............... ................. ............... .............. .
312.5 _ ................ ................. ................. .............. ................ .

312 _ ................ ................. ................. _ ...... ............. .

Temgerature, T (K)

311_ .............................................. . ........................................ . ...................................................... , .......... ..

3 1 .. ........................................................................................................ . ..................................................... ..
I u . 1 e . . . a

 

 

310

 

0 100 200 300 400 500 600 700 800 900
Time, t (s)

Figure 10.20 Finite element analysis results showing the BSCC heart valve temperature
without convection and conduction cooling over 900 seconds of continuous exposure to
the excitation magnetic ﬁeld.

229

Temperature Change, Delta_T = T-T0 (K) (Time=900 5) Max: 3.341

—L

Axial Distance, z (m)
o

 

 

 

-1
-2
-3
.4
0.008 0.01 0.012 0.014 0.016 0.018 -1
Radial Distance, r (m) Min: -1.017

Figure 10.21 Magniﬁed view of the spatial temperature elevation proﬁle of the BSCC
heart valve without convection and conduction cooling at t = 900 s. To = 310 K.

230

Temperature Change Proﬁle, T-TO (T ime=900 3) Max: 3.341
x10'3 1 3.341

 

r - 3.341

 

" 1 3.341

 

 

 

Axisl Distance, 2 (m)
C)

 

 

 

 

 

 

 

 

0.01 0.011 0.012 0.013 0.014 0.015 0.016 0.017 0.018
Racial Distance, r (m) Min: 3.341

Figure 10.22 Temperature proﬁle inside the heart valve at t = 900 s after a continuous

exposure to the excitation signal without conductive and convective cooling. To = 310 K.
(Note that the temperature variation is too small, < 10e-4 degrees, for proper labeling of
the temperature scale)

231

Temperature Change, Delta_T=T-T0 (K) Max: 4.21e-5
-5
x10

 

Axial Distance, 2 (m)

 

 

 

0.009 0.011 0.013 0.015 0.017
Racial Distance, r (m) Min: -1.203e-5

Figure 10.23 Temperature elevation proﬁle at the beginning of the simulation (t = 0.01 5)
showing the effects of numerical noise. The maximum temperature rise at this instance is

AT=4.21><10'5 K.

232

Temperature Change Over Time
3.5 I I I I j I I I

 

\.

2.5

 

 

/- Normal Case
1 . .1” -'—'-Special Case _

 

 

 

Temperature Change, AT (K)
5?
\

 

 

 

 

-0 ' 5 L l l I l l I i
0 1 DD 200 300 400 500 500 700 800 900

Time,t (s)

Figure 10.24 Temperature of the BSCC heart valve with normal (solid line) and
special/worst case (broken line) cooling conditions over 15 minutes. To = 310 K.

The results from Figure 10.18 and Figure 10.24 are summarized in Table 10.3. The data
from this table show that, for the hypothetical worst-case scenario without convective and
conductive cooling, the maximum temperature rise of the prosthetic heart valve is
approximately AT = 3.3408 degrees for 15 minutes of exposure. These results show that
even in the hypothetical worst-case scenario, the maximum temperature rise after 15
minutes of exposure does not exceed 3.5 degrees. Moreover, the results from Figure

10.12 and Figure 10.21 show that there is no temperature rise in the tissues outside of the

233

myocardium under both normal and near-adiabatic (no convection or conduction cooling)

conditions.

The data in Table 10.3 also shows that the temperature of the BSCC heart valve does not
rise beyond AT = 0.003 degrees after 15 minutes of exposure under normal heat transfer
conditions. In practice, a temperature rise prediction of 0.003 degrees is close to the
accuracy limit of the numerical model and should be considered negligible for all
practical purposes. Moreover, in a biological system such as the human body, a
temperature rise of such small magnitude is not perceivable and we do not expect any

adverse effect.

The ﬁnite element analysis results presented in Table 10.3 show that induction heating of
the BSCC heart valve from exposure to the excitation ﬁeld of the gradiometer-based

examination is negligible, and that the system is safe for patient use.

Table 10.3 Summary of maximum temperature elevations of the heart valve after 180 and
900 seconds of continuous exposure, respectively, for the normal and worst-case heat
transfer scenarios.

 

 

 

180 seconds 900 seconds
(T0 = 310 K) (3 minutes) (15 minutes)
Normal cooling 0.002983 degrees 0.002959 degrees
wormase: w“1?°“‘C°“.VeC“°“ and 0.6682 degrees 3.34080 degrees
conduction cooling

 

 

 

 

 

234

CHAPTER 11. PATIENT SAFETY AND REGULATORY

COMPLIANCE

The gradiometer-based prosthetic heart valve evaluation system is an electromagnetic
(EM) diagnostic system that subjects the patient to a low-level radio frequency (RF)
magnetic ﬁeld during the examination. To ensure patient and medical personnel safety,
the operating conditions of the system had been assessed against United States and
international regulatory standards and safety guidelines. This chapter describes the
assessment results and demonstrates compliance of the examination system with the
international health standards. The results also show that exposure to the system’s EM
ﬁeld will not result in any adverse health effects or discomfort to the patient and the

medical staff.

11.1 Health Effects of Radio Frequency EM Field Exposure
In a RF non-ionizing radiation system such as the gradiometer-based examination system,
three principal effects from electric and magnetic ﬁeld exposure may potentially lead to
adverse effects in humans: (1) electrostimulation, (2) RF heating of the biological
tissues, and (3) induction heating of metallic implants. Electrostimulation is the
induction of a nerve pulse (an action potential) in excitable tissues such as nerves and
muscles by an externally applied electrical stimulus. Depending on the amount of
excitation, the biological effects of electrostimulation may range from minor patient
annoyance to detectable impairment of normal physiological functions (e. g., the cardiac

rhythm).

235

Radio frequency heating, or induction heating, includes the heating of the biological
tissue and metallic implant due to exposure to a time-varying EM ﬁeld. Such exposure
induces electric ﬁelds and currents in the tissue and the metallic parts of the implant
according to Faraday ’5 Law. The magnitudes of the induced ﬁeld and current density are

proportional to the frequency of the ﬁeld, rate of change of the applied magnetic ﬂux

density, 51;, and conductivity, 0, of the medium. The induced electric ﬁeld, E, and
t

current density, J, are related by Ohm ’5 Law:

J = a E (11.1)
The ﬂow of the induced currents causes heating of the affected region. The amount of
temperature rise may be obtained by either an in situ measurement (if the location of
interest is accessible) or solving the bioheat transfer equation (10.1) [156] for the region

of interest as demonstrated in the Chapter 10.

In general, exposure to EM ﬁelds may result in an increase in body temperature due to
absorption of energy. The transmitted power and signal frequency are two of the most
important factors affecting absorption as they relate to the penetration depth and
absorption of the RF energy below the skin surface [163]. It is, therefore, important to
ensure patient safety by characterizing, monitoring and, if necessary, limiting the
electromagnetic energy absorption when using a medical diagnostic device such as the

gradiometer-base examination system.

236

11.2 Nature and Duration of EM Field Exposure in the Gradiometer-Based

Examination System

During the heart valve examination, the patient assumes a supine position with the long
axis of the body resting on the plane centered between the excitation coils of a Helmholtz
system. The magnetic ﬁeld is uniform and at its maximum value within an
approximately 40 cm-diameter circular region at the center of this plane. Outside of this
region, the ﬁeld decreases dramatically with increasing distance as shown in Figure 11.1.
This ﬁeld distribution exposes a patient to a non-uniform magnetic ﬁeld with a maximum
exposure per unit volume at the center (around the chest area). For a worst-case
approximation, an uniform magnetic ﬁeld distribution is assumed throughout the entire
Helmholtz coil volume (the compliance volume) with a ﬁeld strength equal to the
maximum value obtained at center of the coils. Note that the actual patient exposure to

the magnetic ﬁeld is signiﬁcantly less that this conservatively estimated value.

The gradiometer-based examination system produces a maximum peak-to—peak magnetic
ﬂux density of approximately 1.5 gauss at an excitation frequency of approximately 65
kHz. During the examination, the excitation ﬁeld is ﬁrst applied when the gradiometer is
placed approximately 10 cm away from the heart valve. At this location, a reference
signal is recorded for approximately 30 seconds. The ﬁeld is then removed and the

gradiometer is advanced into the mitral annulus and placed in front of the outlet strut.

237

Magnetic ﬂux density, norm [T]

 

1.5 r . f

 

1.45 r

r 2-1
W A

E"
N

Magnetic ﬂux density, norm [T]
2* 2‘
H N
U1 U!

E"
...;

 

1.05

 

 

 

-0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8

Figure 11.1 Amplitude of the magnetic ﬂux density is the largest at the center of the
plane positioned mid-way between the two excitation coils.

After the gradiometer is properly-positioned, the excitation ﬁeld is re-applied. The
gradiometer signal is then recorded for 60-90 seconds before the ﬁeld is once again
removed. Depending on the ﬁdelity of the recorded signal, the gradiometer may be
repositioned (and its signal recorded) up to ﬁve times during the examination. Each

gradiometer repositioning may take up to 5 minutes.

11.3 Specific Absorption Rate
The speciﬁc absorption rate, SAR (W/kg), is the radio frequency power absorbed per unit

mass of an object:

238

d (1W (1 dW
SAR=— —— =— —
dt(dm) dtLadV] (112)

where W (J) is the energy absorbed, m (kg) is the mass, V (m3) is the volume, and ,0 (kg/

m3) is the density of the object. When applying to human, the SAR describes the amount

of energy absorbed per unit mass of the affected tissue, organ, or body region.

The SAR can also be related to the electric ﬁeld at a point by

2
SAkzgl—EW—SL (11.3)
p

where O'iS the conductivity (S/m) and ERMS is the RMS electric ﬁeld strength (V/m) of

the tissue.

Alternatively, the SAR can be related to the increase in temperature at a point by

CAT
SAR=-— = 11.4
At t 0 ( )

where AT is the change in temperature (°C), At is the duration of exposure (5), and c is
the speciﬁc heart capacity (J/(kg °C)). Note that (11.4) is valid only under adiabatic
conditions (non—therrnodynamic circumstances) with no heat loss by thermal diffusion,
radiation, convection, or other natural bio-thermoregulation processes (blood ﬂow,

sweating, etc.).

239

11.4 Specific Absorption Rate Calculation
The SAR of the RF energy from the gradiometer-based examination system by the
various body parts had been evaluated to assure patient safety. The induction heating
ﬁnite element analysis (FEA) results from Chapter 9 were used to calculate the
myocardial, whole body and local SARs. Additionally, the whole body average (WBA)
SAR had been estimated using two different approaches. In the ﬁrst approach, the
induced electric ﬁeld due to the RF exposure was obtained from the ﬁnite element model
and applied to (11.3). In the second approach, the WBA SAR was calculated directly
from the FEA results. These SARs were then compared against the regulatory standards

and guidelines to assess compliance.

11.4.1 SAR Inference from Derived Electric Field Strength
Equation (11.3) may be used for estimating the SAR in cases where other, perhaps more
direct, means of obtaining this value do not exist [166]. When applying this equation to

estimate the WBA SAR, a homogenous body conductivity of 0.2 S/m [172] and density

of 1000 kg/m3 may be used. The IEEE standard recommends averaging over a distance

of 5 mm to obtain the magnitude of the RMS electric ﬁeld strength, [ERMSI- For

compliance determination, the SAR value is averaged over 6 minutes [166], [172] of

exposure.

240

To obtain the 5-mm average of the electric ﬁeld in the vicinity of the prosthetic heart

valve, the magnitude of the ﬁeld is averaged over 6 minutes1 along a radial path, LE, near

the outer boundary of the myocardium (from r = 0.0549 to 0.0599 m, at z = 0 m) as
shown in Figure 11.2. The magnitude of the electric ﬁeld along this path is shown in
Figure 11.3. The data in this ﬁgure shows that the induced electric ﬁeld is stronger
towards the exterior of the body muscle (perimeter of the body). This observation may
be extended to the full—body exposure case where the near-surface regions of the body
experience most exposure to the EM energy. As a result, more heat is generated in the
perimeter than the core regions of the body. In general, the RF energy penetration below
the skin surface is a function of frequency and the characteristics of the skin and

underlying tissues.

From the FEA results, the average value of the electric ﬁeld along LE was Em,g = 1.12

V/m, and the corresponding RMS value was ERMS = 0.792 V/m. Substituting ERMS into
(11.3) yields

0.2(0.792)2
1000

=1.24><10'4 (W/kg) (11.5)

SARWBA =

The IEEE standard [163] accepts the WBA SAR estimate obtained using equation (11.5)

for the purpose of compliance assessment. To obtain region-speciﬁc SARs, the ﬁnite

 

1 For the SAR calculations, the 6-minute time-average window is selected from t = 540-
900 s from the ﬁnite element analysis results.

241

element analysis results from Chapter 10 are used. These SAR calculations are described

in the following sections.

i Myocardium
' l
I Atrium

0.03 m /

 

 

0.01 m J Suture

............ l 1
0.06 m 0.04 m \
LE

1 .\

\ Heart Valve
‘f Flange Ring
\

0.03 m Body Muscle Ventricle
1 (Generic)

r—0.06 m——-l

Figure 11.2 The average value of the electric ﬁeld intensity, E, along path L5 is used for
SAR calculation.

 

 

 

 

 

 

 

 

 

 

11.4.2 Region-Speciﬁc SAR Calculations

11.4.2.1 SAR for the myocardium

To calculate the 6-minute average of the total power absorbed by the myocardium, the
inductive heating power density, Q, is integrated over the volume of the myocardium and
averaged over time.

Table 11.1 lists the FE model parameters that are pertinent to the SAR calculations.

242

Magnitude of Electric Field (z=0, r=0.0549-0.0599 m)

 

1.16-

1.15:

1.14"

1.131

IEI (V/m)
:3

1.111

1.1-

1.09 r

1.08 -

 

 

 

4

 

1 .07 1 ‘ ‘ : ‘ ‘
0.0545 0.0555 0.0565 0.0575
Racial Distance, r (m)

0.0585 0.0595

Figure 11.3 The magnitude of the electric ﬁeld, [E , along path LE (2:0, r= 0.0549-0.0599
m) for the ﬁnite element model in Figure 11.2.

 

From the FEA results, the total power absorbed is Pmy = Q(myocardium) = 2.9654x10_6
W. The volume of the myocardium is approximately me = 9.31x10'5 m3, with a

density pmy = 1070 kg/m3 (Table 10.2). Using (11.2), the SAR for the myocardium is

SAR = PM
my pmyme
_ 2.9654x10‘6

— (1070)(9.31x10“5)
= 2.9768x10‘5 (W/kg) (11.6)

 

243

Table 11.1 Selected results from the RF inductive heating FEA from Chapter 9

 

Quantity Myocardium Body muscle

 

3

Volume, V 9.3131x10-5 m 1.239x10-3 m3

 

Total power absorbed, Q 2.9654x10'6 W 2.63O7x10.4 W

 

Total current density, J 0.144898x10-3 A/m2 2,235x10'3 A/m2

 

 

 

 

 

11.4.2.2 Whole Body Average SAR

To obtain the whole body average (WBA) SAR, the total power dissipated in the “body
muscle” region of the ﬁnite element model was calculated. From

Table 11.1, the total power is Pbm = Q(body muscle) = 2.6307x10'4 W, and the volume

of the region was Vb", = 0.001239 m3. The power density dissipated in, or absorbed by,

the “body muscle” region is then given by:

Pbm
me

_ 2.6307x10‘4

0.001239 3
= 0.21233 (W/ m ) (11.7)

 

Qbody =

 

For the worst-case scenario where the entire body is exposed to this power density, with

an average body density [172] of pbody = Wbody / Vbody = 1025 kg/ m3, the worst-case

WBA SAR is given by:

244

Pb l
SARWBA = —me 'Vbody ° Wb d
m 0 y

 

_ Qbody
pbody

_ 0.21233

' 1025

= 2.0715x10‘4 (W/ kg) (11.8)

 

 

where Wbody is the full body weight of the patient. The WBA SARs from equations

(11.5) and (11.8) agree in their orders of magnitude, with the later value approximately
1.6 times the former. These results show that equation (11.5), when used with the whole
body averaged conductivity and density values, provides a convenient estimate of the

SAR and is sufﬁcient for compliancy assessment.

11.4.2.3 Local SAR

The IEEE and IEC standards specify that the power dissipated in a 10-gram mass
(deﬁned as a tissue volume of 10 cm3 in the shape of a cube) is to be used for calculating
the local SAR. From (11.7), the power density dissipated in the “body muscle” region is

Qbody = 0.21233 W/m3. If 100,01: 10 cm3 = 10x10'6 m3, mm, = 10 g = 0.01 kg, the

local SAR is then given by

1
Wlocal

= (0.21233)(10x10'6 )OLOI

= 2.1233x10’4 (W/ kg) (11.9)

 

SARlocal : Qbody ' Vlocal °

245

The myocardial, whole body, and local SARs from (11.6), (11.8), and (11.9),
respectively, are used to access the safety of the examination system in the following

sections.

11.5 Applicable Regulatory Standards and Guidelines

The excitation energy level and patient exposure conditions of the gradiometer-based

examination system had been assessed against the IEEE1 ([163]-[166]) and IEC2 ([167]-

1711) standards, as well as the ICNIRP3 ([172]) and FDA4 ([173]-[175]) guidelines, to

ensure patient and medical personnel safety.

11.5.1 The IEEE Standard (IEEE Std C95.1-2005)

The IEEE Std C95.1-2005 [163] standard recognizes that exposure to different frequency
RF energies result in differing impacts. Therefore, it provides exposure restrictions
called basic restrictions (BRs) to address three distinct frequency bands. The BRs are
expressed in terms of the in situ electric ﬁeld (3 kHz to 5 MHz), speciﬁc absorption rate
(100 kHz to 3 GHz), or the incident power density (3 GHz to 300 GHz). In cases where

values of the physical quantities are difﬁcult or impossible to ascertain, the standard also

 

1 IEEE: Institute of Electrical and Electronics Engineers, Incorporated (www.ieee.org).

2 IEC: Commission Electrotechnique Internationale (International Electrotechnical
Commission) (www.iec.org).

3 ICNIRP: International Commission on Non-Ionizing Radiation Protection
(www.icnirp.org).

4 FDA: United States Food and Drug Administration (www.fda.gov).

246

recommends maximum permissible exposurel (MPE) values. The MPEs are derived or
estimated from the basic restrictions. When comparing with the respective MPEs and
BRs, compliance with the MPEs automatically ensures compliance with the underlying

BRs.

In the following sections, four criteria are used to demonstrate compliance of the
gradiometer-based examination system with the IEEE standard: basic restrictions for the
in situ electric ﬁeld, MPE for the magnetic ﬁeld, peak in situ electric ﬁeld, and peak

external magnetic ﬁeld.

11.5.1.1 Basic restrictions for the in situ electric ﬁeld

Table 11.2 lists the basic restrictions for the in situ electric ﬁeld for various body regions

according to IEEE Std C95.1-2006 [166]. In this table, E0 is the rheobase in situ electric
ﬁeld and fe is a strength-frequency parameter. The maximum allowed induced in situ

electric ﬁeld, E], for each body region is determined from this table according to:

(11.10)
Ei = E0[—f—] f->-fe

 

1 Maximum Permissible Exposure (MPE): The highest level of exposure, with an
additional margin of safety, to which a person may experience without incurring any
established adverse health effects. The MPEs are sometimes referred to as the

reference levels, derived limits, or investigation levels in literatures.

247

Table 11.2 Basic restrictions on exposure for the various regions of the body (after [163],
Table 1), with limits for the gradiometer-based examination system.

 

 

 

 

 

 

 

Limits for the gradiometer-based
IEEE Std (3951-2006 examination system f= 65 kHz)
Action Persons in Action Persons in
l 1 controlled 1 11 controlled
evel environments eve Environ.
EXPosed fe E0 (RMS) £0 (RMS) .1; E1(RMS) E1(RMS)
tissue (Hz) (V/m) (V/m) fe (V/m) (V/m)
Brain 20 5,89x10‘3 1,77x10'2 3250 19.1425 57.525
Heart 167 0.943 0.943 389.222 367.0359 367.0359
Extremities 3350 2.10 2.10 19.4029 40.7463 40.7463
9"” 3350 0.701 2.10 19.4029 13.6015 40.7463
tissues

 

 

 

 

 

 

 

 

 

Substituting the operating frequency f = 65 kHz into (11.10) and applying values from the

left half of Table 11.2 yield the limits for the gradiometer-based examination system.

These limits are expressed as Ei’s on the right half of Table 11.2.

In the gradiometer-based examination system, the excitation magnetic ﬁeld is the
strongest at the center of the Helmholtz coil. As a result, the induced currents are also the
largest when objects or body parts are submerged in this uniform ﬁeld. In the worst-case
scenario, if the region of uniform magnetic ﬁeld was extended to the entire compliance
volume, the maximum electric ﬁeld induction volume would also expand accordingly.
For a patient positioned between the excitation coils, the majority of the induced currents

are conﬁned near the body surface as described in Section 11.4.1. Therefore, the RMS

electric ﬁeld value, ERMS = 0.792 V/m, obtained from Section 11.4.1 for the body
muscle is suitable for use as an upper bound for the induced in situ electric ﬁeld for all

exposed regions of the body. Comparing ERMS with the E values in the right half of is

248

show that the gradiometer-based examination system is compliant with the basic

restrictions for induced in situ electric ﬁeld.

11.5.1.2 MPE for the magnetic ﬁeld
The peak—to-peak magnetic ﬂux density at the center of the gradiometer-based
examination system excitation coils is approximately 0.15 mT (1.5 gauss). The RMS

value of this ﬁeld is

BRMS =1.5x10_4(%][—1-2—] (T)
= 0.5303><10‘4 (T) (11.11)

The corresponding magnetic ﬁeld strength, HRMSs is

1
HRMS = Z BRMS
_ 0.5303x10— 4

47rx10_7
=42.202 (A/m) (1 112)

 

If the uniform magnetic ﬁeld at the center of the Helmholtz coil was expanded to the

entire coil volume, the head, torso and limbs of the patient will also be exposed to BRMS-
Table 11.3 lists the RMS MPEs limits for BRMS and the magnetic ﬁeld, HRMS) for

exposure to the head, torso, and limbs (arms and legs) from the IEEE standard]. A

comparison of the results from (11.11) and (11.12) with those from Table 11.3 show that

 

1 IEEE Std C95.1-2006 Section 4.1: Basic restrictions (BRs) and maximum permissible
exposures (MPEs) for frequencies between 3 kHz and 5 MHz

249

the BRMS and HRMS levels for the gradiometer-based examination system are within the

thresholds recommended by the IEEE standard.

Table 11.3 MPE for the head, torso and limbs: f = 3.35 kHz to 5 MHz (after [166], Tables

 

 

 

 

2 and 3)
Action levell Persons in controlled environments
Body part B
RMS (mT) HRMs (A/m) BRMS (mT) HRMS (A/m)
Head and torso 0.205 163 0.615 490
Limbs 1.13 900 1.13 900

 

 

 

 

 

 

 

11.5.1.3 Peak in situ electric ﬁeld

The IEEE standard states that the temporal peak value of the in situ electric ﬁeld must not

exceed that obtained by multiplying the RMS values in Table 11.2 by J2 . From the

induction heating FEA results, the peak in situ electric ﬁeld for the model in Figure 11.3

is Epeak = 1.169422 V/m. From Table 11.2, the minimal RMS Ei value is the exposure
limit for “Other tissues” (E, = 13.6015 V/m), and the corresponding peak value is

E i, peak = 72036015) = 19.2354 V/m. Comparison of these electric ﬁeld values show
that the peak in situ induced electric ﬁeld, Epeak, is less than the lowest peak restriction

level, Ei,peakv from Table 11.2. These results show that the gradiometer-based

1 IEEE Std C95.1-2006 Section 4.1: Basic restrictions (BRs) and maximum permissible
exposures (MPEs) for frequencies between 3 kHz and 5 MHz

250

examination system satisﬁes the peak in situ electric ﬁeld restriction recommended by the

IEEE standard.

11.5.1.4 Peak external magnetic ﬁeld

The peak-to-peak magnetic ﬂux density for the gradiometer-based examination system is

approximately 1.5x10-4 T (1.5 gauss). This excitation magnetic ﬂux density may be

expressed as B = 30 sin(27: f t) (T), where Bo = 0.75x10-4 T and f = 65 kHz. The time

derivative of this external ﬁeld is B = if- = 27! f BO cos(2ﬂ' f t). The peak value of B is
t

'8' = Zﬂ'f BO (T)

=O.471f (mT) (11.13)

 

. 1 . 1 . - .
and the phase duration , tp, rs t p = 57. For compliance assessment, Bl rs to be

compared with the allowable peak limit on B, denoted as B p , given by
B, =J'2MPEB(2nf) (11.14)

The values for MPEB are obtained from the MPEs in Table 11.3. Table 11.4 lists the B p

values for the head, torso and limbs. Comparisons of the values from (11.13) and Table
11.4 show that the peak external magnetic ﬁeld value in the gradiometer-based

examination system falls well below the limits recommended by the IEEE standard.

 

1 . . . . .
For a periodic waveform, the phase duration 18 deﬁned as the time between zero
crossings

251

Table 11.4 Maximum permissible values, B p , of the time derivative of the external

magnetic ﬁeld for the head, torso and limbs (f = 65 kHz).

 

 

 

 

 

 

 

 

Body Action level (mT) Persons in controlled Gradiometer-based
part environments (mT) system values
Head and
0.2899 (27rf) 0.8698 (272‘f)
torso 0.075 (2711)
Limbs 1.5981 (an) 1.5981 (erf)

 

The assessment of the gradiometer-based examination system with regard to the basic
restrictions for the in situ electric ﬁeld, MPE for the magnetic ﬁeld, peak in situ electric
ﬁeld and the peak external magnetic ﬁeld demonstrate that the system is compliant with
the IEEE Std C95.1-2005 standard, and that exposure to RF energy levels generated by

the system is safe for patients and medical personnel.

11.5.2 The IEC Standard (IEC 60601-2-33)
The IEC standards ([l67]-[171]) establish the general requirements for basic safety and
essential performance of medical electrical equipments, with IEC 60601-2-33 ([167],
[168]) providing particular requirements for the safety of magnetic resonance (MR)
medical diagnostic equipment. Although originally intended for medical MR systems,
the standard nevertheless provides SAR limits that are applicable to the gradiometer-
based examination system. These limits intend to provide protection against:

1. Excessive low frequency ﬁeld variations produced by the gradient system;

2. Excessive radio frequency energy;
A third objective, protection against exposure to static magnetic ﬁelds, does not apply to
the gradiometer-based examination system as it does not utilize any static magnetic

ﬁelds.

252

11.5.2.1 Protection against excessive low frequency ﬁeld variations

The IEC standard establishes limits for a RF system’s gradient output. The objective of
the limits is to prevent cardiac or peripheral nerve stimulation (PNS) due to exposure to
excessive low frequency variations. The limits are expressed in terms of either the

induced electric ﬁeld, E, in the patient, or the time rate of change of the magnetic ﬂux
density, dB/dt (T/s). The limits are functions of the effective stimulus duration, 15,617
(ms), deﬁned as the ratio of the peak-to-peak ﬁeld variation to the maximum value of the

time derivative of the gradient [168]. From equation (11.13), tsﬂﬁr may be written as:

 

t _ 230 _230
“’1 (a) 181
dt max
1
=_ (11.15)
If

For a sinusoidal excitation with frequency f = 65 kHz, the effective stimulus duration is

1W}: 4.897x10'3 ms.

11.5.2.1.1 Limits for prevention of cardiac stimulation

To protect against cardiac stimulation, the induced electric ﬁeld due to the gradient

output of the system must satisfy:

E<

 

t
{1_exp[_ sir J} (11.16)

253

Substituting (11.15) into (11.16) yields

 

 

 

E < 2

+ [ 4.897x10‘3 J}

l—exp —
3
2
<

1.631x10"3

E < 1226.221 (V/m) (11,17)

From the inductive heating ﬁnite element analysis results discussed in Section 11.4.1, the

average induced electric ﬁeld in the patient by the gradiometer-based examination system

is Em,g = 1.12 V/m. This value is well below the IEC limit given by (11.17).

Alternatively, the IEC standard [168] states that for a whole body gradient system such as
the gradiometer-based examination system, equation (11.16) may be replaced by

dB 20

_ <
dt . -
{1_ exp[_ 13:); J} (11.18)

Applying the results from (11.15) to equation (11.18) gives

 

g5 20

<
d‘ { [ 4.897x10—3]}
l-exp — 3

20

1.631x10‘3

d—B- <12262.214 (T/s)

dt (11.19)

 

 

 

<

254

For the gradiometer-based examination system, substituting f = 65 kHz into (11.13) gives

dB

131 = d,

= 61.261 T/s. This value is below the limit speciﬁed by (11.19).

 

 

Comparisons of the induced electric ﬁeld, E, and the time rate of change of the magnetic
ﬂux density, dB/dt, against the limits given by (11.17) and (11.19), respectively,
demonstrate that the gradient output of the gradiometer-based examination system are

within the safety limits set forth by the IEC for the prevention of cardiac stimulation.

11.5.2.1.2 Limits for prevention of peripheral nerve stimulation

To prevent peripheral nerve stimulation, the IEC establishes output limits for the two

types of gradient systems: the whole body and special purpose systems. For each
system, the limits for two operating modes are speciﬁed: the normal operating mode1

(L01) and the ﬁrst level controlled operating mode2 (L12):

L12 = 1.0 rb 1+ 036
ts,eﬁ'

 

(11.20)

 

L01=0.8r,, 1+ 0'36
ts,ejf

 

1 . . . .
In this operational mode, none of the system outputs may cause physrologrcal stress to

patients.

2 The operational mode in which the system must be controlled under medical

supervision as one or more output levels may cause physiological stress to patients

255

In (11.20), rb (T/s) is the rheobase given in Table 11.5. Note that L01, L12 and rb may be

expressed in terms of the induced electric ﬁeld, E (V/m), or the time rate of change of the

magnetic ﬁeld, dB/dt (T/s). From Section 11.5.2.1, the tsﬂﬂr for the gradiometer-based

examrnatron system was found to be tseﬁ: 4.897x10 ms. Substrtutrng this value in

(11.20) yields the L01 and L12 limits in Table 11.5.

From the results in Section 11.4.1, the average induced electric ﬁeld in the patient by the
gradiometer-based examination system is Em,g = 1.12 V/m, and the maximum time rate

of change of the magnetic ﬂux density is 5:; = 61.261 T/s. Comparison of 5an and (‘11—?

with the limits in Table 11.5 show that the gradient output of the gradiometer-based
examination system satisﬁes the safety limits set forth by the IEC for the prevention of

peripheral nerve stimulation.

11.5.2.2 Protection against excessive RF energy

To protect against the effects of excessive RF energy, the IEC standard provides limits
for core and spatially localized temperature elevations and SARs due to exposure to RF
energy. The temperature limits are shown in Table 11.6 and the SAR limits are shown in

Table 11.7.

256

Table 11.5 Gradient output limits for the gradiometer-based examination system at the
operating frequency of f = 65 kHz. (The rb values are obtained from [168].)

 

 

 

 

E (V/m) dB/dt (T/s)
Type of gradient system
rb L12 L01 rb L12 L01
Whole body 2.2 163.929 131.143 20 1490.265 1192.212
Special purpose 2.2 163.929 131.143 N/A - -

 

 

 

 

 

 

 

 

 

From the ﬁnite element analysis results presented in Section 10.3.2, the temperature
elevation of the prosthetic heart valve and the surrounding regions is no more than 0.003
°C after 15 minutes of continuous exposure to the RF energy from the gradiometer-based
examination system. The results also show that there is no RF exposure-related
temperature rise in any other regions of the body. Therefore, the core and spatially
localized temperature elevations are well below the limits prescribed by the IEC as

shown in Table 11.6.

The results from Sections 11.4.1 and 11.4.2 show the whole body SARs from equations

(11.5) and (11.8) are 1.24x10’4 and 2.0715x10'4 W/kg, respectively. The partial body

SAR is 2.1233x10.4 W/kg (11.9). These values satisfy the IEC limits listed in Table

11.7.

To assess the head and local SARs of the examination system, the limits in Table 11.7 are
expressed in terms of temperature elevations [173] in the affected region. Speciﬁcally, in
the normal mode, a head SAR of 3.2 W/kg corresponds to a 1-degree rise in the cranial

temperature [175].

257

Table 11.6 IEC temperature limits for RF exposure (after [169], Table 104)

 

 

 

operating Core Spatially localized temperature limits
Mod Temperature
e Rise, °C Head, °C Torso, °C Extremities, °C
Normal 0.5 38 39 40

 

First level
controlled
Second level

1 38 39 40

 

>1 >38 >39 >40

 

 

 

 

 

 

1 controlled

 

Table 11.7 SAR limits (after [167], Table 105)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Avera in .
. g g 6 mln
time
WHOLE PARTIAL
BODY BODY “3?": LOCAL SAR
SAR SAR
Body region Whole Exposed . .
_9 body body part Head Head Trunk Extrermtles
O ratin
pe jg (WIkg) (WIkg) (WIkg) (W/kg) (Mrs) (Ms)
mode
NORMAL 2 2.10a 3.2 10b 10 20
FIRST LEVEL a b
CONTROLLED 4 4-10 3-2 10 10 20
SECOND a b
LEVEL >4 >(4-1()) >32 >10 >10 >20
CONTROLLED
SHORT TERM The SAR limits over any 10 5 period shall not exceed three times the
SAR stated values
a
The limit scales dynamically with the ratio “exposed patient mass / patient mass”:
Normal Operating Mode:
Partial Body SAR = 10 W/kg — (8 W/kg * exposed patient mass / patient mass)
First Level Controlled Operating Mode:
Partial Body SAR = 10 W/kg - (6 W/kg * exposed patient mass / patient mass)
In cases where the orbit is in the ﬁeld of a small local RF transmit coil, care should be taken to
ensure that the temperature rise is limited to 1 °C.

 

In Table 11.7, a torso SAR Of 10 W/kg, averaged over any 10 grams of trunk tissue,

corresponds to a temperature rise Of 2 degrees [173]-[175]. For extremities, a SAR limit

 

l The output limit beyond which a formal human studies protocol and internal review
board approval are required.

258

of 10 W/kg, averaged over any 10 grams of tissue in the exposed region, corresponds to a
temperature rise of 3 degrees [175]. Since the results in Section 10.3.2 show that the
temperature rise in the regions outside of the myocardium is negligible, it can be
concluded that the gradiometer-based examination system satisﬁes the head and local

SAR limits for all three operation modes in Table 11.7.

The results from Sections 11.5.2.1 and 11.5.2.2 show that the gradiometer-based

examination system is compliant with the IEC standard for RF medical equipment safety.

11.5.3 The ICNIRP Guidelines

Similar to the IEEE standard [166], the ICNIRP guidelines [172] also provide basic
restrictions for exposure to EM ﬁelds. However, in the ICNIRP guidelines, the physical
quantities used to specify the basic restrictions are the current density, SAR, and power
density. Application of these quantities is based on the frequency range of the EM ﬁeld.
Speciﬁcally, between 1 Hz and 10 MHz, induced current density basic restrictions are
provided for preventing nervous system stimulation and adverse effects. From 100 kHz
to 10 GHz, SAR basic restrictions are stated to prevent whole body heat stress and
excessive localized tissue heating. Within the overlapped frequency range Of 100 kHz to
10 MHz, both the current density and SAR restrictions must be satisﬁed. Finally,
between 10 and 300 GHz, power density basic restrictions are provided to prevent

excessive local tissue heating at or near the surface Of the body.

259

The ICNIRP basic restrictions applicable to the gradiometer-based examination system
are those in the 1-100 kHz frequency range. These restrictions specify the maximum
allowable induced current in the exposed region. For an Operating frequency of f = 65

kHz, the basic restrictions are listed in Table 11.8.

Table 11.8 ICNIRP basic restrictions for a time-varying EM ﬁeld at f = 65 kHz

 

Exposure characteristics Current density for head and trunk (mA m-2)(RMS)

 

Occupational 650
General public 130

 

 

 

 

 

From

Table 11.1, the RMS current density in the myocardium is me,RMS = 0.1448978x10.3 x
0.7071 = 0.1236 mA/mz, and that in the body muscle region is mepMs = 2.235x10'3 x
0.707] = 1.5804 mA/mz. Comparing these values with the levels in Table 11.8, the
highest RMS current density induced in the patient body (1.5804 mA/mz) is below the

2 . . . .
130 mA/m basrc restrictlon value for general publlc exposure. Therefore, these results

demonstrate that the gradiometer-based examination system is compliant with the

ICNIRP standards.

11.5.4 The FDA Guidelines

The FDA guidelines ([173]-[175]) align closely with the IEC standard IEC 60601-2-33
([167]-[l71]). They share the same operating mode deﬁnitions, as well as limits for
whole body‘and localized heating, SARs, gradient output, and induced electric ﬁeld

levels [169]. The FDA guidelines, however, imposes additional limits on the RF heating

260

Of the implanted medical devices. Compliance Of the gradiometer-based examination
system with the IEC standard has already been demonstrated in Section 1 1.5 .2. The

remaining FDA requirements are examined here for safety assessment.

11.5.4.1 FDA Limits on Implant Heating

Implant heating can occur in two ways. First, as eddy currents are induced in the human
body by a RF magnetic ﬁeld, a conductor may concentrates these currents and produce
intense localized heating at any exposed “corners” or “points” such as the tip of a wire.
Second, induced eddy currents may result in inductive heating Of the conductor itself.
The FDA guidelines require that the temperature rise of the medical device, without

convective heat transfer, shall not exceed 3 degrees over 15 minutes of exposure [175].

In the gradiometer-based examination system, the ﬂange ring, and the inlet and outlet
struts of the prosthetic heart valve form multiple metallic closed loops. During a heart
valve examination, the heart valve is aligned such that the ﬂange ring is parallel to the
applied magnetic ﬁeld. In this orientation, there is no current induced in the inlet strut
loop, and only minimal eddy currents are induced in the portion of the ﬂange that
intercepts the applied magnetic ﬁeld. In the worst-case scenario when the prosthetic
heart valve is oriented such that the ﬂange ring is orthogonal to the applied magnetic
ﬁeld, a maximum amount of eddy currents will be induced in the ﬂange, resulting in a
maximum temperature rise of the heart valve. This scenario has been investigated in
Chapter 9 for various cooling scenarios. The results in Section 10.3.2 show that, for 15
rrrinutes of exposure with normal temperature regulation, the temperature Of the

prosthetic heart valve does not rise beyond 37.003 °C (a temperature rise of 0.003

261

degrees). In the case when there is no convection cooling (no blood ﬂow), the results in
Section 10.3.3 show that the maximum temperature rise of the prosthetic heart valve is
0.0652 °C after 15 minutes Of exposure. These temperature elevations are less than the 3

°C limit set forth in the FDA guidelines.

The inductive implant heating FEA results showing compliance with the IEC stande
afﬁrms that the gradiometer-based examination system is compliant with the FDA
guidelines for RF medical equipment safety. It must be mentioned that the system cannot
be employed for examining patients with cardiac pacemakers and other implanted

devices containing metal.

11.6 Compliancy Statement
The results presented in this chapter demonstrate that the gradiometer-based examination
system is compliant with the IEEE Std C95.1-2005 [166] and IEC 60601-2-33 [168]
standards, as well as the ICNIRP [172] and FDA [175] guidelines. The RF energy

emission level from the examination system is safe for patients and medical personnel.

262

CHAPTER 12. CONCLUSIONS AND FUTURE WORK

A novel minimally invasive gradiometer-based prosthetic heart valve examination system
has been developed. The system is capable of detecting single-leg separation (SLS)
failures in BSCC heart valves with a reasonable degree of conﬁdence. A clinical-scale
prototype of the system has been constructed, and two gradiometer designs have been
built for use with different patient physical conditions and physiques. An in vitro test
system that can simulate the movement Of the heart valve during the cardiac cycle has
also been constructed. Double-blind in vitro studies with the examination system
prototype on 34 BSCC heart valves with various outlet strut conditions have repeatedly

achieved 100 % accuracy in detecting the SLS cases.

During the development of the examination system, ﬁnite element (FE) techniques were
used to characterize and Optimize various system components such as the magnetic ﬁeld
distribution generated by the excitation coils, design of the gradiometer, as well as to
demonstrate the detection feasibility of the magnetic ﬁeld perturbations in the vicinity of
the outlet strut. Based on these results, the components Of the examination system were

designed.

Since the excitation ﬁeld of the gradiometer-based method induces currents in the ﬂange
and struts of the BSCC heart valve, a FE analysis has also been performed to investigate
the potential temperature rise of the heart valve due to inductive heating effects. Analysis

results show that inductive heating Of the prosthetic heart valve and the neighboring

263

tissues from exposure to the excitation ﬁeld of the gradiometer-based examination system

is negligible, and that the system is safe for patient use.

The FE analysis results also show that the low-level and low frequency magnetic ﬁeld (<

l.5><10-4 T at z 65 kHz) employed by the examination system is compliant with all key

international health and safety standards established by organizations such as the IEEE,
IEC, ICNIRP, and the US FDA. The RF energy emission level from the system is safe

for patients and medical personnel, and will not cause any adverse health effects.

To quantify the reliability of the examination approach when under the inﬂuence Of
system parameter variabilities, a technique for developing the probability of detection
(POD) model for the examination system has been presented. Application of this model
shows that the examination system is highly robust against parameter variability such as
the deviation in the gradiometer initial position from the speciﬁed target location during
an examination. The procedures presented may be adapted for developing PODs for other

sources Of system variabilities.

The clinical-scale system prototype is portable. It is suitable for transportation over long
distances as well as relocation from one operating room to another. A custom surgery
table has also been prepared for use with the examination system. The top Of this table
has been replaced with one made from high-strength non-metallic materials so that it will
not interfere with the excitation magnetic ﬁeld. The examination system prototype, with

the supporting equipment, is ready for animal studies.

264

In conclusion, a minimally invasive, highly accurate gradiometer-based examination
system for detecting SLS failures in BSCC prosthetic heart valve has been developed.

The system is highly reliable and safe for patients and medical personnel.

12.1 Future Work
The gradiometer-based examination system is ready for animal studies. The system
needs to be evaluated in a suitable animal clinic or hospital. An animal use protocol must
be submitted for approval and BSCC heart valves with various outlet strut conditions
need to be implanted into either ovine or bovine models. The cardiac surgeon
responsible for catheterizing the animal subject and conducting the examination should
be familiar with the recommended animal study examination procedures provided in
Appendix B of this thesis. It is crucial that the surgeon understands the principal Of the
examination technique, and be able to position the gradiometer to the prescribed target
location in front of the outlet strut. If feasible, double-blind tests should be performed for

the animal studies.

Upon the successful completion of the animal studies, human trials should commence
immediately. The gradiometer-based examination system may be relocated to a national
test center and BSCC heart valve patients are to be transported to the center for
examinations. A successful human trial will provide BSCC patients with an early

detection mechanism for SLS failures.

265

APPENDIX

266

APPENDIX A. BIOLOGY OF THE HEART AND HEART VALVE

DISORDERS

A.1 The Heart

The cardiovascular system consists Of the heart and the circulatory system. The heart
works as a pump that pushes blood to the organs, tissues, and cells of the body through a
complex network of arteries, arterioles, and capillaries. Blood delivers oxygen and
nutrients to every cell and removes the carbon dioxide and waste products produces by
the cells. The blood is then returned to the heart through venules and veins. If all the
blood vessels in the human body were laid end-tO-end, they would extend for almost
60,000 miles (more than 96,500 kilometers), which is twice the circumference of the

earth.

The heart weighs between 7 and 15 ounces and is approximately the size of two clenched
ﬁsts. On average, the heart beats approximately 100,000 times a day, pumping more than
2,000 gallons of blood. This amounts to nearly 38 million heart beats a year, circulating
more than 730000 gallons of blood. In an average lifetime of a person, the heart pumps

about 1 million barrels Of blood, enough to ﬁll 3.3 supertankers [190].

The heart is located between the lungs in the middle Of thorax (chest), behind and slightly
to the left of the sternum (breastbone). It is a muscular organ enclosed in a double-
layered ﬁbrous sac called the pericardium. The outer layer of the pericardium surrounds

the roots of the heart's major blood vessels and is attached by ligaments to the spinal

267

column, diaphragm, and other parts of the body. The inner layer of the pericardium is
attached to the heart muscle. The extremely narrow space between the pericardium and
the heart is ﬁlled with a watery ﬂuid that serves as a lubricant as the heart moves within
the sac. The walls of the heart are called the myocardium. It is composed primarily of
cardiac muscle cells. The inner surface of the walls that is in contact with the blood
within the cardiac chambers is lined by a thin layer of cells known as endothelial cells, or

endothelium.

The human heart (Figure 12.1) is divided into left and right halves, each consisting Of an
upper chamber called the atrium and a lower chamber called the ventricle. A wall of
muscle called the septum separates the two halves of the heart. Located between the
atrium and ventricle in each half Of the heart is an atlioventricular (AV) valve that
prevents retrograde ﬂow of the blood from the atrium to ventricle. The left AV valve is
called the mitral valve, and the right is called the tricuspid valve. In addition to the AV
valves, two semilunar valves ensure the directional ﬂow Of blood exiting the heart from
the ventricles: The pulmonary valve located at the entrance to the pulmonary trunk, and
the aorta valve located at the entrance to the aorta. These valves permit blood to ﬂow
into the arteries during ventricular contraction but prevent blood from moving in the
Opposite direction during ventricular relaxation. Like the AV valves, the serrrilunar

valves Operate passively accordingly to the pressure differences across them.

268

Aorta

 

 

“U (maln artery
taking blood
Pulmonary X Pulmona'y artery
valve . 1». - 1° "‘9 WY) which takes blood
, ‘ ’ _- 7 to the lungs
Superior vena cava ' /\ . I,
(main vein . . . .- Pumonzy veins 0th
which brings // blood N0 the heart
blood into . tromthe lungs
the heart from
the head and ,' Left atrium
neck) ‘ _ \ ‘
r Aortic valve
R“ ht t ' " ‘
19 a num
) \ Mitral valve
/ Left ventricle

 

Interior vena cava

  
  

 

 

(ma'n vein

which brings
blood into

the heart from Tricuspid
the body) valve

Right ventricle

Figure 12.1 A simpliﬁed cross-section view of the human heart. After [135]

The pumping of blood by the heart is due to the orderly atrial and ventricular contractions
and relaxations known as the cardiac cycle. The cardiac cycle consists of two major
phases: The period of ventricular contraction and blood ejection called systole (Figure
12.2), and the period of ventricular relaxation and blood ﬁlling called diastole (Figure
12.3). The Opening and closing of the AV valves is a passive process resulting from
pressure differences across the valves during the cardiac cycle. During diastole, the
ventricle relaxes and the blood pressure in the atrium is greater than that in the adjacent
ventricle. The connecting AV valve is pushed Open and blood ﬂows from the atrium to
the ventricle. In contrast, during systole when a contracting ventricle achieves an internal
pressure greater than that in the adjacent atrium, the connecting AV valve is forced

Closed. Consequently, blood in the right ventricle is forced into the pulmonary trunk and

269

blood in the left ventricle is forced into the aorta. At an average heart rate of 72 beats per
minute, each cardiac cycle last approximately 0.85, with 0.33 in systole and 0.5s in

diastole.

 

Figure 12.2 During systole, the ventricles contract. The AV vales are closed and blood in
the right ventricle is forced into the pulmonary trunk while blood in the left ventricle is
forced into the aorta. After [136]

D iastole

 

Figure 12.3 During diastole, the ventricles relax. The AV valves are pushed open and
blood ﬂows from the atria to the ventricles. After [136]

270

To prevent the AV valves from being pushed up into the atrium, the valves are fastened
to muscular projections (papillary muscles) of the ventricular walls by ﬁbrous strands
(chordae tendinae). They act only to lirrrit the valves’ movements and prevent them from

being everted.

When Opened, a heart valve Offers very little resistance to ﬂow. In disease states,
however, a valve Opening may become narrowed so that it imposes a high ﬂow resistance
even when opened. In such a state, the contracting cardiac chamber must produce an
unusually high pressure to cause ﬂow across the valve, thus increasing the stress on the

cardiac muscles.

There are no valves at the entrances of the superior and inferior vanae cavae into the right
atrium, and Of the pulmonary veins into the left atrium. However, atrial contraction
pumps very little blood back into the veins because the contraction compresses the veins
at their sites Of entry into the atria, greatly increasing the resistance to backﬂow.
(Accordingly, a little blood is ejected back into the veins, and this accounts for the

venous pulse that can often be seen in the neck veins when the atria are contracting.)

A.2 Heart Valve Disorders
The heart valves can be malformed (due to prolapse) or malfunctioned either in the form
Of leaking (causing regurgitation) or by not Opening adequately and thus partially
obstructing the ﬂow of blood through the valve (causing stenosis). While valve prolapse

is unique to the mitral valve, stenosis and regurgitation can affect any of the heart valves,

271

sometimes simultaneously. These three disorders are described below with respect to the

rrritral valve.

A.2.l Mitral Valve Prolapse

Mitral valve prolapse occurs when the valve cusps (ﬂaps) are abnormally long, ﬂoppy,
billows out into the left atrium and does not close properly when the left ventricle
contracts. In most cases, mitral valve prolapse is harmless, asymptomatic and does not
need to be treated. In a small number Of instances, it can cause severe mitral

regurgitation (leakage Of blood back through the valve) and requires extensive treatment.

A.2.2 Mitral Stenosis or Obstruction

Mitral stenosis, also known as mitral valve obstruction, is a narrowing of the mitral valve
Opening. This prevents adequate blood ﬂow from the left atrium into ventricle during

diastole (Figure 12.4).

A.2.3 Mitral Regurgitation or Insufﬁciency

Mitral regurgitation, or rrritral valve insufﬁciency, is a disorder in which the mitral valve
does not close properly. This causes blood to leak (backﬂow, or regurgitation) into the

left atrium from the left ventricle during systole (Figure 12.5).

272

Narrowed
mitral valve

 

Figure 12.4 Mitral stenosis is the narrowing of the valve opening that reduces blood ﬂow
from the atrium to the ventricle. After [137]

— Leaking
mitral valve

 

Figure 12.5 In rrritral regurgitation, the improper closing Of the valve causes blood to leak
back into the atrium from the ventricle during systole. After [137]

A.3 Treatments of Heart Valve Disorders

Treatments of mitral valve prolapse, stenosis, and regurgitation can range from
monitoring, medications, to surgical repair and replacement Of the diseased valve,

depending on the severity of the case.

A.3.1 Observations and Monitoring

273

Most cases Of mitral valve prolapse and mild, asymptomatic cases of stenosis and
regurgitation do not require treatment. In case of mitral valve prolapse or regurgitation,
echocardiograms are recorded periodically to monitor the enlargement of the left

ventricle.

A.3.2 Medications

In more moderate cases of heart valve disorders, medications may be advised to relief the

symptoms and to prevent complications. Some of these medications include:

o Diuretics (water tablets) to help with breathlessness. They facilitate urine
production in the kidneys to remove excess blood and ﬂuid which may build up in
the lungs or other parts of the body with the “back pressure” from the heart.

0 ACE (angiotensin-converting) Inhibitors are medicines which help to reduce the
amount of work done by the heart and ease symptoms of heart failure.

0 Anti-arrhythmic medication such as a beta-blocker helps to control and lower the
heart rate in case of atrial ﬁbrillation, as well as to reduce palpitations (heartbeat
sensations that feel like pounding or racing, with either normal or abnormal
rhythms).

o Warfarin (anticoagulation) is also advised in cases of atrial ﬁbrillation to prevent
blood clots from forming.

- Antibiotics to prevent endocarditis. Damaged, abnormal, or artiﬁcial heart valves

are Susceptible to a serious bacteria infection called infective endocarditis. A

274

patient should take antibiotics before surgical, dental, or medical procedures to

reduce the risk Of an infection on a valve, even though this risk is small.

A.3.3 Surgical Repair

If medications do not reduce the symptoms satisfactorily, a surgical procedure may be
performed in attempt to correct the heart valve disorder. During valve repair, parts of a
valve are reshaped or removed to make the valve function more efﬁciently. Certain valve
problems can be repaired. For insufﬁciency, extra tissue is removed, or parts of the valve
are strengthened to help it close more tightly. This may be accomplished by sewing a
ring around the Opening of the valve. For stenosis, an obstructing valve can be stretched
open using a procedure called balloon valvuloplasty. In this procedure, a balloon-tipped
catheter is threaded through a vein into the heart. Once inside the valve, the balloon is
inﬂated, separating the valve leaﬂets. Alternatively, a heart surgery may be performed to
separate the fused cusps. One advantage of a heart valve repair Operation is that only the
patient’s own valve tissues are used. If a valve is too badly damaged, surgical

replacement with an artiﬁcial valve maybe the only option.

A.3.4 Heart Valve Replacement

Heart valves that are seriously deformed or degenerated cannot be repaired. In this case,
the Old valve is removed and replaced with a new valve mechanism. The new valve is
attached by suture to a rim of tissue kept from the original valve. The replacement valve

is either a mechanical valve made of plastic and metal or a biologic valve made of tissue,

275

usually from pigs, cows or humans, placed in a synthetic ring. The various types of

prosthetic valves are summarized in the following sections.

Choice of a valve depends on many factors, including characteristics Of the valve. A
mechanical valve lasts longer than a biologic valve but a lifetime therapy with an
anticoagulant (blood thinner) is needed to prevent the formation of blood clots on the
valve. A biologic valve does not last as long as a mechanical valve, but anticoagulants
are rarely required for the patient. Therefore, a patient’s suitability to take anticoagulants
is an important deciding factor. For instance, anticoagulants may not be appropriate for
women of childbearing age because anticoagulants cross the placenta and may affect the
fetus. The patient’s age, activity level, the condition of the heart as well as the location of
the damaged valve are all important factors affecting the choice of the replacement valve.

Typically, mechanical valves are preferred in younger and more active patients.

Heart valve replacement surgery requires general anesthesia of the patient. The heart is
stopped and a heart-lung machine is used to by-pass the heart. The damaged valve is
removed, and the replacement valve is sewn in place. The incisions are closed, the heart-
lung machine is disconnected, and the heart is restarted. The operation may take between
2 to 5 hours. After surgery, the patient is monitored in an intensive care unit for one or
two days before being moved to an in-patient care unit. Full recovery may take anywhere

from 6 to 8 weeks.

276

A.4 Types of Prosthetic Heart Valves

Prosthetic heart valves are either mechanical or tissue valves. The following sections

described the structures and provide examples of each type of these valves.

A.4.1 Mechanical Valves

Mechanical heart valves include the ball valves and disc valves. The disc valves include
the single leaﬂet and bileaﬂet disc valves. The single disc valves can be further divided
into caged disc or tilting disc valves. Examples of these mechanical prosthetic heart

valves are given below.

A.4.2 Ball Valves

Ball valves, also known as ball-in-a—cage valves, were the earliest prosthetic heart valves
implanted in humans. They remain one of the most reliable and durable designs
available. Designed by engineer Lowell Edwards and Dr. Albert Starr, the Starr-Edwards
Silastic ball valve was the world's ﬁrst commercially available replacement mitral valve
The Starr-Edwards model 1260 aortic valve (European model: 1200) and the mitral

model 6120 are shown in Figure 12.6.

The original Starr-Edwards valve suffered from a suboptimal hemodynamics design.
Since the ball had to be larger than the valve Opening to stop the back-ﬂow of blood
when the valve was closed, this left a large ball in the middle of the blood stream that

partially obstructed blood ﬂow. Design improvements were made in later models to

277

reduce the relative size of the occluder ball. The last available models, 1200, 1260 and

6120, discontinued in 2004, had been unchanged since 1968.

 

Figure 12.6 Starr-Edwards Silastic Ball Valves. Top: Mitral Model 1260. Bottom:
Aortic Model 6120.

A.4.3 Disc Valves

There are two kinds of disc valves: Single leaﬂet and bileaﬂet valves. A single leaﬂet
valve can be either a caged disc or tilting disc valve. The earliest disc valves implanted
were single leaﬂet valves and they were very successful. Although the development Of
the bileaﬂet valve did not begin until later, as of 1997 bileaﬂet valves dominated the

mechanical valve market in the United States with a 90% market share [138]

278

Single leaﬂet disc valves. The original BjOrk-Shiley single leaﬂet valve (Figure 12.7) had
a very successful and durable design. Although this valve is no longer sold in the United
States, it is arguably the most controversial single leaﬂet tilting disc valve. Issues with

the BjOrk-Shiley single leaﬂet valve are discussed in Chapter 3.

Single leaﬂet caged disc valves (e.g., the Kay-Shiley TGCD series valve shown in Figure
12.8) are no longer available in the United States. The Medtronic-Hall valve (Figure
12.9) and the Omniscience valve (Figure 12.10) are the two single tilting disc valves still

being used in the US.

 

Figure 12.7 BjOrk-Shiley titling disc valve

 

Figure 12.8 Kay-Shiley TGCD series single leaﬂet caged disc valve.

Bileaﬂet disc valves. The most successful bileaﬂet disc valve is the St. Jude valve

(Figure 12.11) and it is the leading choice for mechanical valve implantations in the

279

United States [138]. Other bileaﬂet valves available in the US. include the Carbomedics

valve (Figure 12. 12) and the On-X valve (Figure 12.13).

 

Figure 12.9 Medtronic-Hall single tilting disc valve.

 

Figure 12.11 St. Jude standard bileaﬂet valve.

280

 

Figure 12.12 Carbomedics CPHV standard bileaﬂet valve.

 

Figure 12.13 On-X bileaﬂet valve.

A.4.4 Tissue Valves

Tissue valves are also called bioprosthetic valves. There are three types of tissue valves:

Animal tissue valves, human tissue valves and Synergraft (synthetic tissue valves).

A.4.4.1 Animal Tissue Valves (xenografts, heterografts)

Animal tissue valves are called xenografts (from the Latin preﬁx “xeno-,” for foreign) or
heterografts (where the preﬁx “hetero-” means “different”). The most commonly used
xenografts can either be of valve tissue, typically porcine or pig valve tissue, or of non-
valve pericardial tissue from bovine or cow. Four examples of the porcine valves are the
Hancock (standard and modiﬁed oriﬁce) valves (Figure 12.14), the Carpentier-Edwards
Porcine valve (Figure 12.15), the St. Jude Toronto Stentless Porcine valve (Figure 12.16),
and the Medtronic Freestyle Aortic Root Bioprosthesis (Figure 12.17). The later two

employ stentless designs intended to improve the hemodynamics of the valves. The only

281

pericardial valve currently available is the Carpentier-Edwards Pericardial valve (Figure

12.18) and its variants.

 

Figure 12.14 Hancock porcine valve.

 

Figure 12.16 St. Jude Toronto Stentless Porcine Valve (SPV).

282

 

Figure 12.17 Freestyle Stentless Aortic Bioprosthesis: End view (left) and lateral view
(right).

 

Figure 12.18 Carpentier-Edwards pericardial valve: Lateral view (left) and top View
(right).

A.4.4.2 Human Tissue Valves (homografts, autografts) and the Ross Procedure

A human tissue valve transplant from a donor person is called a homograft valve.
Homograft valves tend to be durable and have good hemodynarrrics, although it is
uncertain at the time of this writing whether these characteristics are signiﬁcantly better
than ordinary tissue valves. There are generally no rejection issues and
immunosuppressive therapy is typically no required. One drawback of the homograft
valves is that they are more technically challenging to implant than standard tissue
valves. In addition, homograft availability is limited by donor availability. Also, since
the valve is typically cryopreserved in liquid nitrogen until it is used, the valve must be

thawed overnight before the surgery. This means that the patient's valve size must be

283

known beforehand. This information may be ascertained through the use of either

echocardiograms or Magnetic Resonance Imaging (MRI) images.

All autograft is a valve moved from one position to another within the same patient. The
most common autograft procedure is moving the pulmonary valve to the aortic position,

also known as the Ross procedure [139]. Figure 12. 19 explains the Ross procedure.

 
 
 
   

Pulmonary
1 ~~~~~~~~ ~ Artery

I
.Aortic Pulmonary‘

Transplanted \ Valve '

Pulmonary Valve

      
  

Final Result

  
  

 
 

Coronary
Artery Pulmonary

Artery
Autograft

    
   

Figure 12.19 The Ross procedure. The numbers depict the surgery sequence. After [143]

A.4.4.3 Decellularized homograft and xenograft aortic prosthesis (Synergraft)
Although still in its infancy and highly experimental, Synergraft [l44]-[146] is intended

to be the ideal replacement valve Of the future. Synergraft is a tissue-engineered valve

284

made from either synthetic materials, stem cells, or a mixture of both. To create the '
valve, a suitable immunologically inert scaffold is used as a matrix for cellular
population. The scaffold is biodegradable and is eventually replaced by autologous tissue
growth. Some choices of biological scaffold material include ﬁbrin gel and extracellular
matrix, while synthetic matrices include glactin polymers, glycolic acid and polyurethane

[147].

After the scaffold is in place, it is seeded with appropriate cells for repopulation.
Endothelial cells and myoﬁbroblasts from the eventual host may be grown onto the
scaffold. Alternatively, mesenchymal stem cells, or bone marrow progenitor cells, as
well as allogenic stem cells (derived from embryonic tissue) may also be used. These
stem cells can potentially differentiate into cardiovascular cell lines and tissues that are

morphologically compatible with the host organ.

A completely tissue-engineered replacement valve must also posses the strength for
attachment to the recipient’s existing valve apparatus, and the biomechanical properties
to sustain the hemodynamical stress and strain inside the heart. Work [148] is underway
to characterize and develop numerical models for the cellular structure and
biomechanical properties of the mitral valve (leaﬂet), as well as to develop biological and

synthetic materials [149] for the scaffold and valve repopulation.

285

APPENDIX B. ANIMAL STUDY EXAMINATION PROCEDURE

This chapter describes the animal study procedure, provides a list of critical
considerations and recommends several data events to be captured for the gradiometer-
based BSCC heart valve examination system. To ensure success, the animal study must
be conducted by qualiﬁed and experienced medical and engineering personnel to ensure
that the listed critical considerations are addressed and Observed. The list of
recommended data collection events should be followed whenever a listed data source is
available. Minimally, the gradiometer-based examination system outputs must be

recorded.
B.1 The Animal Study Procedure

B.1.1 Animal Support

During the heart valve examination, the animal is supported by a custom surgery table
(Figure 12.20) which provides height, roll, pitch, and yaw angle adjustments. With the
animal secured in the supine (or lateral decubitus, depending on the surgeon’s preference)
position, the table top of the surgery bed is inserted into the region between the excitation
coils (see Figure 12.21). Note that the table top is custom constructed from non-metallic

components that do not interfere with the magnetic excitation ﬁeld.

B.1.2 Position and Orientation of the Test Animal

With the aid Of a mobile ﬂuoroscope, identify the position and orientation of the heart

valve. Adjust the surgery bed to the proper height to ensure that the heart valve is located

286

at the center of the excitation coils. Orient the animal (by adjusting the surgery bed) to
align the heart valve ﬂange with the axis of the portable excitation coils. Note that the
proper orientation of the heart valve within the excitation coils is critical to the success of
this examination technique. The cardiologist must provide accurate orientation

information and supervise the alignment of the heart valve with the coils.

 

Figure 12.20 The custom surgery table for the gradiometer-based heart valve examination
system

r) - 17' *

 

B.1.3 Calibration and Zeroing of Measurement Instruments

Prior to inserting the animal between the excitation coils, position a guide catheter into
the left ventricle under ﬂuoroscopy. Advance a gradiometer via the catheter lumen to a

position approximately 10 cm away from the outlet strut. Remove the ﬂuoroscope and

287

position the mobile excitation coils to surround the animal such that the heart valve is at
the center of the coils. (It is recommended to move the excitation coils rather than the
surgery bed as the animal will be attached with life-support and anesthesia equipment at
this point.) Energize the excitation coils and calibrate the test equipment at this
gradiometer location to remove the effects of background noise from the operating room.
Switch Off and remove the excitation coils, and reposition the portable ﬂuoroscope over

the test animal.

 

Figure 12.21 Table top of the custom surgery bed is positioned between the excitation
coils through a side opening during examination

288

B.1.4 Positioning of the Gradiometer

Under ﬂuoroscopy, advance the gradiometer into the left ventricle and position the distal
end in the mitral annulus. Due to cardiac activities, the relative distance between the
gradiometer and the outlet strut will change in an oscillatory fashion. At its closest point,
the gradiometer should be inside the outlet strut loop (Figure 12.22) and as close to the
strut plane as possible without crossing the plane (Figure 12.23). Figure 12.24 shows the

Optimal location Of the gradiometer in relation to the outlet strut.

as

Figure 12.22 Optimal position of the gradiometer inside the outlet strut loop

2&1: 5 mm

Figure 12.23 The gradiometer should be placed as close to the outlet strut loop as
possible without crossing the plane of the loop

 

 

Figure 12.24 Optimal position of the gradiometer in relation to the outlet strut

289

B.1.5 Examination and Data Acquisition

After the gradiometer is properly positioned, remove the ﬂuoroscope and energize the
external low-ﬁeld excitation coils. Record the gradiometer output signals for 10 or more
cardiac cycles. Switch Off and remove the excitation coils and readjust the position of the
gradiometer under ﬂuoroscopy. Reposition the excitation coils over the animal and
reenergize the system. Record the gradiometer system output signals again. Repeat the
gradiometer position adjustment three to ﬁve times, depending on signal quality, and

record the gradiometer system outputs at each position.

B.1.6 Animal Recovery

After the tests, remove the sensor and the sheath and repair the arterial incision site.

Extubate and recover the animal.

B.2 Critical Considerations

There are a few important considerations that must be satisﬁed to ensure the success the
of gradiometer-based examination technique:
(a) The heart valve ﬂange must be aligned with the center (z-) axis of the excitation
coils
(b) The outlet strut is to be positioned at either the due north or due south direction
with respect to the center axis of the excitation coils
(c) The gradiometer must be positioned inside the outlet strut loop at the location

shown in Figure 12.22 and Figure 12.24

290

(d) The gradiometer must be placed as close to (within 3 mm) the plane of the outlet

strut loop as possible without crossing the plane.

The lead cardiologist is expected to provide accurate information regarding the
orientation and location of the valve in relation to the distal end of the catheter. He or she

is also expected to position the gradiometer at the desired location.

B.3 Recommended Data Capture Events

The following is a list of recommended data capture events and items for both record
keeping and analyses purposes. These items should be recorded whenever the data
source is accessible and a mean of recording is available. At a minimum, the inspection

system outputs must be recorded.

(a) Date, time, and location of the animal study

(b) Names and ranks or titles Of the investigating personnel

(c) Digital photos of the operating room and instruments

((1) Digital photos of the test animal before the study

(e) Digital photos of the test animal after catherization

(f) Digital photos Of test animal on the custom examination bed, surrounded by the
excitation coils

(g) Video recording of the ﬂuoroscope cinegram during the entire procedure

(h) Recording of the test animal ECG

(i) Capture of ﬂuoroscope image at each gradiometer position

291

(j) Settings of the gradiometer-based system instruments (both the excitation and
measurement systems)

(k) Inspection system outputs for 10 or more cardiac cycles

(1) A second measurement of the system outputs in part (k), if necessary

(m) Repeat part (k) as the gradiometer position is readjusted (three to ﬁve times,

depending on the signal quality)

292

BIBLIOGRAPHY

293

BIBLIOGRAPHY
[1] Kraus, John Daniel, Electromagnetics, New York, New York: McGraw-Hill, 1973.

[2] Sadiku, Mathew N.O., Elements ofElectromagnetics, 2nd Edition, New York, New
York: Saunders College Publishing/Harcourt Brace College Publishers, 1994, p 412.

[3] Balanis, Constantine A., Advanced Engineering Electromagnetics, New York, New
York: John Wiley & Sons, 1989.

[4] BjOrk, Viking O., “A new tilting disc valve prosthesis,” Scandinavian Journal of
Thoracic and Cardiovascular Surgery, Vol. 3, 1969, pp. 1-10.

[5] Moulton, A.L., Lincoln, 8., Attar, S., Green, D. C., Sequeira, A., and McLaughlin, J.
S., “The convexo-concave modiﬁcation of the BjOrk-Shiley prosthesis: signiﬁcant
reduction of thromboembolism (abstract),” Journal of American College of
Cardiology, Vol. 5, 1985, p 391.

[6] Kuntze, C. E., Eijgelaar, A., Ebels, T., and Homan Van Der Heide, J .N., “Rates of
Thromboembolism with Difﬁcerent Mechanical Heart Valve Prostheses:
Randomised Study,” The Lancet, Vol. 333, Issue 8637, March 11, 1989, pp. 514-517.

[7] BjOrk, Viking O., Henze, A., and Hindmarsh, T., “Radiopaque marker in the tilting
disc of the BjOrk-Shiley heart valve. Evaluation of in vivo prosthetic valve function
by cineradiography,” Journal of Thoracic and Cardiovascular Surgery, Vol. 73,
1977, pp. 563-569.

[8] Bjﬁrk, Viking O., “The improved BjOrk-Shiley tilting disc valve prosthesis,”
Scandinavian Journal of Thoracic and Cardiovascular Surgery, Vol. 12, 1978, pp.
8 1-84.

[9] Chronology of the ij'rk-Shiley Cardiac Prostheses, Shiley, Incorporated, Irvine, CA,
1986.

[10] Shiley, Inc., presentation in FDA and the Medical Device Industry, Hearing Before
the Subcommittee on Oversight and Investigations of the Committee on Energy and
Commerce, US. House of Representatives, One Hundred First Congress, Second
Session, Ser. NO. 101-127, Feb. 26, 1990, p 175.

[11] Brubakk, O., Simonsen, S., Kallman, L., and Fredriksen, A., “Strut fracture in the
new BjOrk-Shiley mitral valve prosthesis,” The Thoracic Cardiovascular Surgeon,
Vol. 29, 1981, pp. 108-109.

[12] Lindblom, D., BjOrk, V.O., and Semb, B.K.H., “Mechanical failures of the BjOrk-

Shiley valve: incidence, clinical presentation, and management,” Journal of
Thoracic and Cardiovascular Surgery, Vol. 92, 1986, pp. 894-907.

294

[13] Fielder, John H., “Ethical Issues in Biomedical Engineering: The BjOrk-Shiley
Heart Valve,” IEEE Engineering in Medicine and Biology, March 1991, pp. 76-78.

[14] Chandler, J., Knight, R., and Strom, J ., “Broken Heart Valves,” Emergency
Medicine, Cahners Publishing Co.: Highlands Ranch, CO, Vol. 24, June 15, 1992.

[15] Hedger, Philip, “Single-leg separations In explanted Bjork-Shiley convexo-concave
heart valves,” Lancet, Vol. 343, NO. 8892, January 29, 1994, pp. 294-295.

[16] Fielder, J. H., “Defects and Deceptions—The BjOrk-Shiley Heart Valve,” IEEE
Technology and Society Magazine, Fall 1995, pp. 17-22.

[17] Wieting, David. W., Eberhardt, Allen. C., Reul, H., Breznock, E. M., Schreck, S. G.,
and Chandler, J. G., “Strut fracture mechanisms of the Bjork-Shiley convexo-
concave heart valve,” Journal of Heart Valve Diseases, Vol. 8, No. 2, March 1999,
pp. 206-217.

[18] Guo, G. X., Xu, C. C., and Hwang, N. H. C., “The closing velocity of Baxter-
Duromedics heart valve prostheses,” American Society for Artificial Internal
Organs (ASAIO) Transactions, Vol. 36, NO. 3, 1990, pp. 529-532.

[19] Eberhardt, Allen C., Ward, Mark A., Lewandowski, Simon J., Inderbitzen, Rebecca,
Wieting, David W., “Relationships between closure delay, pressure, velocity, and
outlet strut forces of the BjOrk-Shiley heart valve,” Advances in Bioengineering,
ASME, Vol. 26, 1993, pp. 547-550.

[20] Rau, Giinter, Reul, Helmut, Eichler, Michael, Schreck, Stefan, and Wieting, David
W., “The Effect of Left Ventricular dP/dt on the In Vitro Dynamics of the BjOrk-
Shiley Convexo-Concave Mitral Valve,” The Journal of Heart Valve Disease, Vol.

4. Suppl. 1, 1995. pp. 817-820.

[21] Schreck, S., Inderbitzen, Rebecca, Chin, H., Wieting, David W., Smilor, M.,
Breznock, E., and Pendray, D., “Dynamics of BjOrk-Shiley Convexo-Concave
Mitral Valves in Sheep,” The Journal of Heart Valve Disease, Vol. 4, Suppl. 1,

1995, pp. $21-$25.

[22] Chandran, Krishnan B., Lee, Chong-sun, Aluri, Srinivas, and Dellsperger, Kevin C.,
“Pressure Distribution Near the Occluders and Impact Forces on the Outlet Structs
of BjOrk-Shiley Convexo—Concave Valves during Closing,” The Journal of Heart
Valve Disease, Vol. 5, No. 2, Mar 1996, pp. 199-206.

[23] Structural Acoustics, Incorporated, Acoustic Characterization of Outlet Strut

Condition and Loading, a research proposal submitted to the Bowling-Pﬁzer
Supervisory Panel, July 26, 1994, p. 20 (access through the Bowling-Pﬁzer

295

Document Repository web site: www.bowling-pﬁzer.com/repository.
Title=03001 1.).

[24] BjOrk, Viking O., “Metallurgic and design development in response to mechanical
dysfunction of Bj'Ork-Shiley heart valves,” Scandinavian Journal of Thoracic and
Cardiovascular Surgery, Vol. 19, 1985, pp. l-12.

[25] Sacks, S. H., Harrison, M., Bischler, P. J., Martin, J. W., Watkins, J ., and Gunning,
A., “Metallurgical analysis of failed BjOrk-Shiley cardiac valve prostheses,”
Thorax, Vol. 41, Feb 1986, pp. 142-147.

[26] ROckelein, G., Breme, J ., and van der Emde, J ., “Lethal blockage of a BjOrk-Shiley
artiﬁcal heart valve caused by strut failure — The metallurgical aspect,” The
Thoracic Cardiovascular Surgeon, Vol. 37, 1989, pp. 47-51.

[27] Swieten, van H. A., Mol, de B. A., Defauw, J. J., Overkamp, P. J ., and Verrneulen,
F. E. E., “Metallurgical analysis of the BjOrk-Shiley convexo-concave valve

prosthesis to assess the case of late outlet strut fracture,” Surgery for Heart Valve
Disease,” Bodnar, E., editor, ICR Publishers: London, 1990, pp. 616-627.

[28] Xiao, K., and Appleby, A. J ., “Stress Corrosion Cracking in BjOrk-Shiley Convexo-
Concave Prosthetic Heart Valves Due to Random In Vivo Electrochemical Pulsing,”
International Journal of Artiﬁcial Organs, Vol. 19, No. 8, August 1996, pp. 477-
486. (Center for Electrochemical Systems and Hydrogen Research, Texas
Engineering Experiment Station/Texas A & M University, College Station, USA)

[29] Wenzel, Thomas G, Manning, Charles R., Chandler, James G., and Williams David
F., “Welding Metallurgy's Putative Inﬂuence on BjOrk-Shiley Convexo-Concave
Valve Outlet Strut Failures,” Journal of Heart Valve Diseases, Vol. 8, No. 2, March
1999, pp. 218-231.

[30] Sethia B., Quin R. O., and Bain, W.H., “Disc embolization after minor strut fracture
in a BjOrk-Shiley nritral valve prosthesis,” Thorax, Vol. 38, 1983, pp. 390-391.

[31] Ibarra F., Gutierrez, A., Martinez, F., Carreras, L., Lopez, C., and Alonso-Lej, F.,
“Fracture of the outlet strut of a BjOrk-Shiley mitral prosthesis: emergency

Operation with survival,” Journal of Thoracic and Cardiovascular Surgery, Vol. 87,
1984, pp. 315-318.

[32] Garcia-De-Castillo, H., Larrousse-Perez, E., Murtra-Ferre, M., and Soler-Soler, J .,
“Strut fracture and disc embolization of a BjOrk-Shiley mitral valve prosthesis,”
American Journal of Cardiology, Vol. 55, 1985, pp. 597-598.

[33] Davis, P. K., Myers, J. L., Pennock, J. L., and Thiele, B. L., “Strut fracture and disc

embolization in BjOrk-Shiley rrritral valve prostheses: diagnosis and management,”
The Annul of Thoracic Surgery, Vol. 40, 1985, pp. 65-69.

296

[34] Miccolo M. L., Lemole, G. M., and Spagna, P. M., “Late strut fracture and disc 23 in
a 27mm BjOrk-Shiley mitral prosthesis,” American Heart Journal, Vol. 110, 1985,
pp. 898-899.

[35] Khalil, Y., Sethia, B., Quin, R. 0., and Bain, W. H., “Disc and strut embolization
after minor strut fracture in a BjOrk-Shiley rrritral valve prosthesis,” Thorax, Vol.
40, 1985, pp. 158-159.

[36] Hendel, P. N., “BjOrk-Shiley strut fracture and disc escape: literature review and a
method Of disc retrieval,” The Annals of Thoracic Surgery, Vol. 47, 1989, pp. 436-
440.

[37] Mass, A. H. E. M., van Swieten, H. A., Knaepen, P. J ., Defauw, J. J., Brute], de la
Riviere A., and van Herpen, G., “Successful emergency surgery after strut fracture
of BjOrk-Shiley mitral heart valve,” Ned Tijdschr Geneeskd, Vol. 130, 1986, pp.
835-838.

[38] Tanaka, M., Abe, T., Takeuchi, E., Watanabe, T., and Tamaki, S., “Intraoperative
echocardiography of a dislodged Bjork-Shiley mitral valve disc,” The Annals of
Thoracic Surgery, Vol. 51, 1991, pp. 315-316.

[39] Personal electronic mail correspondence with David E. Miller, Chairman of the
Bowling-Pﬁzer Supervisory Panel, Nov 2004 and March 2005.

[40] Kallewaard, Marjon, Algra, Ale, and van der Graaf, Yolanda, “Welder identity, weld
date, and the risk of outlet strut fracture in BjOrk-Shiley convexo-concave valves:
the Dutch cohort study,” Heart, Vol. 76, 1996, pp. 510-512.

[41] Walker, Alexander M., Funch, Donnie P., Sulsky, Sandra L, and Dreyer, Nancy A.,
“Manufacturing Characteristics Associated with Strut Fracture in BjOrk-Shiley 60°
Convexo—Concave Heart Valves,” Journal of Heart Valve Diseases, Vol. 4, NO. 6,
November 1995, pp. 640-648.

[42] Walker, Alexander M., Funch, Donnie P., Sulsky, Sandra 1., and Dreyer, Nancy A.,
“Shop order fracture rate as a risk factor for strut fracture in BjOrk-Shiley CC 60
degrees heart valves,” Journal of Heart Valve Diseases, Vol. 6, NO. 3, May 1997,
pp. 264-267. ‘

[43] Kallewaard, Marjon, Algra, Ale, Defauw, Jo, and van der Graaf, Yolanda, “Which
Manufacturing Characteristics are Predictors of Outlet Strut Fracture in Large
Sixty-Degree BjOrk-Shiley Convexo-Concave Mitral Valves? “ The BjOrk-Shiley
Study Group, Journal of Thoracic and Cardiovascular Surgery, Vol. 117, 1999, pp.
766-775.

297

[44] Omar, Rumana Z., Morton, Linda S., Beime, Mairead, Blot, William J., Lawford,
Patricia V., Hose, Rodney, and Taylor, Kenneth M., “Outlet strut fracture of BjOrk-
Shiley convexo-concave valves: Can valve-manufacturing characteristics explain
the risk?” Journal of Thoracic and Cardiovascular Surgery, Vol. 121, NO. 6, 2001,
pp. 1143-1149.

[45] Omar, Rumana Z., Morton, Linda S., Murad, Shahed, and Taylor, Kenneth M., “Use
of ﬂexibility tests in the manufacturing process of 60° BjOrk-Shiley convexo-
concave valves and the risk of outlet strut fracture,” Journal of Thoracic and
Cardiovascular Surgery, Vol. 126, 2003, pp. 832-836.

[46] Advanced Computational and Engineering Services (ACES, Columbus, Ohio),
Evaluation of Hook to Well Separation on Tip Loading, a research proposal
approved by the Bowling-Pﬁzer Supervisory Panel, May 2004. (Available from the
Bowling-Pﬁzer Document Repository: http://www.bowling-pﬁzer.com/repository.)

[47] van Der Graaf, Y., Waard, F. D., van Herwerden, L. A., and Defauw, J ., “Risk of
strut fracture of BjOrk-Shiley valves,” The Lancet, Vol. 339, No. 8788, February
1992, pp. 257-61.

[48] Ericsson, A., Lindblom, D., Huysmans, H. A., Thulin, L. I., Scully, H. E., Bennett, J.
G., Osterrneyer, J ., and Grunkemeier, G. L., “Strut fracture with BjOrk-Shiley 70o
convexo-concave valve: An international multi-institutional follow-up study,”
European Journal of Cardio-thoracic Surgery, Vol. 6, 1992, pp. 339-346.

[49] Grunkemeier, J. D., Chandler, J. G., Miller, C., Jamieson, W. R. E., and Starr, A.,
“Utilization of manufacturer’s implant card data to estimate heart valve failure,”
Journal of Heart Valve Diseases, Vol. 2, 1993, pp. 493-503.

[50] Walker, Alexander M., Funch, Donnie P., Sulsky, Sandra 1., and Dreyer, Nancy A.,
“Patient Factors Associated With Strut Fracture in BjOrk-Shiley 60° Convexo-
Concave Heart Valves,” Circulation, Vol. 92, No. 11, December 1995, pp 3235-
3239.

[51] Piehler, Jeffery M., Blackstone, Eugene H., Bailey, Kent R., Sullivan, Michael B.,
Pluth, James, R., Weiss, Noel S., Brookmeyer, Ron S., and Chandler, James G.,
“Reoperation on prosthetic heart valves: Patient-speciﬁc estimates of in-hospital
events,” European Journal of Thoracic and Cardiovascular Surgery, Vol. 109, NO.

1, 1995, pp. 30-48.

[52] de Mol, Bas A. J. M., Overkamp, Peter J ., van Gaalen, Geerda L., and Becker, Anton
B., “Non-destructive assessment of 62 Dutch BjOrk-Shiley convexo-concave heart

valves,” European Journal of Thoracic and Cardiovascular Surgery, Vol. 11, 1997,
pp. 703-709.

298

[53] Omar, Rumana Z., Morton, Linda S., Halliday, D. A., Danns, E. M., Beime, M. T.,
Blot, W. J ., and Taylor K. M., “Outlet strut fracture of BjOrk-Shiley convexo-
concave heart valves: the UK cohort study,” Heart, Vol. 86, 2001, pp. 57-62.

[54] UK Cohort, Patient F ollow-up to Improve Estimation of Risk of Strut Fracture,
2002-2003, sponsored by the Bowling-Pﬁzer Supervisory Panel, 2003.

[55] Lindblom, D., Rodriguez, L., and BjOrk, V.O., “Mechanical failure of the BjOrk-
Shiley valve: updated follow-up and considerations on prophylactic
rereplacement,” Journal of Thoracic and Cardiovascular Surgery, Vol. 97, 1989,
pp. 95-97.

[56] Birkmeyer, J. D., Marlin, C. A., and O’Connor, G. T., “Should patients with BjOrk-
Shiley valves undergo prophylactic replacement?” Lancet, Vol. 340, 1992, pp. 520-
523.

[57] Marlin, C. A., Birkmeyer, J. D., and O’Connor, G. T., “The BjOrk-Shiley dilemma,”
Annal of Thoracic Surgery, Vol. 55, 1993, pp. 1361-1364.

[58] van der Meulen, Jan H. P., Steyerberg, Ewout W, van der Graaf, Y., et al., “Age
thresholds for prophylactic replacement of Bj'Ork-Shiley convexo-concave heart
valves: a clinical and economic evaluation,” Circulation, Vol. 88, 1993, pp. 156-
164.

[59] de Mol, Bass A., Kallewaard, M., McLellan, R. B., van Herwerden, L. A., Defauw,
J. J ., and van der Graaf, Y., “Single-leg strut fractures in explanted BjOrk-Shiley
valves,” Lancet, Vol. 343, January 1994, pp 9-12.

[60] Steyerberg, Ewout W, van der Meulen, Jan H. P., van Herwerden, L. A., et a1.
“Prophylactic replacement of BjOrk-Shiley convexo-concave heart valves: an easy-
to-use tool to aid decision-making in individual patients,” Heart, Vol. 76, 1996, pp.
264-268.

[61] Kallewaard, Marjon, Defauw, JO, and van der Graaf, Yolanda, “Psychological
distress among recipients of B jOrk-Shiley convexo-concave valves: the impact of
information,” Heart, Vol. 78, 1997, pp. 577-580.

[62] van Gorp, M. J ., Steyerberg, E.W., Kallewaard, Marjon, and van der Graaf, Yolanda,
“Clinical prediction rule for 30—day mortality in BjOrk-Shiley convexo-concave
valve replacement,” Journal of Clinical Epidemiology, Vol. 56, NO. 10, 2003, pp.
1006-1012.

[63] van Gorp, M. J ., Steyerberg, E.W., and van der Graaf, Yolanda, “Decision
Guidelines for Prophylactic Replacement of BjOrk-Shiley Convexo-Concave Heart
Valves: Impact on Clinical Practice,” Circulation, Vol. 109, No. 17, May 4, 2004,
pp. 2092-2096.

299

[64] Lindblom, D., “Management of patients with Bj'Ork-Shiley prosthetic valves,”
British Heart Journal, Vol. 68, 1992, p. 249.

[65] Hiratzka, Loren F., Kouchoukos, Nicholas T., Grunkemeier, Gary L., Miller, D.
Craig, and Scully, Hugh E., “Outlet Strut Fracture of the BjOrk-Shiley 60°
Convexo-Concave Valve: Current Information and Recommendations for Patient
Care,” Journal of American College of Cardiology, Vol. 11, 1988, pp. 1130-1137.

[66] Bowling-Pﬁzer Supervisory Panel, “Guidelines to Assess Patients with BjOrk-Shiley
Convexo-Concave Heart Valves for Elective Explantation,” sent as an attachment to
a “Dear Doctor” letter, August 1997 (http://www.bowling-
pﬁzer.com/guidelines.htm).

[67] Bowling-Pﬁzer Supervisory Panel, “Amended Guidelines to Assess Patients with
BjOrk-Shiley Convexo-Concave Heart Valves for Elective Explantation,” sent as an
attachment to a “Dear Doctor” letter, March 2000 (http://www.bowling-
pﬁzer.com/guidelines.htm).

[68] Bowling-Pﬁzer Supervisory Panel, “The 2003 Amended Guidelines to Assess
Patients with BjOrk-Shiley Convexo-Concave Heart Valves for Elective
Explantation,” sent as an attachment to a “Dear Doctor” letter, October 2003
(http://www.bowling-pﬁzer.corn/guidelines.htm).

[69] Eberhardt, A., “Finite Element Stress Analysis,” Accident Reconstruction Analysis,
Inc., 5801 Lease Lane, Raleigh, NC 27613, Oct, 1989.

[70] Chia, Raymond, “Finite Element Analysis of Vibrations of the BjOrk-Shiley
Convexo—Concave Heart Valve,” Proceedings of the 1994 IEEE Seventh
Symposium on Computer-Based Medical Systems, June 1994, pp. 48-52.

[71] Goldstein, J. A., Zucker, R. P., and Lee, B. Y., “Echocardiographic demonstration of
outlet strut fracture Of a BjOrk-Shiley nritral prosthesis,” Journal of the American
College of Cardiology, Vol. 7, 1986, pp. 949-952.

[72] Chandra, Rajiv, Bilsker, Martin, Myerburg, Robert J ., and Kessler, Kenneth M.,
“Echocardiographic diagnosis of outlet strut fracture of a BjOrk-Shiley prosthesis in
the mitral position,” The American Journal of Cardiology, Vol. 58, Issue 11, 1986,
pp. 1117-1118.

[73] Assanelli, D., Aquilina, M., Marangoni, S., MOrgagni, G. L., and Visioli, O.,
“Echophonocardiographic evaluation Of the BjOrk-Shiley mitral prosthesis,”
American Journal of Cardiology, Vol. 57, 1986, pp. 165-170.

[74] Min, Pu, Vandervoort, Pieter M., Greenberg, Neil L., Bibawy, George A., Eberhardt,
Allen C., Schreck, Stefan, and Thomas, James, D., “Assessment of The Closing

300

Dynamics In BjOrk-Shiley Mitral Prosthesis By Digital Echocardiography: In Vitro
Validation,” Computers in Cardiology 1994, Sept. 1994, pp. 349-352.

[75] Mehlman, D. J. and Resnekov, L., “A guide to radiographic identiﬁcation of
prosthetic heart valves,” Circulation, Vol. 57, 1978, pp. 613-23.

[76] Mehlman, D. J ., “A guide to "radiographic identiﬁcation of prosthetic heart valves: an
addendum,” Circulation, Vol. 69, 1984, pp. 102-105.

[77] Cesmeli, Erdogan, Powell, Kimberly A., Greenberg, Neil L., and Comhill, Fredrick,
“An Automated Temporal Alignment Technique for the Translational and
Rotational Correction Of Digital Radiographic Images of BjOrk-Shiley Heart
Valves,” Proceedings of Computers in Cardiology I 993, Sept 1993, pp. 619-622.

[78] Powell, K. A., Comhill, J. F., Nissen, S. N., Chandler, J. G., Abolfathi, A. H.,
LaPrestO, E. L., Herderick, E. B., Mueller, K., Cesmeli, E., and Chandra, S.,
“Digital radiographic analysis of single-leg separations (SLS) in the outlet struts of
BjOrk-Shiley convexo-concave valves,” Computer Assisted Radiology, Lemke, H.
U., Inamura, K., Vannier, M. W., Editors, Springer-Verlag: Berlin, Germany: 1993.

[79] Abolfathi, Amir H., Hirsch, Jamie Lee, Nissen, Steven B., Wieting, David W., and
Chandler, James G., ”Effect of Valve Proﬁling and Increasing kVp on In Situ
Radiographic Detection of Single Leg Separations of the BjOrk-Shiley Convexo-
Concave Heart Valve,” Engineering in Medicine and Biology Society, I 994.
Engineering Advances: New Opportunities for Biomedical Engineers. Proceedings
of the 16th Annual International Conference of the IEEE, Vol. 1, Nov 1994, pp.
528-529.

[80] Cesmeli, E., Powell, Kimerly A., and Comhill, J. Fredrick, “An Automated Method
To Detect Single-Leg Separation In Radiographic Images Of BjOrk-Shiley 60° And
70° Convexo-Concave Heart Valves,” Proceedings of the EMBC ’95, 1995 Annual
Conference of IEEE Engineering in Medicine and Biology Society, Vol. 1, 1995,
pp. 427-428.

[81] O’Neill, William, W., Chandler, James G., Gordon, Richard E., Bakalyar, Donovan
M., Abolfathi, Amir H., Castellani, Mark D., Hirsch, Jamie Lee, Wieting, David
W., Bassett, Joseph S., Beatty, Karen C., Soltis, Mary Ann, Timmis, Gerald G, and
Grines, Cindy L, “Radiographic Detection of Strut Separations in BjOrk-Shiley
Convexo-Concave Mitral Valves,” The New England Journal of Medicine, Vol.
333, No. 7, Aug 1995, pp. 414-419.

[82] Chandler, James G., Hirsch, Jamie Lee, O’Neill, William, W., Oesterle, Stephen N.,
Miller, D. Craig, Kennedy, John A., and Faichney, Alan, “Radiographic Detection
of Single Strut Leg Separations as a Putative Basis for Prophylactic Explantation of
BjOrk-Shiley Convexo-Concave Heart Valves,” World Journal of Surgery, Vol. 20,
No. 8, Oct 1996, pp. 953-960.

301

[83] Powell, Kimerly A., Obuchowski, Nancy A., Mueller, Klaus D., LaPresto, Eric,
Hwang, Charles G., Hirsch, Jamie Lee, Nissen, Steven E., and Cornhill, J. Fredrick,
“Quantitative Detection of Outlet Strut Separations in BjOrk-Shiley Convexo-
concave Mitral Valves,” Circulation, Vol. 94, 1996, pp. 32510-3256.

[84] Gilchrist, Ian G, Cardella, John F., Fox, Preston S., Pae, Walter E., Sabe, Ahmed R.
Kunselman, and Hopper, Kenneth D., “Radiographic detection of single-leg fracture
in BjOrk-Shiley Convexo-Concave prosthetic valves: A phantom model study,”
American Heart Journal, Vol. 133, 1997, pp. 197-202.

[85] Vrooman, Henri A., Maliepaard, Chris, Van der Linden, Leo, Jessurun, Emil R.,
Ludwig, J. Werner, Plokker, H. W. Thijs, Schalij, Martin J ., Weeda, Hans W. H.,
Laufer, Jaap L., Huysmans, Hans A., and Reiber, Johan H. C., “Quantitative
Assessment of the Presence of a Single Leg Separation in BjOrk-Shiley
Convexoconcave Prosthetic Heart Valves,” Investigative Radiology, Vol. 32, No. 9,
September 1997, pp. 540-549.

[86] Hopper, Kenneth D., Gilchrist, Ian C., Landis, J. Richard, Abolfathi, Arrrir H.,
Localio, A. Russell, Wilson, Ronald P., Pae, Walter E., Jr., Kunselman, Allen R.,
Wieting, David W., Grifﬁth, James W., Pierce, William S., Potok, Paul S.,
TenHave, Thomas R., and Chandler, James G., “In Vivo Accuracy Of Two
Radiographic Systems In The Detection Of BjOrk-Shiley Convexo-Concave Heart
Valve Outlet Strut Single Leg Separations,” Journal of Thoracic and
Cardiovascular Surgery, Vol. 115, Mar 1998, pp. 582-586.

[87] Brendzel, A., Rambod, E., Jorgensen, S. M., Reyes, D. A., Chmelik, M. S., and
Ritrnan, E. L., “Three-Dimensional Imaging of Fractures in Outlet Struts of BjOrk-
Shiley Convexo-Concave Heart Valves by Micro-Computed Tomography In-
Vitro,” Journal of Heart Valve Diseases, Vol. 11, 2002, pp. 114-120.

[88] Leilyn Perri, “Hershey Is The Primary Site In The Country TO Test Heart Valves,”
Ofﬁce of University Relations, Pennsylvania State University, November 7, 2000
(http://www.psu.edu/ur/2000/heartvalveshmc.html).

[89] Weber, Herman J. “Twenty-First Report of the Special Masters/Trustees,” Bowling-
Pﬁzer Litigation (case no. C-1-91-256), United States District Court, Southern
District of Ohio, Western Division, Nov 1, 2004, Appendix 2.

[90] Stein, P. D., Sabbah, H. N., Albert, D. E., and Snyder, J. E., “Spectral Signature of
the Opening Sound of the BjOrk-Shiley Convexo-Concave Valve as a Potential
Indicator of Strut Fracture,” American Journal of Noninvasive Cardiology, Vol. 1,

1987. pp. 369-372.

302

[91] Koymen, H., Altay, B. K., and Ider, Y. Z., “A study of prosthetic heart valve
sounds,” IEEE Transaction on Biomedical Engineering, Vol. 34, No. 11, 1987, pp.
853-863.

[92] Walker, David. K., Scotten, Lawrence N., “Discrimination in vitro between the
acoustic emissions from BjOrk-Shiley Convexo-Concave valves with and without a
broken minor strut,” Medical & Biological Engineering and Computing, Vol. 29,
1991, pp. 457-464.

[93] Walker, David K., and Scotten, Lawrence N., “Discriminant Analysis Of Acoustic
Emissions From Bjork-Shiley Heart Valves With And Without Broken Minor
Struts,” Proceedings of the Annual International Conference of the IEEE
Engineering in Medicine and Biology Society, 1991, Vol. 13, Nov 1991, pp. 2107-
2108.

[94] Plemons, Terry D., Hovenga, M., Reeder, H., Dow, J. J ., Inderbitzen, Rebecca, Chia,
Raymond, Wieting, David W., “Classiﬁcation of Bjork-Shiley convexo-concave
valve status by detection of the intact outlet strut resonant frequency,” Proceedings
of the 15th Annual lntemational Conference of the IEEE Engineering in Medicine
and Biology Society, Oct 1993, pp. 389—390.

[95] Nygaard, Hans, Inderbitzen, Rebecca, Hasenkam, J. Michael, Wieting, David W.,
Paulsen, P. K., “Measurement of sounds generated by mechanical aortic and mitral
heart valve prostheses,” Proceedings of the I 994 IEEE Seventh Symposium on
Computer-Based Medical Systems, J une 1994, pp. 55-60.

[96] Durand, Louis-Gilles, Stein, Paul D., Grenier, Marie-Claude, Henry, J. W.,
Inderbitzen, Rebecca S., and Wieting, David W., “In vitro and in vivo low
frequency acoustic analysis of BjOrk-Shiley convexo-concave heart valve opening
sounds,” Proceedings of the 1994 IEEE Seventh Symposium on Computer-Based
Medical Systems, June 1994, pp. 61-66.

[97] Eberhardt, Allen C., Chassaing, Charles E., Inderbitzen, Rebecca S., Wieting, David
W., “Acoustic evaluation of progressive failure in BSCC heart valves,” Proceedings
of the 1994 IEEE Seventh Symposium on Computer-Based Medical Systems, June
1994, pp. 112-118.

[98] Plemons, Terry, Schreck, Stefan G., and Inderbitzen, Rebecca S., “Stability of the
BSCC Heart Valve Outlet Strut Resonant Frequency Under Changing Physiological
Conditions and Observation Times,” Proceedings of the 1994 IEEE Seventh
Symposium on Computer-Based Medical Systems, June 1994, pp. 119-124.

[99] Plemons, Terry, Dow, J. J., Wieting, David W., and Chandler, James, G., “Multiple
Feature Acoustic Classiﬁcation of BSCC Artiﬁcial Heart Valves,” Proceedings of

the 1994 IEEE Seventh Symposium on Computer-Based Medical Systems, June
1994, PP. 125-130.

303

[100] Scarbrough, Kent, and Inderbitzen, Rebecca S., “Application of Time-Frequency
Analysis to the Classiﬁcation of BjOrk-Shiley Conxevo-Concave Heart Valve
Condition,” Proceedings of the I 994 IEEE Seventh Symposium on Computer-Based
Medical Systems, June 1994, pp. 131-138.

[101] Candy, J. V., and Jones, H. ~E., “Processing of prosthetic heart valve sounds for
single leg separation classiﬁcation,” Journal of the Acoustical Society of America,
Vol. 97, No. 6, June 1995, pp. 3663-3673.

[102] Candy, J. V., and Jones, H. E., “Classiﬁcation of prosthetic heart valve sounds: A
parametric approach,” Journal of the Acoustical Society of America, Vol. 97, No. 6,

June 1995. pp. 3675-3687.

[103] Sava, H. P., McDonnell, J. T. B., and Grant, P. M., “Spectral composition of
closing sounds produced by mechanical prosthetic heart valves,” Proceedings of
the 17h Annual Conference of the IEEE Engineering in Medicine and Biology
Society, 1995, Vol. 1, Sept 1995, pp. 125—126.

[104] Eberhardt, Allen C., Chassaing, Charles E., Ward, Mark A., and Lewandowski,
Simon J ., “Acoustic Characterization of Mechanical Valve Condition and Loading,”
Journal of Heart Valve Diseases, Vol. 4, NO. 6, Nov 1995, pp. 649-659.

[105] Durand, Louis-Gilles, Grenier, Marie-Claude, Inderbitzen, Rebecca S., Wieting,
David W., and Stein, Paul D., “Low Frequency Analysis of Opening Sound for
Detection of Single Leg Separation of BjOrk-Shiley Convexo-Concave Heart
Valves,” Journal of Heart Valve Diseases, Vol. 4, Suppl. I, 1995, pp. 832-837.

[106] Dow, J. J ., Plemons, Terry D., Scarbrough, Kent, Reeder, Hugh, Hovenga, Mike,
Wieting, David W., Chandler, James G., “Acoustic Assessment of the Physical
Integrity of Bjork-Shiley Convexo-Concave Heart Valves,” Circulation, Vol. 95,
1997, pp. 905-909.

[107] de Mol, Bas A., Cromheecke, Manon E., Groen, Janny G., Faber, Gerard, van der
Heiden, Maurits S. and Ongkiehong, Leo, “The Complexity of External Acoustic

Detection of Defects in BjOrk-Shiley Convexoconcave Heart Valves,” Artiﬁcial
Organs, Vol. 25, NO. 1, Jan 2001, pp. 63-67.

[108] Shawkat, S., Petersen, S. A., and Bailey, J. S., “A new technique for the assessment
of mechanical prosthetic heart valves: Analysis of their ultrasonic clicks,” in
Surgery for heart valve disease, Bodnar, E., editor, ICR Publishers, London, 1989,
pp. 96—104.

[109] Rambod, Edmond, Fatemi, M., Sahn, David J ., and Greenleaf, J. E, “An Innovative
Technique to Detect the Single Leg Separation (SLS) in the BjOrk—Shiley Convexo-

304

 

concave (BSCC) Mechanical Heart Valve,” Journal of the American College of
Cardiology, Vol. 37, 2001, p. 136.

[110] Rambod, Edmond, “Application of Novel Technologies in Detection of SLS in the
BSCC Heart Valves: VibrO-acoustography, X-Ray Micro-CT and Numerical
Modeling,” Report: BioQuantetics, Inc., Signal, CA, Nov 12, 2001. (Available
from the Bowling-Pﬁzer Document Repository: http://www.bowling-
pﬁzer.com/repository. Title = 020001.)

[111] Rambod, Edmond, “Step towards clinical implementation of an ultrasound burst-
spectrography technique for detection Of single leg separation (SLS) in cycling
BjOrk-Shiley Convexo-Concave (BSCC) Heart Valves,” a research proposal
submitted to the Bowling-Pﬁzer Supervisory Panel, Jan 21, 2003. (Available from
the Bowling-Pﬁzer Document Repository: http://www.bowling-
pﬁzer.com/repository. Title=240008.)

[112] van Neer, P. L. M. J ., Bouakaz, A., Vlaanderen, E., de Hart, J., van de Vosse, F. N.,
van der Steen, A. F. W., de Jong, N., “In-vitro development Of a novel method to
noninvasively establish BScc valve integrity,” Proceedings of the 2005 IEEE
lntemational Ultrasonic Symposium, Rotterdam, The Netherlands, Sept 18-21,
2005.

[113] Lepelaars, Eugene S. A. M., van Ooijen, Willem D. R., T ijhuis, Anton G., and de
Mol, Bas A. J. M., “Detecting Fractures in Thin-Wire Structures,” Proceedings of
the 18’” Annual lntemational Conference of the IEEE Engineering in Medicine and
Biology Society, Amsterdam, 1996, pp. 159-160.

[114] Lepelaars, Eugene S. A. M., van Ooijen, Willem D. R., Tijhuis, Anton G., and de
MOI, Bas A. J. M., “Detecting fractures in artiﬁcial heart valves,” Digest of the
IEEE Antennas and Propagation Society lntemational Symposium, 1997, Vol. 2,
July 1997, pp. 1198-1201.

[115] Lepelaars, Eugene S. A. M., “Transient Electromagnetic Excitation of Biological
Media by Circular Loop Antennas,” Ph.D. Thesis, Eindhoven University of
Technology, Eindhoven, The Netherlands, 1997, Ch. 8.

[116] Lepelaars, Eugéne S. A. M., van Ooijen, Willem D. R., and Tijhuis, Anton G., “On
the Feasibility of Detecting Flaws in Artiﬁcial Heart Valves,” IEEE Transactions
on Microwave Theory and Techniques, Vol.48, Issue 11, Nov 2000, pp. 2165-2171.

[117] Tijhuis, Anton G. and Visser, Hubrecht J ., “Development of a Catheter-Based
Antenna for Detecting Flaws in BjOrk-Shiley Artiﬁcial Heart Valves,” a research
proposal submitted to the Bowling-Pﬁzer Supervisory Panel, Feb 27, 2003,
Schedule A. (Available from the Bowling-Pﬁzer Document Repository:
httpzl/www.bowling-pﬁzer.com/repository. Title = 240001.)

305

 

[118] Amin, Viren, “Development and Evaluation of Differential Eddy Current Method
for Detection of Outlet Strut Fracture in BjOrk-Shiley Valves. Progress Report: First
animal test using differential eddy current approach,” a report submitted to the
Bowling-Pﬁzer Supervisory Panel, March 3, 2004. (Available from the Bowling-
Pﬁzer Document Repository: http://www.bowling-pﬁzer.com/repository. Title 2
290039.)

[119] Udpa, Satish S., “New Electromagnetic Methods for the Evaluation of Prosthetic
Heart Valves,” Journal of Applied Physics, Vol. 91, NO. 10, 2002, pp. 1-5.

[120] Nair, Naveen V., Ramakrishnan, Sridhar, Clifford, Robert, Li, Yue, Majumdar,
Shantanu, Chan, Shiu C., Ramuhalli, Pradeep, Udpa, Lalita, and Udpa, Satish S., “A
Beat Frequency Electromagnetic Acoustic Transduction Technique for Detecting
Strut Fractures in Prosthetic Heart Valves,” (Proceedings of the 10th lntemational
Workshop on Electromagnetic Nondestructive Evaluation, East Lansing, Michigan,
June 1-2 2004), Electromagnetic Nondestructive Evaluation (IX), Burke, VA: IOS
Press, Inc., 2004.

[121] Ramakrishnan, Sridhar, Nair, Naveen V., Clifford, Robert, Majumdar, Shantanu,
Chan, Shiu C., Li, Yue, Ramuhalli, Pradeep, Udpa, Lalita and Udpa, Satish S., “An
Electromagnetic Acoustic Transduction Technique for Detecting Strut Fractures in
Artiﬁcial Heart Valves,” Proceedings of the IEEE Electra/Information Technology
(EIT) Conference 2005, Lincoln, NE, May 22-25, 2005.

[122] Chan, Shiu C., Li, Yue, Udpa, Lalita, and Udpa, Satish S., “Electromagnetic
Techniques for Detecting Failures in Artiﬁcial Heart Valves” (Proceedings of the
8th lntemational Workshop on Electromagnetic Nondestructive Evaluation,
Saarbriicken, Germany, 2002), Electromagnetic Nondestructive Evaluation (VII),
Kojima, F., Takagi, T., Udpa, Satish S., and Pavo, J., Editors, Burke, VA: 108
Press, Inc., 2003.

[123] Chan, Shiu C., Clifford, Robert, Majumdar, Shantanu, Nair, Naveen V.,
Ramakrishnan, Sridhar, Li, Yue, Ramuhalli, Pradeep, Udpa, Lalita, and Udpa,
Satish S., “Novel methods for detecting fractures in prosthetic heart valves,” Insight
(Journal of the British Institute of Non-Destructive Testing), Vol. 47, No. 1, Jan
2005, pp. 15-19. (Paper was also presented at the 16th World Conference on
Nondestructive Testing, Montreal, Canada, Aug 30-Sep 3, 2004.)

[124] Chan, Shiu C., Oka, Mohachiro, Downey, Adam, and Udpa, Satish S., “Minimally
Invasive System for the Evaluation of Prosthetic Heart Valves” (Proceedings of the
9th lntemational Workshop on Electromagnetic Nondestructive Evaluation, Saclay,
Paris, France, May 2003), Electromagnetic Nondestructive Evaluation (VIII),
Sollier, Thierry, Prémel, Denis, and Lesselier, Dominique, Editors, Burke, VA: IOS
Press, Inc., 2004, pp. 247-254.

306

 

[125] Nondestructive Testing Handbook, 3rd Edition, Volume 5: Electromagnetic
Testing, Satish S. Udpa (technical editor) and Patrick 0. Moore (editor), American
Society for Nondestructive Testing: Columbus, OH, 2004.

[126] Blitz, Jack, Electrical and Magnetic Methods of Nondestructive Testing, 2nd
Edition, Springer-Verlag: New York, Nov 1997.

[127] Wieting, D. W., Breznock, E. M., Kafesjian, R., and Stobie, R., “Exercising sheep:
An in vivo model for assessing the durability of pyrolytic carbon heart valves,” The
lntemational Journal of Artificial Organs, Vol. 13, NO. 9, 1990, p. 608.

[128] Breznock, E., Pendray, D., Kinoshita, G., Wieting, D., Kafesjian, R., and Litzman,
R., “The Ovine Model for In Vivo Chronic Mitral/Tricuspid Valve Studies,”
Journal of the American Society for Artiﬁcial Intemal Organs (ASAIO), Vol. 42,
1997,p.27.

[129] Nave, C. Rod, HyperPhysics, Georgia State University, 2005,
http://hype rphysics.phy-astr. gsu. edu/hbase/magnetic/helmholtz. html

[130] Biddlecombe, C., Heighwar, B., Simkin, J. and Trowbridge, C., “Methods For
Eddy Current Computation in Three Dimensions,” IEEE Transactions on
Magnetics, Vol. 18, No. 2, 1982, pp. 492-497.

[131] Salon, S. and Peng, J ., “3-D Eddy Current Using a Four Component Finite Element
Formulation,” IEEE Transactions on Magnetics, Vol. 20, No. 5, Sept, 1984, pp.
1992-1994.

[132] Bardeen, James M. and Press, William H., “Radiation ﬁelds in the Schwarzschild
background,” Journal of Mathematical Physics, Vol. 14, No. 7, 1973 pp. 7-19.

[133] Fleiss, Joseph L., Statistical Methods for Rates and Proportions, 2'"! Edition, New
York: John Wiley & Sons, 1981, pp. 1-18.

[134] Woolson, Robert F ., and Clarke, William R., Statistical Methods for the Analysis of
Biomedical Data, 2"“ Edition, New York: John Wiley & Sons, 2002.

[135] Patient UK, Understanding the heart,
http://www. patient. co. uk/showdoc/2 7000350/.

[136] Texas Heart Institute at St. Luke’s Episcopal Hospital, Heart Information,
http://www.tmc.edu/thi/valves.html.

[137] Merck & CO., Inc., The Merck Manual of Medical Information, 2'“ Home Edition,
Ch. 28 (Online version: http://www.merck.com/mmhe/sec03/ch028/ch028a.html).

307

 

[138] Cedars-Sinai Medical Center Division of Cardiology, Cedars-Sinai Medical Center
Prosthetic Heart Valve Information. (http://www.csmc.edu/pdf/Heart_Valves.pdf)

[139] Morse, D., Steiner, R.M., Fernandez, J ., Guide to Prosthetic Valves, Springer-
Verlag, New York, 1985.

[140] Baxter-Edwards Museum of Heart Valve Design.

[141] Evansville Heart Center, Evansville, IN, Image gallery/Prosthetic Heart Valve
Gallery, http://www.evansvilleheartcenter.com

[142] St. Jude Medical, Inc., Toronto SPV® Valve,
http://www.sjm. com/devices/device.aspx ?name= Toronto+SP V%26%23 I 74 %3b+ V
alve&location=us&type= I 9.

[143] Kouchoukos, N.T., Davila-Roman, V.G., Spray, T.L., Murphy, SF, and Perrillo,
J .B., “Replacement of the Aortic Root with a Pulmonary Autograft in Children and
Young Adults with Aortic-Valve Disease,” New England Journal of Medicine, Vol.
330, No. 1, Jan 6, 1994, pp. 1-6.

[144] Kincaid, Edward H., Atala, Anthony, and K011, Neal D., “Tissue Engineered Heart
Valves: A Potential Cure for Valvular Heart Disease,” Cardiothoracic Surgery
Network
(http://www.ctsnet.org/sections/innovation/valvetechnology/articles/article-
12.htrnl).

[145] Elkins, Ronald C., Dawson, Patti B., Goldstein, Steven, Walsh, Steven P., Black,
Kirby S., “Decellularized human valve allografts,” The Annuals of Thoracic
Surgery, NO. 71, 2001, pp. $428-$432.

[146] Bechtel, J. F., Muller-Steinhardt, M., Schrrridtke, C., Brunswik, A., Stierle, U.,
Sievers, H. H., “Evaluation of the decellularized pulmonary valve homograft
(SynerGraft),” Journal of Heart Valve Diseases, NO. 12, 2003, pp. 734-739.

[147] Guan, J ., Fujimoto, K. L., Sacks, M. S., Wagner, W. R., “Preparation and
characterization Of highly porous, biodegradable polyurethane scaffolds for soft
tissue applications,” Biomaterials, Vol. 26, 2005, pp. 3961-3971.

[148] Connolly, J. M., Alferiev, 1., Clark, J. N., Eidelman, N., Sacks, M. S., Palmatory,
E., Lu, 2., Kronsteiner, A., DeFelice, S., Xu, J ., Ohri, R., Narula, N., Vyavahare, N.
and Levy, R. J., “Triglycidyl amine crosslinking of porcine aortic valve cusps or
bovine pericardium results in improved biocompatibility, biomechanics, and
calciﬁcation resistance: Chemical and Biological mechanisms,” American Journal
of Pathology, Vol. 166, NO. 1, 2005, pp. 1-13.

308

 

[149] Engelmayr, G. C., Jr., Rabkin, E., Sutherland, F. W. H., Schoen, F. J ., Mayer, J. E.,
Jr. and Sacks, M. S., ”The independent role of cyclic ﬂexure in the early in vitro
development of engineered heart valve tissue,” Biomaterials, Vol. 26, No. 2, 2005,
pp 175-187.

[150] Personal conversations with Dr. Donald C. Harrison of the Bowling-Pﬁzer
Supervisory Panel (summer, 2003) (http://www.bowling-
pﬁzer.com/supervisor.htm).

[151] Pﬁzer Incorporated (formerly Shiley Laboratories, Incorporated), Bjdrk-Shiley
Convexo-Concave Prosthetic Heart Valve Technical Schematics, New York: NY,
drawing no.: Dp02-0119, Rev. E. document no.: 809018449.

[152] Pﬁzer Incorporated (formerly Shiley Laboratories, Incorporated), ij’rk-Shiley
Convexo-Concave Prosthetic Heart Valve Technical Schematics, New York: NY,
drawing no.: Dp02-0120, Rev. K. document no.: 809018497.

[153] Pﬁzer Incorporated (formerly Shiley Laboratories, Incorporated), ij'rk-Shiley
Convexo-Concave Prosthetic Heart Valve Technical Schematics, New York: NY,
drawing no.: Dp02-0121, Rev. D. document no.: 809018566.

[154] Cardima, Incorporated, On-line Product Catalog, Naviport® guide catheter product
web page (http://www.cardima.com/company/products/products_naviport.html)

[155] Tian, Yong, Udpa, Satish S., Nair, Naveen, and Ramakrishnan, Sridhar, “Modeling
of Electromagnetic Heating Effects During In Vivo Testing of Prosthetic Heart
Valves,” Proceedings of the 2006 IEEE lntemational Magnetics Conference
(INT ERMAG 2006), San Diego, CA., May 8-12, 2006.

[156] Pennes, Harry H., “Analysis of Tissue and Arterial BlOod Temperatures in the
Resting Human Forearm,” Journal of Applied Physiology, Vol. 1, 1948, pp. 93-122.

[157] Wissler, Eugene H., “Pennes’ 1948 paper revisited,” Journal of Applied
Physiology, Vol. 85, 1998, pp. 35-41.

[158] Tangwongsan, Q, Will, I. A., Webster, J. G., Meredith, K. L. Jr., and Mahvi, D.
M., “In Vivo Measurement of Swine Endocardial Convective Heart Transfer

Coefﬁcient,” IEEE Transactions on Biomedical Engineering, Vol. 51, 2004, pp.
1478-1486.

[159] Haynes lntemational, Inc., HA YNES® 25 Alloy ( h305 7), 2004.

[160] MatWeb Material Property Data website, http://www.matweb.com.

[161] Bronzino, Joseph D., editor-in-chief, The Biomedical Engineering Handbook, 2nd
ed., Vol. II, CRC Press: Boca Raton, FL, 2000.

309

 

[162] Bronzino, Joseph D., editor, The Biomedical Engineering Handbook, 3rd ed.,
Tissue Engineering and Artiﬁcial Organs, CRC Press: Boca Raton, FL, 2006.

[163] IEEE Std C95.1-1982, ANSI Standard for Safety Levels with Respect to Human
Exposure to Radio Frequency Electromagnetic Fields, 3 kHz to 100 GHz.

[164] IEEE Std C95.], 1999 Edition, IEEE Standard for Safety Levels with Respect to
Human Exposure to Radio Frequency Electromagnetic Fields, 3 kHz to 300 GHz.

[165] IEEE Std C95.1b-2004, IEEE Standard for Safety Levels with Respect to Human
Exposure to Radio Frequency Electromagnetic Fields, 3 kHz to 300 GHz—
Amendment 2: Speciﬁc Absorption Rate (SAR) Limits for the Pinna.

[166] IEEE Std C95.1-2005, IEEE Standard for Safety Levels with Respect to Human
Exposure to Radio Frequency Electromagnetic Fields, 3 kHz to 300 GHz.

[167] IEC 60601-2-33, IEC Standard Medical Electrical Equipment—Part 2-33:
Particular Requirements for the Safety of Magnetic Resonance Equipment for
Medical Diagnosis, Second Edition, 2002.

[168] IEC 60601-2-33, IEC Standard Medical Electrical Equipment—Part 2-33:
Particular Requirements for the Safety of Magnetic Resonance Equipment for
Medical Diagnosis (including Amendment 1), Consolidated Edition 2.1, 2006.

[169] IEC 60601-1, IEC lntemational Standard Medical Electrical Equipment—Part 1:
General Requirements for Basic Safety and Essential Performance, Edition 3.0,
2005.

[170] IEC 60601-1-2, IEC lntemational Standard Medical Electrical Equipment—Part 1-
2: General Requirements for Safety—Collateral Standard: Electromagnetic

Compatibility—Requirements and Tests (include Amendment 1), Edition 2.1, 2005.

[171] IEC 60601-l-SER, IEC lntemational Standard Medical Electrical Equipment—All
Parts, Edition 1.0, 2006.

[172] ICNIRP (lntemational Corrnnission on Non-Ionizing Radiation Protection),
“Guidelines for Limiting Exposure to Time-Varying Electric, Magnetic, and
Electromagnetic Fields (Up to 300 GHz),” Health Physics, Vol. 74, No. 4, April,
1998, pp. 494-522.

[173] Zaremba, Loren A., “FDA Guidelines for Magnetic Resonance Equipment Safety,”

44th Annual Meeting of the American Association of Physicists in Medicine
(AAPM), Montreal, Quebec, Canada, July 14-18, 2002.

310

 

[174] Zaremba, Loren A., “(Abstract) FDA Guidelines for Magnetic Resonance
Equipment Safety,” Medical Physics, Vol. 29, No. 6, June 2002, pp. 1302.

[175] Zaremba, Loren A., “Criteria for Signiﬁcant Risk Investigations Of Magnetic
Resonance Diagnostic Devices,” US. Food and Drug Administration, Rockville,
MD, July 14, 2003.

[176] Javor, Edware R. and Anderson, Theodore, “Design of a Helmholtz Coil for Low
Frequency Susceptibility Testing,” 1998 IEEE lntemational Symposium on
Electromagnetic Compatibility, Vol. 2, Aug 1998, pp. 912-917.

[177] Anderson, Theodore, “Design of a Helmholtz coil for susceptibility testing using
variational calculus and experimental veriﬁcation,” 1999 IEEE lntemational
Symposium on Electromagnetic Compatibility, Vol. 2, Aug 2-6, 1999, pp. 601-604.

[178] see [129]

[179] Chan, Shiu C., Oka, Mohachiro, and Udpa, Satish S., “A Minimally Invasive
Electromagnetic Technique for the Evaluation of Prosthetic Heart Valves,” Paper
Summaries of the ASNT 16th Annual Research Symposium, The American Society
for Nondestructive Testing, Wyndham Orlando Resort, Orlando, FL, March 26-30,
2007.

[180] Zeng Z., Xuan L., Sun Y., Udpa L., and Udpa S., "Probability of detection model
for gas transmission pipeline inspection," Research in Nondestructive Evaluation,
Vol. 15, No. 3, July-Sep. 2004, pp. 99-110

[181] Zeng, Zhiwei, Applications of POD Studies and Robust Design to Electromagnetic
NDE, Ph. D. Dissertation, Iowa State University, Ames, IA, 2003.

[182] Rajesh, S. N., Udpa, L., and Udpa, S.S., “Numerical Model Based Approach for
Estimating Probability of Detection in NDE Applications,” IEEE Transactions on
Magnetics, Vol. 29, No. 2., March 1993, pp. 1857-1860.

[183] Lawless, J. F., Statistical Models and Methods for Lifetime Data, New York, New
York: John-Wiley and sons, 1982.

[184] Grivich, Matthew, 1., and Jackson, David P., “The Magnetic Field of Current-
Carrying Polygons: An Application of Vector Field Rotations,” American Journal
of Physics, Vol. 68, No. 5, May 2000, pp. 469-474.

[185] Blade], J. Van, Electromagnetic Fields, New York, New York: McGraw-Hill Book
Company, 1964, Chapter 6.

311

[186] MAGNET FORMULAS web site
(http://www.netdenizen.com/emagnet/Offaxis/iloopoffaxis.htm)

[187] Montgomery, D. Bruce, and Terrell, J ., “Some Useful Information for the Design
of Air-Core Solenoids,” published under Air Force Contract AF 19(604)-7344.

[188] Abramowitz, Milton, and Stegun, Irene, Handbook of Mathematical Functions,
New York, New York: Dover Publications, 1964, Chapter 17.

[189] Hancock, Harris, Lectures on the Theory of Elliptic Functions, New York, New
York: J. Wiley and sons, 1910.

[190] eHealthMD.com web site
(http://www.ehealthmd.com/library/heartattack/HA_whatis.htrnl).

[191] Rafael, C. Gonzalez, and Wintz, Paul, Digital Image Processing, 2nd Ed., Addison
Wesley publishing, 1987, pp. 360-361.

[192] Mandayam, S., Udpa, L., Udpa, S. S., and Lord, W., “Monte Carlo Methods for
Modeling Magnetostatic NDE Phenomena: A Feasibility Study,” IEEE
Transactions on Magnetics, Vol. 32, No. 3, May 1996, pp. 1425-1428

[193] Zhang, Y., Zhang, Y., Udpa, L., and Udpa, S. S., “Probability of detection model
for Pipeline inspection,” Review of Progress in Quantitative Nondestructive

Evaluation, Vol. 16, pp. 1307-1314, 1997

[194] Personal meeting with Dr. James Thomas, Section Head, Cardiovascular Imaging,
Heart and Vascular Institute, Cleveland Clinic, on October 16, 2003.

[195] Personal correspondence with Dr. Donald Moore, Division of Pediatric Cardiology,
Vanderbilt University School of Medicine, Nashville, TN, on March, 2004.

312

 

   

11111111111111111