. . l. .. 7.27.?“ n. g .22.». .. . .: . . . . draws. maul . Wm??? . . . . '2 .. t . . .. .7 u u... wfiflv aw .. *p‘cfl unstab- nflvn EL. . .ny fifiwfii gawk; 2%, . g or} . .V any. ~ 0... 254...! s. ”Mam. ”fit-in? . A 4 - ‘ D. ‘ Eirznbbfiud . . . v I J “I' t v .. .cfl ‘ 1 .f QLIJ. I." ‘ . . -I’\ . I“ .r .. l v! ‘Vu‘n .8; SJ“ . . . ‘ Imoxtont . . y ..c.. .Ov afiwMH-HJ...‘ “WWW“ o1»-.. . . x .. . , .ohunwmmumdfivurr...“ . v.1..x..._.b.n1.-,; f . V Divoaoflr. .9 Jr“; .1.“ . . v “swam”. 21-min). WWW“. I‘ 1 u I «4 V (I .. t. I ‘ . .. ‘ rmvnw.‘ 1...»...r ‘ . , . I3k»¢sufii1m%fivflnfiwtvh¥ . .iWcfi .. ‘ntlhzfv . . . ‘ q ‘ r h. 1.3.. ”fill... .. 3|: 4...“. I. IV 11:»? .90 .l., . till-t .0»... ‘cltuyil “830.... v .55...“ ‘Ifli - i 1‘.\‘.... .wilt‘oi'nluuooJ ll ‘9 u r 1...“-.1! I. L.‘.‘l.u..ilv. 3o. ‘13 “If (HUN-Vat. .u...,.‘ II a. , £11k ll . (n -lv Indut‘lr . I . (Pickl- Drfé , ".598 IlllllllllllllllllllHilllllllllllllllllllll|l|l||||l|lllllll 31293 015819026 LIBRARY Mlchigan State University This is to certify that the thesis entitled BIOMECHANICAL ASSESSMENT OF HTR AS A PREMOLDED CRANIOPLASTY IMPLANT MATERIAL presented by Michael Lynn Esch has been accepted towards fulfillment of the requirements for M 5 degree in M SM fight/545374 Dr. Roger Haut Major professor Date 12/5/96 07639 MS U is an Affirmative Action/Equal Opportunity Institution PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. T——__——_—_————'T DATE DUE DATE DUE DATE DUE L___“ l.— IEII l- L—‘ll l ' l —T___l LELA- MSU le An Affirmative Action/Equal Opportunity Institution m1 BIOMECHANICAL ASSESSMENT OF HTR" AS A PREMOLDED CRANIOPLASTY IMPLANT MATERIAL By “ Michael Lynn Esch A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Materials Science and Mechanics 1996 ABSTRACT BIOMECHANICAL ASSESSMENT OF HTRdb AS A PREMOLDED CRANIOPLASTY IMPLANT MATERIAL By Michael Lynn Esch The influence of bone and soft tissue growth into a porous implant material was demonstrated with mathematical analysis and biomechanical testing. Fifteen premolded cranioplasties were performed on eight laboratory beagles. Nine of these implants were made of HTR", a commercially available porous PMMA/PHEMA composite, the remaining six were made from Cranioplastic®, solid PMMA. These implants were retrieved and biomechanically tested to determine the effect of biological fixation on flexure rigidity of the device after a minimum of 6 months implantation. Theoretical models were used to demonstrate a 230% increase in rigidity when the implant margin is rigidly fixed by bony tissue. Histological samples from the bone implant interface demonstrated more soft tissue than bone growth into the porous structure. Material property comparisons of pre- and post-implanted HTR devices showed no significant increase in Young’s Modulus or ultimate tensile strength. The HTR" material did not demonstrate the expected clinical osseoconductivity and subsequent material property increase in this surgical application. Dedication Page This research is dedicated to my family who have offered the love, support and encouragement to complete the program. iii ACKNOWLEDGMENTS US Surgical Corporation for funding this research. Drs. Andy Shores, PhD., DVM, Dermot Ievens, DVM, Charlie DeCamp, DMV for their surgical work in the clinical phase of the research. Jane Walsh for her histology work iv TABLE OF CONTENTS LIST OF TABLES ............................................. vii LIST OF FIGURES ........................................... viii INTRODUCTION / PROJECT OVERVIEW ......................... 1 DESIGN CRITERIA ............................................ 5 Cranial Vault Anatomy ..................................... 6 Surgical Practices for Filling Cranial Defects ................... 11 Forms of Cranial Vault Failure .............................. 12 Clinical Manifestations and Sources of Cranial Vault Failure .............................. 14 Properties of Porous HTRo ................................. 16 EXPERIMENTAL PROCEDURES ................................ 21 Implant Design .......................................... 22 Mechanical Testing ....................................... 25 Three - Point Beam Bending Specimens ...................... 28 Analysis of Full Implant Geometry under Quasi-static Load .............................. 29 Analysis of Harvested Shell Specimens ....................... 30 Histology .............................................. 3 1 MATHEMATICAL MODELING ................................. 32 Three Point Beam Bending Model ........................... 33 Plate Model ............................................ 36 Finite Element Analysis ................................... 41 Onlay Model ...................................... 42 Inlay Model ....................................... 44 RESULTS .................................................. 45 Three-Point Beam Bending of HTR ' Prior to Implantation ................................ 46 Full Implant Model Prior to Implantation ..................... 47 Three Point Beam Bending of HTR After Implantation .................................. 51 Strength of Materials Parametric Models ...................... 53 Finite Element Analysis of the HTR Implant ................... S4 Histological Analysis of the Retrieved HTR and Cranioplastic Implants ....................... 59 Gross Morphology Observations ....................... 59 Microscopic Observations ............................ 63 CONCLUSION .............................................. 7 1 APPENDIX A ................................................ 74 Surgical Protocol ......................................... 74 BIBLIOGRAPHY ............................................. 78 vi Table 1 Table 2 Table 3 Table 4 Table 5 Table 6 Table 7 LIST OF TABLES Critical Cranial Bone Properties ........................... 10 Cranioimplant material by animal. ......................... 24 Material Properties of Non-Implanted HTR Beam Samples in 3-Point Bending Analysis ...................................... 46 Mechanical Properties of HTR Based on Plate Geometry in Quasi- Static Bending Analysis ................................. 48 Harvested HTR Implants in 3-point Beam Bending ............ 52 Material Properties used in FEM ........................... 55 Results from FEA Onlay and Inlay Models ................... 57 vii LIST OF FIGURES Anatomy of Human Cranial Vault ......................... 7 Location of a Temporal/Parietal Craniotomy on a Canine Skull . . 7 Graphic Representation of Canine Calvaria in Cross Section ..... 8 Failure Modes Related to Impact Area .................... 13 Distribution of Linear and Depressed Fractures Related to Cranial Insult“°’ ............................................ 14 Types of Head Injury Lesions ............................ 15 Micrograph of HTR“ (25X). Note contact between spheres . . . . 19 Detailed Implant Geometry ............................. 23 Test Set Up Used for Beam and Plate Analysis (Plate Shown) . . . 27 Final Test Coupon Geometry ............................ 28 Symmetrical Bending of Flat Plate about Neutral Axis. ........ 39 Finite Element Model -- Initial Constraints ................. 43 Typical Output from Three Point Beam Bending Test ......... 45 Failed HTR Beam Sample in Three Point Beam Bending ....... 47 Failure mode of Implant Geometry under Central Load ........ 49 Stress-Strain Plot of HTR Plate Sample .................... 50 Comparison of Inlay and Onlay FEM Models ............... 56 Strains (ex and 62) associated with the caudal surface of the ”inlay" implant ............................................. 58 Caudal Surface of HTR Explant .......................... 6O Interface of Cranioplastic Cranioimplant ................... 6O Bone/Implant Interface of Resected HTR Cranioimplant ....... 61 Poor Fit of Implant in Craniotomy ........................ 62 Micro Air Bubbles in Cranioplastic Material ................ 62 Soft Tissue Penetration to Approximately 3 Bead Depth ....... 63 Soft Tissue in Contact with Cranioplastic Material but Not Adhered ............................................ 64 viii INTRODUCTION / PROJECT OVERVIEW Current surgical practices for filling large cranial defects include the use of autograft, allograft and alloplastic implants. The mechanical properties of materials used for such procedures must be capable of withstanding the normal physiological demands placed on the skull, as well as demonstrating sufficient biocompatability and ease of surgical implantation. Mechanical failure of a cranioimplant may lead to catastrophic clinical consequences. Hard Tissue Replacement, HTR”, is a porous acrylic composite made from two commonly used and commercially available bioplastics. The material is fabricated in a micro-beaded fashion consisting of an inner core of polymethylmethacrylate (PMMA) and an outer coating of polyhydroxylethylmethacrylate (PHEMA). Previous research has shown that in combining these two materials, the complex is biocompatible, nontoxic, 'Product marketed by US Surgical Corporation, Norwalk, CT. 1 2 noninflammatory, noncarcinogenic, and facilitates bone and soft tissue integration('”"'5’*. HTR" was developed and patented by Drs. Arthur Ashman and Paul Bruins("7'8'°’ in the early eighties and made commercially available for dental applications by United States Surgical Corporation, Norwalk, Connecticut. Ashman, a maxillofacial surgeon, has published several papers on successful dental procedures utilizing HTR" as a maxillofacial bone grafting material. These surgical procedures have involved the use of particulate HTR“ in bone grafting following a tooth extraction and for general reconstruction of the alveolar ridgem'”. Evaluation of alternate biomaterials for use as cranioimplants involve a review of laboratory and clinical properties of the candidate material, as well as a definition of success/failure criteria. The critical factors (e. g. ultimate tensile strength, Young’s Modulus, fracture toughness, immunotoxogenesity, biocompatability, etc.) associated with this type of implant are very broad. This investigation is designed to explore several of them. Numbers in parenthesis denote references in bibliography. 3 The hypothesis of this research was that bone will grow into the pores of HTR, strengthen the device and stabilize the implant at the implant bone interface. This research will attempt to show the relationship of geometrical rigidity to the ingrowth of bone and soft tissue into the pores of the HTR material via three point bending analysis of pre- and post-implanted devices. The intent of this research was to assess the biomechancial efficacy of using HTR as a premolded cranioimplant material. A brief summary of work follows. Maw Prior research on HTR has not determined the material properties when it is molded into an implant geometry. It was an objective of this research to determine the change in mechanical properties after a period of 6 months implantation. A simplified beam model, sectioned from the central third of the implant and tested in a three point bending fixture, was used to determine the material properties. WW Strength of material’s models were developed to demonstrate the effect of bone growth into the HTR boundary at the bone/implant interface. Due to the flatness of the implant geometry, a simplified flat plate analysis was used to determine the effect of boundary conditions on the rigidity of the implant. 4 MW Finite Element Modelling (FEA) techniques were used to determine the influence of specific geometrical features on the implant. In the design phase, a “positioning” feature was added to the implant design in to aid in implantation. This geometrical feature added a sharp corner at the implant bone interface. The FEA models were used to determine the stress concentration related to this feature. Two models, were developed and compared to determine the difference in stress distribution. Wm Ten rigidly supported samples of the non-implanted HTR cranioimplants were loaded to failure under a central load. The failure mechanism was typical of a simply supported plate under a central load. The fracture planes originated under the load applicator and “bisected” the 4 edges of the implant geometry. 315191935 The two outer thirds of the fifteen explanted devices were used for histological analysis. The use of HSLE stain will label the type of biological reaction to the HTR and PMMA implants. These preparations will be used to the osseoconductive properties of the materials. DESIGN CRITERIA The primary function of a cranioimplant is the surgical replacement of cranial bone. According to Rawlings‘lo’, the ideal cranioimplant material would: 1) capable of being vascularized for osseointegration, 2) light in weight, 3) traceable on x-ray, 4) similar thermal expansion properties as the host tissue, 5) physically stable (nonionizing, noncorrosive, nonbiodegradable, etc.), 6) inert and biocompatible, 7) aesthetically pleasing, 8) similar biomechanical properties as the skull section it is to replace, 9) easily shaped by the surgeon, IO) relatively inexpensive, and 11) readily available clinically. In short, the cranioimplant material must satisfy biocompatability, biomechanical strength and surgical procedure constraints. HTR“ has been shown by previous researchers and clinicians‘1'2'3'4'5’ to satisfy the last two of these three constraints, but has not been evaluated against the biomechanical requirements of a cranioimplant material. To determine the required biomechanical properties, HTR1D will be compared against Cranioplastic", a currently available PMMA used by clinicians to repair cranial defects. The 6 comparison of HTR and Cranioplastic will yield the clinical difference between the two materials in this application. Cranial Vault Anatomy The main function of the eight cranial bones which form the human cranial vault, Figure 1, are to enclose and protect the brain and the organs of sight, hearing and balance”‘"”. Failure of the braincase is determined to be either fracture of the cranial bone or excessive deflection of the bone which causes brain impingement (a form of closed head injury). This research will be limited to the assessment of large cranial bone defects of the temporal/parietal region of the canine skull (Figure 2). fit. ,1 Figure 1 Anatomy of Human Cranial Vault .41., Ji-tm Jia “is i ’ ' ' ‘1 1' ‘ Figur 2 Location of a Temporal/Parietal Craniotomy on a Canine Skull 8 The cross sectional structure of cranial bone shown in Figure 3, has been compared to an "engineering sandwich structure'm'”). The bones consist of stiff inner and outer layers of cortical (compact) bone which "sandwich" between them the softer porous diploe layer. In general, the inner and outer tables offer the stiffness of the structure while the porous middle layer is the major source of the blood supply and nutrients for the bone. In the canine model there is great variation in the thickness of the cranial bone. The skull bones in the region of the external sagittal crest are thicker due to a thicker diploe layer. The overall thickness of the canine skull thins as the diploe layer Cortical Bone AAA A: . A .5." AAAAAAAAAAtAAAAAAAiA A A A A J AA A A A; AAAAAAAAAAAAAAAA AAAAAA AA AAA 1% A‘.‘ Figure 3 Graphic Representation of Canine Calvaria in Cross Section 9 thins laterally from the crestus’ (Figure 3). Oklund et. al."” observed that unfilled craniotomies measuring 17 mm (.675 inches) in diameter in the canine skull showed greater bony growth in the areas of the thicker diploe regions. This difference in thickness results in a great variation in the strength of the cranial bones and ultimately affects the rate“) and extent of bony fixation of porous cranial implants. In the human skull there are variations in thickness‘”’, but they are not as extreme as seen in the canine model”). The average human skull thickness of an adult male (Figure 1) is 7.0 mm (0.272 inches) and the average diploe layer is 2.7 mm (0.108 inches)‘”’. Gurdjian‘m has shown the thinner cross section cranial bones have an outer table that is thinner than the inner cortical bone table. The amount of static stress required to cause fracture of the human cadaveric skull, with the scalp covering intact, was reported by Gurdjian‘m to range from 3.00 MPa (435 psi) to over 6.20 MP3 (900 psi). Similarly, he reported the fracture strength of the uncovered, relatively dry skull to average 0.28 MPa (40 psi). The ability of the hair and scalp to distribute the load on the cranial bone results in an approximate 10 fold increase in static load carrying capacity. Woodm’, from research utilizing strain gages on cadaveric cranial bones, found the modulus of elasticity to be 12.4 GPa (1.8 X 106 psi) and the 10 temporal/parietal region to have an average ultimate tensile strength (UTS) of 68.9 MPa (10,000 psi). Oklund‘m found the yield stress (oyidd) and ultimate tensile strength of the uncovered canine skull to be 21.7 MPa (3,150 psi) and 25.1 MPa (3,640 psi) respectively when measured near the external sagittal Cl’CSt. Failure of human cranial bones has been shown by Woodm’ to be rate sensitive. He determined Young’s modulus and breaking stress (0y) to be rate sensitive, while energy absorbed to failure was shown to not be a function of strain rate. Table 1 contains some of the critical bone properties reported in the above literature. Table 1 Critical Cranial Bone Properties Parameter Value Eflmed ’0’ Impact Rate( 1 3) Game)” 3.00-6.20MPa No oymdcndmu” 0.28MPa No II 0mm" 68.9MPa Yes I Eu” 12.4GPa Yes Oms(canine)(1s) 25.1MPa No 11 Surgical Practices for Filling Cranial Defects The earliest report of an attempt to fill a large cranial defect was reported by Grant et.al.(39’ in 1670. In this case, a male patient was treated with a canine xenograft to repair a large cranial defect. Interestingly, though the procedure was a landmark surgical event, the cranioplasty was later removed at the insistence of the church. In 1920, Dr. Paul Lecéne‘”), a noted French surgeon, presented cranioplasty procedures using a) metal implants (gold or aluminum), b) an allograft, c) an intercostal cartilage graft, d) a tibial osteoperiosteal graft, and e) and a pedicle osteoperiosteal autoplasty. In the more recent literature, the use of hydroxylapatitewm", madreporic coral‘3”, hydroxylapatite and plaster of Paris composite”), PMMA and autogenous bone have been presented. Despite their limited success, the primary clinical outcome of resorbable materials has been unsatisfactory cosmetic results due to loss of primary implant shape in vivo. The concept of utilizing porous materials for implants has been to facilitate the induction of bone (osseoinduction) between the implant material and the host 12 bone as the primary mode of implant fixation. Theoretically, the implant that is firmly affixed by bony ingrowth will be stabilized by viable tissue providing a stable and physiologically active form of implant fixation. Besides theoretical implant/bone fixation, the use of 'off the shelf implant materials enables the surgical team to readily access the volume of implant material needed at the time of surgery. Though autogenous graft, retrieved from the ribs, iliac crest or cranial bone flap is resilient to the resorption mechanism, the harvesting of the graft requires a second incision, lengthens surgical time and increases the potential of surgical site morbidity. The use of autogenous bone is best suited for small and medium sized defects‘”) while for large defects, the use of metal plates‘36'37'38’ and PMMA‘”) are indicated in the literature. Forms of Cranial Vault Failure The Society of Automotive Engineers (SAE) Information Report on Human Tolerance to Impact Conditions”) defines two modes of cranial bone fracture. A Linear Fracture results from a distributed load which produces gross compression of the cranium resulting in a region of tensile loading away from the point of load. Crack initiation is often associated with a stress riser in the tensile stress field from which the crack travels in a straight line to the location of the load”). This type of cranial vault failure is a disruption of the l3 mechanical integrity of the skull, but there may be no permanent violation of the brain space. A Depresse cranium producing localized failure of the cranial vault. The SAE recognizes a 43/4 (in)2 for “clean pm through raw Figure 4 Failure Modes Related to Impact Area contact area of 12.90 cm2 (2 in?) to be the transition from a distributed load which may produce a linear fracture to a concentrated load which produce depressed fractures (Figure 4). A contact area of 4.84 cm2 (3/4 inz) is considered to be the transition from the depressed fracture to a punch through failure of the cranial bone”). The punch through failure is believed to be the result of the compression of the diploe layer and shearing of the inner and outer cortical bone tablesm'). For the cranioimplant, the punch through failure mechanism will potentially be the most serious loading condition which it must withstand. 14 Clinical Manifestations and Sources of Cranial Vault Failure Gennarelli“°), in a review of clinical case histories of head injury patients, has categorized the common accidents or load cases which have produced cranial injury (Figure 5). In a retrospective study of 434 hospitalized head injury patients, Gennarelli‘”) categorized the patient population by injury type. Head Injury Type Related to Cause Vault Fractures (Linear Fracture) m Depressed Fractures (Closed Head Injury) 'Fllls' refer to 'ellp and fell' accidents 'VehldeOceupent' lnfolvu a pluneter In In automobile Tedeurlln' are non-vehicle occupentu Involved In vehicle accident ‘Aneultu' Involved lnjurlouu euulle Percent Figure 5 Distribution of Linear and D8pressed Fractures Related to Cranial Insult“°’ 15 Other than fracture of the cranial bones, excessive deflection of the skull can present clinical symptoms. This second form of cranial vault failure does not cause fracture of the bony braincase, but results in injury to the brain itself (Figure 6). This is a form of "Closed Head Injury" where there has been damage to the brain, but there has not been penetration into the brain cavity. Often, brain damage is caused by severe acceleration ("whiplash") or excessive deflection (without fracture) of the cranial bones. Advani‘zo’ identifies a 0.635 mm (0.025 inch) deflection of the human skull to be the critical deflection at l. Scalp A B. C D II. Skull A B. C D. E. . Bruise (leakage of blood from a vessel into adjacent tissue). Abrasion (traumatic removal of some outer layers of scalp) . Laceration (cutting injury. tearing of scalp) . Avulsion (extreme laceration causing peeling of whole scalp) . Suture separation (diastasis, more often in younger skulls) Indentation (e.g. ping-pong “fracture", usually in younger skulls) . Linear fracture (may occur at points remote from impact location due to tensile stresses generated by sudden return of skull to its original shape after deformation) Depressed fracture (may be accompanied by perforation, fragmentation or comminution of skull) Crushed skull (massive comminution, usually due to extreme static loading) Ill. Extracebral Bleeding (Focal or Diffuse) A. Subarachnoid hemorrhage 8. C. Epidural hematome (with skull fracture in 90% of cases) Subdurel hemetoma (with skull fracture in 50% of acute cases. Usually due to torn bridging veins) Figure 6 Types of Head Injury Lesions the impact pole (i.e. point of load). Although this failure mechanism was not 16 evaluated in the study, the material and geometry of the cranioimplant must be sufficiently stable to prevent this magnitude of deflection. It must be as effective at supporting the load as viable cranial bone in the region of the cranial defect. Properties of Porous HTR” When compared to solid PMMA, the mechanical properties of HTR“ are compromised by the fully porous structure of the material. Although implantable porous biomaterials often facilitate ingrowth of soft tissue and/or bone, the mechanical properties are initially below those of solid implants. This is not an important feature, unless the implant is placed in a stress bearing environment. Ashman's worku'”) was primarily utilizing HTR as a "non-structural" allograft. The cranioimplant can be considered a "structural" graft, since it must be able to withstand impact loading in trauma situations. The material properties of porous materials will affect both tissue ingrowth, as well as clinical failure of the device. There has been significant research in the area of joint arthroplasty to determine the optimal pore size of implantable porous materials. Pilliar‘m, a 17 pioneer in orthopedic porous coatings, reported the optimum pore size for osseointegration into porous Titanium alloy and Cobalt-Chromium- Mollybidium alloy implants to be 50 - 400 um, based on canine femur studies. He further reported that for porous implants, mineralization of tissue ingrowth will eventually occur provided the following conditions are metzué’ 1. A snug fit of the implant in the bone and good implant-bone contact. 2. Good bone quality around implant which will facilitate quicker ingrowth rates. 3. An optimal pore size (> 50 am) to encourage bony ingrowth. Pore sizes less than 50 um are too small for osseointegration, but do allow fibrous tissue ingrowth.“°'2"25’ At the other extreme, porous implants with pore sizes greater than 400 pm require longer implantation times to achieve bony ingrowth. Pilliar‘m and Bobyn et.al.‘2°’ reported that porous Titanium femoral implants for total hip arthroplasty, with pore sizes in the 50 - 400 um range produced optimum fixation of 17 MPa in a push-out test 8 weeks post surgery. Implants in the 400 - 800 um range took 16 weeks to achieve this level of fixation. The pore size of HTR” is controlled to be 200 pm”), which will place it in the range that favors osseointegration. 18 The rate of ingrowth is a function of the bone quality into which the material is implanted. This not only includes healthy viable bone versus pathologic bone, but also the type of bone. Pilliar‘”) has documented greater rates of ingrowth when the implant is placed in cancellous bone versus cortical bone. This has also been reported by Oklund et. al.“” in her work with canine cranial implants, which showed greater tissue ingrowth occurring in regions of thicker diploe layers. The mechanical properties of porous PMMA (~150 um pore size) have been studied by Klavvitter et. alfzs) and Taylor et. aw“) in an effort to assess the effect of a porosity. Solid PMMA has a compressive yield strength (Yeompmsm) 5.6 times greater than that of the porous PMMA‘Z‘”, while Youngs modulus (Ecompmsion) was 5.8 times greater for solid versus porous PMMA‘“). Intuitively, the decrease in mechanical properties are a function of the contact area between spheres in the porous materials as shown (Figure 7). It follows 19 that a decrease in the pore size of a porous material would lead to an increase in the yield strength”), due to the resulting increase in the total area of contact. The reduced area of contact between spheres plays an even greater role when the porous material is placed in a tensile stress field. For porous PMMA, Klawitter et. alfm found the ultimate tensile strength (UTS) to be approximately 55% of the compressive yield strength. The compressive modulus of elasticity (Emmpmon) was found to be approximately three times the tensile elastic modulus (Eumsion). This resulting reduction in mechanical properties is intuitively due to the cross sectional area of the specimen in 12M 2M ' wolf‘s" Figure 7 Micrograph of HTR® (25X). Note contact between spheres 20 properties is intuitively due to the cross sectional area of the specimen in tension being reduced to the sum of the contact area between spheres along the fracture plane. Based on these studies of porosity, HTR“, used as a cranioimplant material, will be most susceptible to failure when cranial loading conditions place the implant in a tensile stress field (e.g. a "punch throug " failure mode). The implant will be most vulnerable to loosening or potentially fracture during the initial period, 2 weeks post-surgery“). Assuming the conditions defined by Pilliar‘”) are satisfied, infiltration of bone and soft tissue into the pores of the material will potentially strengthen the material and stabilize the implant in the cranial defect. EXPERIMENTAL PROCEDURES A protocol was developed to compare the biomechanical properties of HTR” before and after a period of implantation. The biomechanical strength assessment of porous implants was based on bending models for plates and beams. Testing to determine the material properties of the solid PMMA implants was dropped from this investigation due to inconsistencies in the material during fabrication. When manufactured, a significant number of air bubbles were left in the PMMA which influenced the material properties. For this study, the solid Cranioplastic implants were used for histological evaluation only. A three point bending fixture was used to assess the strength of implanted and non-implanted specimens, in order to quantify the effect of bone or soft tissue integration into the material pores. The effect of tissue ingrowth on rigidity of the implant was studied using classic Strength of materials plate modeling techniques. Finite Element Models (FEM) were used to assess the effect of 21 22 osseointegration into the curved implant geometry, verify the assumption that the geometry behaved as a flat plate, and determine the influence of specific geometrical features. Implant Design The development of the cranioimplant was performed in conjunction with Drs. D. Ievens and C. DeCamp, Veterinary Surgeons at the Michigan State University Small Animal Veterinary Clinic, East Lansing, Michigan. In order to assess osseointegration versus natural bone regeneration, a cranial defect too large to heal spontaneously (greater than 17 mm (0.663") diameter)”” was selected. To develop the implant, a temporal/parietal craniotomy, 30 mm X 17 mm (1.170" x .663"), was performed on a cadaveric beagle skull. A template was formed by filling the craniotomy with self-polymerizing PMMA in the doughy state and molding it to match the radius of curvature of the canine skull. Once cured, the polymer template was removed from the craniotomy and trimmed to a uniform and moldable geometry. Utilizing polycarbonate lens material, local eye glass grinders used this PMMA template to manufacture the final implant prototype. This resulted in an implant geometry with smooth and consistent radii of curvature (Figure 8). US Surgical used this prototype to 23 develop a compression mold for the manufacture of the premolded Cranioplastic” and HTR® cranioimplants. Four suture holes were added to the corners of the implant in order to interopertatively stabilize the device in the craniotomy. The resorbable sutures were used to supplement the biological fixation in the early'weeks post-cranioplasty. (m. . . _‘ l— : 30.I mm _ . _ ._ I6.9 mm—L—-l R 40.82 mm W 22.00 mm F —( 7.03 mm) _.., w (51 L-.. l—( 5.03 mm) 2-4 mm R ”9.65 mm) R 54.4 mm R ”7.65 mm R 50.0 mm 4 Figure 8 Detailed Implant Geometry 24 Clinical Phase Eight laboratory beagles were selected and kenneled at the Michigan State University, East Lansing, Michigan. A total of 15 cranioplasties were performed. Bilateral craniotomies were filled with HTRm and Cranioplastic” implants contralaterally in 6 animals. One animal received bilateral HTR” implants and the first animal received a single HTR“ implant (Figure 8). A standard midline incision was used to expose the calvaria. The 30 mm x 17 mm (1.170" x .663") temporal/parietal craniotomy was performed with a high speed craniotome. To ensure accurate sizing of the craniotomy, a 30 mm x 17 mm (1.170" x .663") PMMA template was used to ensure line-to-line contact Table 2 Cranioimplant material by animal. ANIMAL LEFT RIGHT HARVl HTR HARV2 HTR HTR HARV3 HTR PMMA HARV4 HTR PMMA HARVS HTR PMMA HARV6 HTR PMMA HARV7 HTR PMMA HARV8 HTR PMMA 25 at the implant-bone interface. Four 2 mm (.078") diameter holes were drilled with a high speed drill at the corners of the craniotomy approximately 3 mm (.117") from the edge of the implant. Resorbable sutures were used to secure the implant to the host bone and minimize motion at the implant-bone interface (see Appendix A for Dr. Ieven's detailed surgical procedure). The animals were euthanized by lethal intravenous injection 6 months post- cranioplasty. The cranioimplants were harvested with approximately 10 mm (0.397") of cranial bone to preserve the soft tissue and bony interfaces of the cranioplasty material. Mechanical Testing Molded HTR implants were used to determine the mechanical properties of the material. A 30 mm X 8 mm (1.170" X 0.318") beam sample was created by removing the central third of the implant. A simple three-point beam bending analysis of these beam samples was used to determine Young’s Modulus and ultimate tensile strength of HTR. In a second set up using the same test fixture, full geometry implants were loaded to failure for a qualitative 26 analysis of the fracture pattern and secondarily, to calculate the material properties using a more complex plate model. This second test was used for comparison purposes due to the increased complexities of the plate model versus the beam analysis. Beam analysis was used to compare non-implanted and implanted material properties. The testing fixture used in both beam and plate studies consisted of a 12.6mm (0.50 inch) thick aluminum plate firmly supported at the four corners to prevent any movement (Figure 9). To model the surgical environment, a 30 mm x 17 mm (1.170" x .663") hole was cut in the center of the plate to reflect size and shape of the craniotomy. The central load was applied with a 9.5 mm (0.375 inch)‘"” diameter polished steel ball bearing, positioned at the apex of the test specimen. The selection of a load applicator of this shape and size was to simulate a worst case cranial load situation in which a potential "punch - through" cranial vault failure may occur. The rounded applicator was chosen to decrease the sensitivity of the test setup to alignment between the load applicator and the test specimen. A crosshead speed of 0.625 mm/sec (0.025 inch/sec) (é = 2.35%/sec) was used for loading in the quasi-static case. The samples were tested to failure in each test. 27 i he ' v r h "I Figure 9 Test Set Up Used for Beam and Plate Analysis (Plate Shown) Research by Skowronskim) demonstrated that soaking HTR‘D in lactated Ringer's solution at body temperature stabilizes the material properties and increases the compliance of the material. Therefore, to model the material in the implanted environment, each of the specimens were soaked at 37°C in lactated Ringer's for at least 72 hours prior to testing. 28 Three - Point Beam Bending Specimens From 10 samples of premolded HTR® implants, the central 8mm section was obtained by shaping the full implant geometry using 240 grit sandpaperm) on a stationary belt sander (Figure 10). In order to limit the thermal stress effects, care was taken to frequently soak the specimen in clean lactated Ringer’s Solution during the shaping process. 29 Measurements were taken to determine the actual width (b) and thickness (d) of the test sample. The beam samples were mounted on the testing fixture. The samples were tested to failure in a quasi-static manner using an Instron model 1331 servo-hydraulic testing machine. The 9.5 mm (0.375 inch) diameter polished steel ball bearing was used as the load applicator. Force was applied at a displacement rate of 0.625 mm/sec (0.025 inches/sec). Force output from the 100 pound load cell and displacement data from the Instron testing machine were captured by a Nicolet Digital Oscilloscope. This data was transferred to a IBM PC for analysis. Analysis of Full Implant Geometry under Quasi-static Load Due to the concern about fixture dependant stress concentrations, each of the full geometry implants was potted in an epoxy resin mixture to ensure the implant was fully supported on all four edges. After the 72 hours soak in lactated Ringer's at 37° C(23), each of the 10 HTR implants was individually wrapped in cellophane (SaranwrapT') and subjected to an 8 mm Hg vacuum. The wrap protected the implant from the potting chemicals, while forming an air-tight seal that was fully conforming to the implant. The cellophane wrap was sprayed with a silicone lubricant to prevent the potting resin from 30 adhering. With the resin still setting, the implant was removed and excess resin was trimmed from beneath the implant. This ensured the implant was completely seated along a fully conforming 30 mm X 17 mm (1.170" X 0.663") rim, thereby modelling the theoretical clinical cranioimplant (see Figure 9). The thickness (d) of each implant was measured at the implant center before testing. The samples were tested to failure in a quasi-static manner using an Instron model 1331 servo-hydraulic testing machine. The 9.5 mm (0.375 inch) diameter polished steel ball bearing was used as the load applicator. Force was applied at a displacement rate of 0.625 mm/sec (0.025 inches/sec). Force output from the 100 pound load cell and displacement data from the Instron testing machine were captured by a Nicolet Digital Oscilloscope. This data was transferred to a IBM PC for analysis. Analysis of Harvested Shell Specimens The Fifteen full shell implants were recovered from each of the eight test animals. The cranioimplants were harvested with approximately 10 mm (0.397") of cranial bone to preserve the soft tissue and bony interfaces of the 31 cranioplasty material. All of these implants were then sectioned into three equal longitudinal segments. The central 8mm section, sectioned from the implant using diamond blade histological sectioning saw, was used determine the mechanical properties of the HTR after a 6 month period of implantation. In order to directly compare the pre- and post-implanted properties, the same three point beam bending test fixture and the protocol described in “Three- Point Beam Bending Specimens” was used. Histology Once harvested, the implants were sectioned into three longitudinal sections, the outer two sections which contained the significant portion of the bone/implant interface, were used for histological examination. These samples were potted in clear PMMA following standard histological sample preparation techniques. Representative sections of the bone/implant interface were prepared with H&E stain for analysis of the bone and soft tissue structures. MATHEMATICAL MODELING Complex approaches have been developed to model the human cranial vault‘41'“). Often these studies have been attempts to model the mechanisms of head injury in automobile accidents. The majority of these models have described the cranium as a liquid filled shell or half sphere. The design of the cranioimplant used in this study should be modelled as a shell or segment of a sphere. The use of Kirchoff-Love approximations for thin shell structures is limited to geometries that have an outside radius of curvature (pomidc) that is less than 5 times the thickness (i.e. pomsidc/ (powde- p ilnsidc) < 5). For geometries that do not meet this criteria, straight beam and flat plate modeling techniques accurately predict the behavior of the geometry. PMMA and PHEMA were assumed to be Hookian materials with isotropic material properties“ 1’. Further, the PMMA/PHEMA composite was assumed to behave as a homogeneous material due to the strong chemical bond formed between these two materials during fabrication‘z”. The geometry of the 32 33 cranioimplant used in this study (Figure 8) was modeled using a simplified straight beam and flat plate strength of materials approach. Three Point Beam Bending Model Eight millimeter (0.318 in) beam samples were sectioned from the central third of the implant and were modeled as a rigidly fixed beam under a central load. The beam model, rather than the flat plate model, was used to analyze the material properties of unimplanted and implanted HTR. The model was selected for the simplicity and reproducibility of the test set up and consistent and predictable failure mode. The beam was modeled as a rigidly fixed rather than simply supported beam due to the inherent notch sensitivity of the porous material. Prior testing supported the assumption that beam failure would occur before the end boundary conditions influenced the behavior of the beam. The flexure formula for a rigidly fixed beam is: P13 m“ 1921151r [1] 34 Also, I = 11";- for a rectangular cross section, and length (1) equals 30mm(1.170") due to the geometry of the test fixture. Thus, with force and deformation data measured from the lab testing, equation [1] can be rewritten to solve for Young's Modulus: Pl 3 Pl 3 g [2] 19261 166(bd3) e The bending stress distribution is described by: My OMug'T [3] where y is the distance from the neutral axis to the point of interest. In the three-point bending case, the maximum tensile stress is on the caudal implant surface at the center of the load where M = P1/4 and y = d/2 , thus the equation is reduced to: 3P1 0 = 4 35 In the analysis of the results, equation [4] was used to determine the Ultimate Tensile Strength (UTS) of the HTR material. Substituting equation [4] into equation [2], the strain component (6) can be resolved to: Equations [4] and [5] were used to normalize the force deformation information from the testing protocol and develop a Stress-Strain plot for the material in a three point beam bending test. Young's Modulus was calculated as the mathematical slope of a straight line through points on the stress-strain plot that correspond to test initiation and maximum tensile stress (or UTS). The strain value at UTS is defined as the Maximum Strain (emu). The "energy absorbed at failure" was calculated by mathematically summing the area under the curve of the stress - strain plot between theoretical zero (test initiation) and UTS. This area was calculated (integrated) using a computer subroutine based on the Runge-Kutta Method. 36 Plate Model In developing the theoretical plate model of the implant geometry, several assumptions were made regarding the boundary conditions at the bone/implant interface of the in viva device. Mathematically, by changing these boundary conditions, the implant was modeled for two extreme cases: a) no tissue integration, and b) total osseointegration. To simulate immediate post-operation and prior to any soft tissue or bone integration, the implant was modeled as a flat plate, simply supported along all four edges. Assuming a perfect surgical technique, this is the initial postoperative condition since the implant is able to flex freely around the periphery of the craniotomy which is assumed to be a straight, sharp and rigid edge. To further simplify the model, the implant was assumed to be a plate of constant thickness. This is based on the assumption the effect of the 2.5 mm (0.098") rim at the implant/craniotomy interface (Figure 8) is negligible. The validity of this assumption is tested in the Finite Element Analysis which is described later. 37 Timoshenkom’ has fully developed the model of a simply supported flat rectangular plate of uniform thickness under a central concentrated load. Using the dimensions shown in Figure 8, the maximum deflection of a simply Supported flat plate is,(28) 2 on mu: P“ itanham— "’ [6] 21t3Dm-1 m3 coshza where, a =rmtb [7 .. 2.. 1 E a3 D= m [8] 12(l-v2) In the case of the cranioimplant geometry (b/a = 1.8), the equation reduces tozas) Pa 2 6m=0.01620 [9] 38 or, simplified for the cranioimplant geometry, P 6 =l.245 “ E [10] m Timoshenko‘zs) further defines the bending moments generated in the simply supported plate under a concentrated central load to be: P a Mmax=E[(l+v)log:+ 1] [I I] 01' M =0.1593P [12] 39 The radius of the load, c, is equal to the radius of the steel ball used in the experimental procedure, 4.76 mm (0.1875 inches). The influence of load P, on the bending moments produced in the plate is shown in Figure 11. Under b P //— I ’////// Tension «ALL EDGES ABE W SIMPLY sapeoarec Tension Figure 11 Symmetrical Bending of Flat Plate about Neutral Axis. ‘——U these loading conditions, the geometry of the cranioimplant will experience the greatest bending moment along the shortest dimension ("a" as shown in Figure 11) of the implant. 40 To quantify the effect of bony tissue ingrowth (osseointegration) into the pores of HTR®, the boundary conditions of the model was changed to be completely rigid at all implant-bone interfaces. Again from Timoshenko‘zs’ plate theory, the model of a rectangular plate perfectly fixed on all four edges under a concentrated central load is: M =-0. l 667P and Pa2 a u=0.0786 or, simplified for the cranioimplant geometry, P 6 =0.537 “ E m tension [13] [14] [15] In this case, it was assumed that the junction at the bone implant interface was completely rigid with significant osseointegration into the pores of HTR”. 41 Finite Element Analysis The use of Finite Element Analysis (FEA) techniques enabled the geometry of the implant to be more fully developed at the bone implant interface. This was specifically intended to validate the assumption that the 2.5mm (0.098") rim on the implant at the implant/craniotomy interface had minimal effect on the rigidity of the implant and could be neglected in the strength of materials plate models. The FEA modelling techniques more clearly defined the influence of these geometrical features of the implant and their ability to transfer stress. Clinically, the geometry of the rim serves as a "positioning feature " to ensure the cranioimplant was easily placed in the craniotomy by the surgeon. This offered a “positive lock” that the surgeon could use to reproducibly position the cranioimplant in the surgical site. The potential disadvantage of this feature was that it served as a stress concentration due to the abrupt change in geometry. Placed in a strong tensile stress field, this could lead to failure of the device at the implant-bone interface. 42 The FEA model was based on the prior assumption that all materials behave as Hookian materials with isotropic properties. The model geometry was developed based on the print dimensions (Figure 8). By creating two separate models, "onlay" (without the rim) and "inlay" (with the rim), the stabilizing effect of this implant feature was shown through comparison of the model responses. The implant was modeled using AN SYS (version 5.0) Finite Element Analysis software package (AN SYS, Inc., 201 Johnson Road, Houston, PA 15342-1300). 405 20-node brick elements”) were used to mesh the model. The central, steady-state load of 110 N (40 lbf) was distributed equally over 4 central nodes. Onlay Model In the "onlay" model, all inferior (caudal) surface nodes of the perimeter elements were vertically constrained (Y- displacements). This is where the implant-bone interface was assumed to be of perfect geometry and fully conforming (Figure 12). The remaining degrees of freedom (X and Z 43 displacements and X, Y and Z rotations) where left unconstrained in order to model the initial clinical condition immediately post-operation. With these constraints, the implant slides along the cephalad surface of the cranial bone, since there are no biological adhesions at the implant-bone interface. This effectively eliminates the load carrying capability of the rim feature. Both “inlay” and “onlay” models were loaded with the same central loading scheme. Figure 12 Finite Element Model -- Initial Constraints 44 Inlay Model Clinically, the difference between the "onlay" and "inlay" models was assumed to be associated with the boundary conditions at the implant/bone interface. The "inlay" model was developed by changing the material properties of the perimeter elements on the caudal surface of the "onlay" model to those of bone (E = 12.4 GPa). This line of nodes was 2.5 mm (0.098") superior to the caudal surface which is the height of the rim on the implant. The model geometry emulated the effect of the osteotomized bone adjacent to the rim of the implant. Being a theoretical model, it assumes complete and intimate contact between the implant and bone, which is not clinically feasible. This modelling technique further assumes that there are no significant effects related to the unmodelled cranium due to the significant differences in Young's Modulus of HTR” and cranial bone. For this modelling approach, the bone interface was assumed to be completely rigid in all six degrees of freedom. The intent of this model was to isolate the effect of the positional feature and did not involve modelling the effects of osseointegration. RESULTS Due to the variability in the sample piece dimensions, the force-deformation output from each of the testing setups was normalized to generate stress-strain data . These data points were transferred to graphical format (Figure 13). Strms Strain Curve for HTR Beam in Quasi-Static 3-Point Bending 160- ' I Strain (mm/mm) 21 he.“ Figure 13 Typical Output from Three Point Beam Bending Test 45 46 Three-Point Beam Bending of HTR Prior to Implantation In three point beam bending experiment, failure was defined to correspond to maximum stress during the test (Table 3). The test coupons failed with the initiation of a crack in the tensile field (caudal surface) along the axis of the Table 3 Material Properties of Non-Implanted HTR Beam Samples in 3-Point Bending Analysis Young ' 5 Max Max Energy Sample Modulus Stress Strain 22:13:13? (3‘): (MPa) (MPa) (mm/mm) 21HTR 77.74 1.02 0.02 0.11 22HTR 60.74 1.23 0.02 0.05 23HTR 97.45 2.67 0.035 0.13 24HTR 71.89 1.62 0.025 0.08 25HTR 59.48 1.57 0.025 0.07 26HTR 79.28 2.21 0.03 0.13 27HTR 75.27 1.57 0.02 0.08 28HTR 91.53 2.64 0.015 0.09 29HTR 72.65 2.05 0.030 0.12 30HTR 65.32 1.10 0.020 0.05 Average 75.1 1.8 0.02 0.09 Dsetvainaiairodn 11.6 0.6 0.006 0 . 03 47 load applicator. The crack propagated until it intersected the top (cephalad) surface. There were no failures at the sample fixture interface. J .r . .. .. .T Figure 14 Failed HTR Beam Sample in Three Point Beam Bending Full Implant Model Prior to Implantation Output from the simply supported plate experiment (Table 4) was normalized to stress and strain values due to the dimensional variation between samples. Table 4 Mechanical Properties of I-ITR Based on Plate 48 Geometry in Quasi-Static Bending Analysis Youngs Max Max Ene rgy Sample Modulus Stress Strain fijiiifié‘)’ (MPa) (MPa) (mm/mm) 40HTR 55.91 1.4 0.024 0.069 41HTR 57.19 1.77 0.018 0.071 42HTR 42.32 0.95 0.016 0.026 43HTR 83.71 3.38 0.011 0.074 44HTR 49.02 2.78 0.021 0.096 45HTR 60.25 2.05 0.022 0.077 46HTR 69.32 1.71 0.020 0.048 47HTR 45.94 1.37 0.016 0.033 48HTR 77.91 1.68 0.016 0.038 49HTR 70.6 1.42 0.013 0.026 SOHTR 80.8 2.18 0.023 0.101 Average 63.0 1.88 0.018 0.060 Standard 13.7 0.66 0.004 0.026 Deviation 4 9 Failure of the plate coupon demonstrated perpendicular linear fracture planes initiating at the center of the plate. The planes approximately bisected the edges of the implant geometry (Figure 15). Crack initiation was on the caudal surface and slightly off center of the axis of the load applicator. Figure 15 Failure mode of Implant Geometry under Central Load 50 The Stress - Strain plot for the plate model demonstrated multiple stress peaks which may be related to the initiation of the different fracture planes (Figure 16). Visual inspection does not indicate timing of fracture planes. 1.8 1 I . . I Sample 41 htr 1.5 - _ i 1.2 ' / " g 0.9 ~ ~ .3 0.0 - - 0.3 - - “V I . . . .016 .032 .048 straln (mmlmm) Figure 16 Stress-Strain Plot of HTR Plate Sample A direct comparison between Table 3 and Table 4 highlight some variations in the material properties of HTR. There was a slight decrease in Young’s Modulus (75.1 MPa to 63.0MPa) (p<.05). The Maximum Stress decreased 51 (1.77MPa to 1.88 MPa"). Maximum Strain increased (.024 mm/mm to .0182 mm/mm) (p<.02) and Energy Absorbed at Failure increased (0.09 I to 0.0601) (p<.05). Three Point Beam Bending of HTR After Implantation Results of the Three Point Beam Bending experiment with the harvested HTR beam is given in Table 5. The analysis protocol is described under the quasi- static Three Point Beam Bending section above. "' The standard Student T-test statistical analysis is not applicable to this comparison due to the large difference in the standard deviations of the results. 52 Table 5 Harvested HTR Implants in 3-point Beam Bending Youngs Max Stress Max Strain Energy Sample Modulus (MPa) (mm/mm) Absnbait° Failure (J) (MPa) HARVlHTR 72.74 2.11 0.025 0.0425 HARVZHTRR 79.61 4.06 0.06 0.100 HARVZHTRL 25.26 2.08 0.100 0.087 HARVBHTR 15.434 1.32 0.085 0.053 HARV4HTR 34.53 2.36 0.070 0.080 HARVSHTR 79.87 2.93 0.035 0.040 HARV6HTR 63.45 1.03 0.025 0.050 HARV7HTR 59.78 .91 0.150 0.020 HARVBHTR 67.35 .71 0.020 0.028 Average 55.3 1.9 0.06 0.06 Standard 22.7 1.1 0.04 0.03 Deviation The failure mechanism of the explanted material was similar to the non- implanted specimens with crack initiation occurring on the caudal surface of the implant along the axis of the load impactor. In all samples, at least one of the implant bone interfaces was disrupted by the diamond saw blade used in the histological sectioning process. In order to 53 balance the sample in the test fixture, both bone fragments were removed from the HTR beam sample prior to testing. A general observation of the bone implant interface was that the strength was low and required only light force to disrupt it. The separation of the bone from the implant was a “clean” fracture of the interface. There were no instances of HTR fracture at the interface. A direct comparison between Table 3 and Table 5 highlight several changes between the pre-implanted and harvested HTR material properties. There was a decrease in Young’s Modulus (75.1 MP3 to 55.3 MPa) (p<.05). The maximum stress was unchanged (1.8 MP3 to 1.9 MPa‘). Maximum strain increased (.02 mm/mm to .06 mm/mm) (p<.02) and energy absorbed at failure increased (0.03 I to 0.061) (p<.02). Strength of Materials Parametric Models The two mathematical modelling techniques were used to compare the effect of rigid biological ingrowth on the flexure rigidity of the porous HTR device. A ’ The standard Student T-test statistical analysis is not applicable to this comparison due to the large difference in the standard deviations of the results. 54 parametric comparison of the two strength of materials models demonstrate the stabilizing (decreased motion) effects of osseointegration into the HTR boundary pores. For comparison, by dividing the deflection equation (Equation [10],Page 38) for a simply supported plate by the deflection equation (Equation [15],Page 40) for a rigid, fully supported plate demonstrates an increase in rigidity of 230% for the fully supported model. In this analysis, the increase in flexure rigidity is related only to the boundary conditions of the plate model and is based on the assumption the material properties of the plate are consistent between the two models. Potential material property improvements related to the complete osseointegration of a porous structure is not considered in this result. Finite Element Analysis of the HTR Implant At the request of US Surgical, the finite element models were run using the material properties contained in Table 6. The analysis outlines the stress and strain profiles associated with the two bone implant interfaces. Due to the boundary conditions used, both models are “osseointegrated” at the boundary elements which is equivalent to the rigidly fixed strength of materials model. The comparison between the graphical stress profiles of the “onlay” and 55 Table 6 Material Properties used in FEM Material E v <30) HTR® 70 MPa .25 0 (17) (I7) Cranral Bone 12 GPa 35 “inlay” models (Figure 17) show there is no significant stress concentration associated with the "rim” (positioning feature) of the geometry. The tensile stress field on the caudal surface of the two models is similar in stress distribution. The “inlay” model presents a smaller area of the higher tensile stress region than the "onlay” model. 56 Figure 17 Comparison of Inlay and Onlay FEM Models 57 There was in increase in flexure rigidity associated with the “inlay”model of 13% (Table 7). The strain profiles of the “inlay” model (Figure 18) demonstrate the load transfer between the implant and bone interface. These same results also show the lack of sensitivity of the design to the notch stress concentration at the “rim” feature. Table 7 Results from FEA Onlay and Inlay Models Maximum Maximum Maximum Deflection Stress Strain (mm) (MPa) (mm/mm) Inlay 2.164 10.85 0.11 Model Onlay 2.486 11.36 0. 12 Model 96 Change 13% 4.5% 8.3% 58 Figure 18 Strains (ex and 62) associated with the caudal surface of the "inlay" implant 59 Histological Analysis of the Retrieved HTR and Cranioplastic Implants Gross Morphology Observations Retrieved samples of HTR and Cranioplastic had developed a soft tissue sheath across the cephalad and caudal surfaces. The caudal surface tissue was highly vascularized (Figure 19). This tissue appeared to be a continuation of the periosteal membrane of the adjacent cranial bone. By gross observation, these tissues appeared to be interdigitated into the surface of the HTR implant (Figure 19). Conversely, it was apparent these tissues were not adhered to the smooth surface of the Cranioplastic device (Figure 20) The strongest bony/soft tissue adhesions to the HTR implant were observed at the superior-posterior pole though the strength of the interface was not measured. In most cases, this interface was disrupted during the microtorning process (Figure 21). 60 “1.9,, ‘ V k”; -‘~hn‘u:' . Figure 20 Interface of Cranioplastic Cranioimplant 61 It was apparent from the gross observations that the implant was not fully 3% Figure 21 Bone/Implant Interface of Resected HTR Cranioimplant conforming to the curvature of the skull resulting in an ill fit of the implant in the craniotomy (Figure 22) . In Figure 23, the presence of a significant number of air bubbles in the Cranioplastic material is easily observed. 62 Figure 23 Micro Air Bubbles in Cranioplastic Material 63 Microscopic Observations Ingrowth of soft tissue into the pores of HTR was demonstrated at the bone implant interface. There was no apparent enhanced osseointegration outside Q ., A Figure 24 Soft Bead Depth on to Approximately 3 Tissue Penetrati of normal bone healing response. The presence of striated soft tissue was observed to penetrate to approximately a three bead depth (Figure 24). The presence of striated soft tissue is observed to be in contact with the surface of the Cranioplastic material though there is no penetration or adhesion to the material (Figure 25). 64 Figure 25 Soft Tissue in Contact with Cranioplastic Material but Not Adhered DISCUSSION The use of HTR as a premolded cranioplasty implant has been reviewed by theoretical and biological means in this research. The porous structure of HTR offers the clinical benefit of potential osseointegration and biological fixation. Thus, this HTR material can be incorporated into the host cranial bone and serve as a scaffold for bone remodelling. Though the mathematical models in this research predicted (based on assumptions) the stabilizing effect of osseointegration on the cranial implant, the biomechanical and histological evaluation did not corroborate this predicted result. As stated by Rawlings et. alum, the ideal material for a cranioimplant would have similar material properties as the host cranial bone. In the case of I-ITR, the material properties even after a 6 month period of implantation, are significantly less than those of cranial bone. The HTR material, used as a cranioimplant in large cranial defects, does not appear to facilitate the 65 66 osseoconductivity that is required to increase the material properties in this application. This may indicate that the material is better suited to be used a defect filler versus a load bearing implant. BEAM AND PLATE Experimentally, the non-implanted plate model resulted in lower material properties than the beam model (Ebcam= 75.1MPa vs. Eplate= 63.0MPa). This is most likely due to the more complex stress field generated in the plate model, as demonstrated by a comparison of the fracture patterns. Relative to the beam, the plate model exposes more of the bead-to-bead contact areas to a tensile stress field. This potentially leads to earlier crack initiation due to the increased number of “stress concentrations” along the fracture plane. Consequently, the plate model resulted in lower material properties for the HTR material. The fracture pattern in the beam and plate model followed the predicted planes of failure‘sz’. This is significant in establishing that there are no geometrical features of the implant design which influence the fracture pattern. In particular, the influence of the “positioning” ridge on the undersurface of 67 the implant and the presence of the drill holes in the four corners did not affect the failure mode of the geometry. These two features do not appear to weaken the implant design which was an initial concern. This interpretation of the fracture plane in the plate model supports the use of the simply supported flat plate strength of materials approach in modelling. By using the beam model in the three point bending test protocol to compare pre- and post-implanted devices, the effect of the porous material was minimized. The use of full implant geometries to compare pre- and post- implant material properties would have increased the scatter in the results due to the variable performance of the porous structure in a tensile field. PRE- AND POST-IMPLANT The intent of using the porous HTR material as a cranioplasty implant is that bone will grow into the pore of the implant and strengthen the material matrix. Hence, this should increase the material properties of the HTR. This concept is similar to the engineering principles used in carbon fiber reinforced composites. Clinically, this would mean the HTR material becomes more “bone like” through bone ingrowth, following a period of implantation. In this 68 investigation, comparison of the pre- and post-implanted three point beam bending specimens demonstrated a decrease in stiffness of the matrix (75.1MPa vs. 55.3 MPa). It is unclear why this reduction was observed. Though not investigated, this reduction may be related to the degradation of the PHEMA/PMMA interface during the period of implantation. If the material properties of the plastics which form HTR were stable in-vivo, there should have been a similar result between the pre- and post-implanted samples. A more significant importance of the biological fixation (soft tissue and bone) may have been seen if the size of the defect was reduced to one that would “heal spontaneously”. In this case, the bone and soft tissue would have bridged the defect as long as the HTR did not prevent the formation of bone. This would be beneficial, assuming the osseointegration will improve mechanical properties of the HTR implant. The effect of bone ingrowth may have been different in another clinical model. The primary difference in clinical experience reported by Ashman et. al.‘6'7'8'9’ and this research is the influence of the surgical site. As supported by the reported oral-maxillofacial experience, HTR placed in a rich cancellous bone structure will act as a scaffold for bone formation. In the cranioplasty application, the host bone 69 does not offer the same amount of cancellous bone structure and consequently appears less able to support osseointegration. FEM The Finite Element models demonstrated the effect of the stabilizing feature of the implant design. The ability of this modeling tool to predict the stress distribution in the complex plate model enabled the comparison of the effect of an onlay and inlay design. The results shown the “inlay” model to have a 13% lower maximum deflection and a 4.5% decrease in maximum stress. This result is best case since it is a model of the perfect surgical procedure. Consequently, the assumption of perfect bone implant interface is effective for modelling but is not practical clinically. Since the implant was premolded and not a custom design for each canine, there were errors in fit between the implant and craniotomy. This reduces the effect of the positional feature in the in-vivo performance of the prothesis. Eliminating the positional feature from the strength of materials analysis of the beam and plate make these models more conservative and consequently offer a higher margin of safety. 7O HISTOLOGY The histological results demonstrated that bony tissue did not integrate the interstical spaces of the porous material, as hypothesized. This may be related to the type of bone at the surgical site. As demonstrated by Oklund st. 511/”) in canine cranial defects, new bone forms more vigorously around cancellous bone. In the craniotomy model, the majority of bone close to the implant was cortical bone, which does not remodel rapidly. Though the Cranioplastic material demonstrated superior material properties, the disadvantage of the solid material was demonstrated by macro-histology results of this research. The body appears to only encapsulate the solid PMMA implant with a non-adherent soft tissue envelope. This may contraindicate the use of resorbable sutures for this style of implant whereas the porous structure would eventually be stabilized by the soft tissue ingrowth. CONCLUSION The clinical and biomechanical benefit for the use of HTR as a premolded cranioplasty material, was not demonstrated in this research. Though the material appears to be tolerated by the host cranial bone and soft tissue, the ingrowth of bone into the porous structure of the material was limited or not present. Without the potential for strengthening of the porous material through osseointegration, the design of the implant should be reconsidered. The following recommendations can be made: 1. Us: osseoinductive agent in HTR formulation: The physical characteristics (hydrophilic material and pore size) of HTR are correct for osseoconductivity. In the cranioplasty application, due to the type of bone around the defect, the addition of an agent that would induce the formation of bone within the HTR matrix may lead to the intended result of the porous structure. Materials such as Bone Morphogenic Protein (BMP) and TGF-B may improve the bone development in the cranioimplant. 71 72 2. our heei 0f_‘1mlo. u .. teurostu ureof he H I R is present enly at the bene-implant interfaee: To take advantage of the desirable material properties of solid PMMA with the benefits of a porous interface, modify the implant design to be based on a “core” of solid PMMA. This could be done by applying the HTR material (approximately 3 bead layers thick) to the surface of the premolded Cranioplastic implant that would contact with the bone. This implant design would have higher material properties (improved biomechanical performance) due to the solid PMMA COI’C . h HTR m ri 1 in r 1i i n : The material properties of HTR may be more suited for other orthopaedic applications such as non-stress bearing grafting or the basis of a drug delivery system. In the latter, the porous nature of the material would allow rapid elution of drug therapy while the material serves as a temporary spacer block (e.g. treatment of an infected knee joint). The use of porous implants in treating skeletal disorders or trauma could be appealing due to the potential for osseointegration. In clinical practice, the positive benefits which include bony incorporation of the implant, may not 73 outweigh the inherent reduction in material properties due to the porous structure of the material. APPENDD( A Surgical Protocolf (ULAR #01-89-403-01) Eight mature beagles, four male and four female, were admitted to the animal care facility at Michigan State University. All animals were given a thorough physical examination and complete neurological examination. No abnormalities were detected. Anesthesia was induced in each dog using 2% thiopental intravenously and lidocaine 2 mg/kg intravenously. The animal was intubated and a surgical plane of anesthesia was maintained using isoflurane. After induction dexamethasone was administered at 2mg/kg intravenously. A catheter was placed in the dorsal pedal artery for direct arterial pressure measurement. Intermittent positive pressure ventilation was begun after intubation. Oxygen saturation was measured using a pulse oximeter and end tidal CO2 was ’ Surgical Protocol developed by Dr. Dermot Ievens, DVM, Michigan State University, East Lansing Michigan 74 75 measured continuously after intubation. The skin was clipped over the surgical site to include both ears, the dorsum of the nose and the dorsal proximal half of the neck. The site was prepared for aseptic surgery. The animal was placed in dorsal recumbency, the muzzle being taped to a head stand. The surgical site was again scrubbed 3 times. The surgical site was draped in a sterile manner. A transverse skin incision was made extending from a point caudal to one pinna across the dorsum of the skull, to a point caudal to the contralateral pinna. A similar skin incision was made rostral to both pinnae. These incisions were connected with a longitudinal incision over the dorsal midline. These incisions were deepened to include the superficial pre-auricular and post-auricular muscles. The skin and auricular muscles were reflected to expose the temporalis muscle and facia. The temporalis fascia and muscle were incised close to their attachments bilaterally and the muscle was elevated away from the temporal fossa using a periosteal levator. Using a slow- speed hand drill, a hole was drilled in the caudal dorsal aspect of the frontal bone for placement of the transduced of an intracranial pressure monitoring kit. Intracranial pressure was then continuously monitored. Prior to placement of the intracranial pressure monitor, Lasix was administered at 1-2.5 mg/kg intravenously. Using a sterile template and surgical marker, the areas of the proposed cranial defects were outlined bilaterally. Prior to creation of the 76 intracranial defect, end tidal CO2 was reduced to 22-25 mmHg using intermittent positive pressur-e ventilation. Intracranial pressure was consistently below 10 mmHg at this time. Mannitol was administered at .25- .5 g/kg slowly intravenously at the time of craniotomy. Using a perforator and craniotome, the area of the parietal bone and caudal frontal bone outline was removed. Holes were placed on each corner of the defect for suture placement. On one side, the defect was filled using a premolded HTR® cranioimplant. The implant was held in place using monofilament polypropylene sutures and non-metallic surgical clips, making use of the previously drilled four holes. The contralateral side was filled using a premolded PMMA implant. This implant was anchored as described above. Each defect was filled prior to creation of the contralateral defect. Intracranial pressure was continuously monitored during craniectomy and implant placement, as an indicator of excessive cerebral trauma. The temporalis muscle and facia were sutured in place using simple interrupted 2-0 monofilament polyglyconate sutures. Prior to final closure of the temporalis muscle and fascia on the right side, the intracranial pressure monitor was removed. At this time, end tidal CO2 had returned to approximately 35 mmHg. The superficial preauricular and postauricular muscles were sutured using simple interrupted 3-0 monofilament 77 polyglyconate sutures. The skin was closed using surgical staples. During the procedure Keflin was administered at 10 mg/lb intravenously every 2 hours. The animal was recovered at the intensive care unit of Michigan State University Veterinary Clinical Center. Butorphanol (0.2 mg/kg) was administered postoperatively if the dog appeared painful in any way. The dogs were usually ambulatory one hour postextubation. Each dog was hospitalized for a minimum of one week at Michigan State University Veterinary Clinical Center prior to return to Laboratory Animal Care Services. Complete physical and neurological examination once monthly. At 6 months postoperatively, each dog was anesthetized using pentobarbital intravenously. Each animal was intubated. Nuclear magnetic resonance imaging of the skull was performed at the Clinical Center, Michigan State University. The dogs were then recovered at the intensive care unit, Veterinary Clinical Center, Michigan State University. Each dog was euthanized at 6 months postoperatively and the implants and surrounding calvarium were harvested. Euthanasia was performed by means of lethal intravenous injection. BIBLIOGRAPHY lA. Ashman and P. Bruins, "HTR (Hard Tissue Replacement) For Edentulous Ridge Augmentation," The New York Journal of Dentistry, 53, No. 8 (1983), pp. 387 - 392. 2A. Ashman and P. Bruins, "Prevention of Alveolar Bone loss Postextraction with HTR"D Grafting Material," Oral Surgery, 60, No. 2, (1985), pp. 146 - 153. 3A. Ashman, S. Neuvvirth, and P. Bruins, "The HTR Molded Ridge for Alveolar Augmentation -- An Alternative to the Subperiosteal Implant, Autogenous Bone Graft, of Injectable Bone Grafting Materials", Oral Implantology, 12, No. 4 (1986) 4B. Eppley, A. Sadove, R. German, "Evaluation of HTR polymer as a craniomaillofacial graft material," Plastic Reconstructive Surgery, 86, No. 6, pp. 1085-1092. 5M. Amler, R. LeGeros, "Hard Tissue Replacement (HTR) polymer as an Implant Material," Journal of Biomedical Material Research, 24, No. 8 (1990), pp. 1079-1089. 6United States Patent, 4,535,485 7United States Patent, 4,547,327 8United States Patent, 4,547,390 9United States Patent, 4,728,570 78 79 10C. Rawlings III, R. Wilkins, I. Hanker, N. Georgiade, and I. Harrelson, "Evaluation in Cats of a New Material for Cranioplasty: A Composite of Plaster of Paris and Hydroxyapatite," J Neurosurg, 69 (1988), pp. 269-275. ”GJ. Tortora, Principles of Human Anatomy, 2"d ed. (New York: Harper & ROW, 1980). 12P. Lowenhielm, "The Anatomy and Physiology of Head and Neck Injuries" in The Biomechanics of Impact Trauma, ed. B. Aldman and A. Chapon (Amsterdam: Elsevier Science Publishers B. V., 1984), pp. 93 - 101. ‘3]. Wood, "Dynamic Response of Human Cranial Bone," I. Biomechanics, Vol. 4, No. 1 (1971), pp. 1 - 12. 1”R. Hubbard, "Flexure of Layered Cranial Bone," I. Biomechanics, Vol. 4, No. 1(1971), pp. 251 - 263. 15S. Oklund, D. Prolo, R. Gutierrez, and S. King, "Quantive Comparisons of Healing in Cranial Fresh Autografts, Frozen Autografts and Processed Autografts, and Allografts in Canine Skull Defects," Clinical Orthopaedics and Related Research, No. 205 (1986), pp. 269-291. 16R.M. Pilliar, "Powder Metal - Made Orthopedic Implants with Porous Surface for Fixation by Tissue Ingrowth," Clinical Orthopaedics and Related Research, No. 176 (1983), pp. 42 - 51. ”ES. Gurdjian, Impact Head Injury: Mechanistis, Clinical and Preventative Correlations (Springfield: Charles C. Thomas Publishing, 1975) pp. 60 - 61. ”Human Tolerance to Impact Conditions as Related to Motor Vehicle Design -- SAE 1885 APR80, SAE Information Report (Warrendale,PA: Society of Automotive Engineers, 1980), pp.1 - 7. 19T. Gennarelli, "Clinical and Experimental Head Injury" in The Biomechanics of Impact Trauma, ed. B. Aldman and A. Chapon (Amsterdam: Elsevier Science Publishers B. V., l984),pp. 103 - 115. 20S. Advani, AA. Ommaya, and W. Yang: Head Injury Mechanisms. Chapter 1 in: Human Body Dynamics, Ed. D. Ghista. Oxford U. Press. 1982. 80 21AA. Ommaya, "The Head: Kinematics and Brain Injury Mechanisms" in The Biomechanics of Impact Trauma, ed. B. Aldman and A. Chapon (Amsterdam: Elsevier Science Publishers B. V., 1984), pp. 117 - 125. 22W Goldsmith and AA Ommaya, "Head and Neck Injury Criteria and Tolerance Levels" in The Biomechanics of Impact Trauma, ed. B. Adman and A. Chapon (Amsterdam: Elsevier Science Publishers B. V., 1984), pp. 149 - 194. 23D. Skowronski, 1988, personal communication. 24D.F. Taylor and EB. Smith, "Porous Methyl Methacrylate as an Implant Material," I. Biomed. Mater. Res. Symposium,No. 2,part 2 (New YorkzIohn Wiley 8L Sons, Inc., 1972) pp. 467-479. 25].]. Klawitter,A.M. Weinstein, and LI. Peterson, "Fabrication and Characterization of Porous-Rooted Polymethylmethacrylate (PMMA) Dental Implants," I Dent Res, 56, No. 4 (1977), pp. 385-393. 26I.D. Bobyn, R.M. Pilliar, H.U. Cameron, and CC. Weatherly, "The Optimum Pore Size for the Fixation of Porous-Surfaced Metal Implants by the Ingrowth of Bone," Clinical Orthopaedics and Related Research, No. 150 (1980), pp. 263 - 270. 27R. Cytermann, "A New Way to Investigate the Dependence of Elastic Moduli on the Microstructure of Porous Materials," Powder Metallurgy International, 19, No.4 (1987), pp. 27 - 34. 28$. Timoshenko and S. Woinowsky-Krieger, "Theory of Plates and Shells," 2"d ed. (New York: McGraw-Hill, 1959). 29H. Parisch, "Nonlinear Analysis of Shells Using Isoparametric Elements," paper presented at "The Winter Annual Meeting of The American Society of Mechanical Engeering, Washington, DC, November 15 - 20, 1981. 30RB. Welsh, R.M. Pilliar, and I. Macnab, "Surgical Implants, The Role of Surface Porosity in Fixation to Bone and Acrylic," The Iournal of Bone and Ioint Surgery, 53-A, No. 5 (1971), pp. 963 - 977. 81 31F. Roux, D. Brasnu, B. Loty, B. George, and G. Guillemin, "Madreporic Coral: A New Bone Graft Substitute for Cranial Surgery," I Neurosurg, 69 (1988), pp. 510-513. 32V. Matukas, I. Clanton, K Langford, and P. Aronin, "Hydroxylapatite: An Adjunct to Cranial Bone Grafting," I. Neurosurg,69 (1988), pp. 514-517. 33P. Manson, W Crawley, and I. Hoopes, "Frontal Cranioplasty: Risk Factors and Choice of Cranial Vault Reconstructive Material," Plastic and Recontructive Surgery, 77, No. 6 (1986), pp. 888-900. 3”R. Holmes and H. Hagler, "Porous Hydroxyapatite as a Bone Graft Substitute in Cranial Reconstruction: A Histometric Study," Plastic and Reconstructive Surgery, 81, No. 5 (1988), pp. 662-671. 3SASTM F-622, "Standard Specification for Preformed Cranioplasty Plates That Can Be Altered," American Society for Testing and Materials, 1989 Standards. 36ASTM F-452, "Standard Specification for Preformed Cranioplasty Plates," American Society for Testing and Materials, 1989 Standards. 37ASTM F-SOO, "Standard Specification for Self-Curing Acrylic Resins Used In Neurosurgery," American Society for Testing and Materials, 1989 Standards. 38P. Lecéne, "Cranioplasty and Cranial Prothesis," translated by Ian T Iackson, M.D., Plastic and Reconstructiove Surgery, 78, No. 4 (1986), pp. 530-535. 39F. Grant, N. Norcross, "Repair of Cranial Defects by Cranioplasty," Ann Surgery, 110 (1989), pp. 488-512. 408. Schmitt, I. Schreiner, R. Abraham, "Construction of Large Cranial Implants," Military Medicine, 153 (1988), pp. 582-585. “A. Engin, A. Engin, "Survey of Theoretical and Experimental Mechanics applied to Head Injury," Shock and Vibration Digest, 7 (1975), No. 3. 42A. Engin, A. Engin, "Survey of the Dynamic Response of Spherical and Spheroidal Shells," Shock and Vibration Digest, 7 (1975), No. 3. 82 43L. Salman, L. Kinney, "Clinical Response of Hard Tissue Replacement (HTR) polymer as an Implant Material in Oral Surgery Patients," Iournal of Oral Implantology, 18, No. l, (1992), pp. 24-28. “S. Isaksson, "Aspects of Bone healing and Bone Substitute Incorporation. An Experimental Study in Rabbit Skull bone Defects," Swedish Dental Iournal Suppliment, 84 (1992), pp. 1-46. ”S. Shahmiri, 1. Singh, S. Stahl, "Clinical Response to the Use of the HTR Polymer Implant in Human Intrabony Lesions," International Iournal of Periodontics and Restorative Dentistry, 12, No. 4 (1992), pp. 294-299. 46F. Louise, A. Borghetti, "New Developments in Synthetic Bne Replacement Materials," Current Opinions in Dentistry, 2 (1992), pp. 97-103. 47R. Yukna, R. Greer, "Human Gingival Tissue Response to HTR Polymer," Iournal of Biomedical Material Research, 26, No. 4 (1992), pp. 517-527. 48A. Pearsall, R. Spears, M. Chokshi, "The Ultrastructural Architecture of the Tisue/Hard-Tissue Replacement Replacement Interface," Iournal of Oral and Maxillofacial Surgery, 50, No. 4 (1992), pp. 375-385. ”T. Najjar, W Lerdrit, I. Parsons, "Enhanced Osseointegration of Hydroxylapatite Implant Material," Oral Surgery, Oral Medicine, Oral Pathology, 71, No. 1 (1991), pp. 9-15. soS. Stahl, S. Froum, D. Tarnow, "Human Clinical and Histologic responses to the Placement of HTR Polymer particles in 11 Intrabony Lesions," Iournal of Periodontology, 61, No. 5 (1990), pp. 269-274. 5‘ A. Sadegh, G. Luo, S. Covvin, "Bone Ingrowth: An Application of the Boundary Element Method to Bone remodeling at the Implant Interface," I. Biomechanics, 26, No. 2 (1993), pp. 167-192. 52 I. Colton, “Multiple Fracture of Plates Under Localized Impulsive Loading,” I. of Applied Mechanics, 58, No. 3 (1976), pp. 33-38. HICHIGRN STRTE UNIV. LIBRRRIES lllllllll ll llllll lllllllllllllllllllllllllllll ll 1 ll 31293015819026