A STUDY OF ’16 CANINE HEMANGIOSARCOMAS RECORDED AT MMHEGAN STAYE UNEVERSITY FROM ”56 TO 1965 Thesis for the Degree of M. 5. MICHIGAN STATE UNWERSITY Ernest E. McCcnneIi 1.965 +————- LIBRA R Y Michigan 3.21%.: University flat-At ABSTRACT A STUDY OF 76 CANINE REMANGIOSARCOMAS RECORDED AT MICHIGAN STATE UNIVERSITY FROM 1956 TO 1965 by Ernest E. MbConnell This thesis is a retrospective study of 76 canine hemangiosaro comas diagnosed in the Department of Pathology from 1956 to 1965. The study was divided into 3 basic parts: (I) occurrence and gross characteristics, (2) microscopic characteristics, and (3) radio~ graphic characteristics of the tumor. Hemangiosarcomas occurred with greater incidence in the large breeds of dogs, particularly the Boxer. There was a sex ratio of approximately 2 males to 1 female. The mean age was 9 years, a little earlier in males (8.5 years) and a little later in females (10.7 years). The most common primary sites were the subcutis and the spleen. If metastasis occurred the most common site‘was the lung. The tumor was difficult to diagnose without histologic examination. Hemorrhage was an almost constant finding. The neoplastic endothelial cell was generally spindle shaped, resembling a short fibroblast. The nucleus was generally oval and of moderate size (9 x 12 microns), often containing large eosino= philic single and multiple nucleoli. There was a tendency for the cytOplasm of 1 cell to fuse with the next, presenting a syncytial type of arrangement rather than individualized cells. There were both normal and abnormal mitoses. Host of the nuclei had a moderate Ernest E. McConnell amount of hyperchromatism. The cellular arrangement varied from normal appearing vascular cells to undifferentiated cellular sheets arranged in whorls or clumps with little vessel formation. The use of a reticulum stain revealed a fine reticular meshlike network, even in the most anaplastic areas of the tumor. There was a variable amount of necrosis, especially associated with hemorrhages. The amount of stroma varied from large amounts in the vascular forms to scanty in the more anaplastic cellular forms. Invasion of blood vessels occurred in 29 of 59 hemangiosarcomas. There were inflamma= tory cells in many of the tumors, particularly those which had an eroded surface. Two radiographic characteristics were observed. First, the metastatic lesions in the lungs presented a ”snowflake" pattern. Secondly, osteolysis occurred in those neOplamms that were in or near bone. A STUDY OF 76 CANINE HEMANGIOSARCOMAS RECORDED AT MICHIGAN STATE UNIVERSITY FROM 1956 TO 1965 By Ernest E. McConnell A THESIS Submitted to Michigan State University in partial fulfilhment of the requirements for the degree of MASTER OF SCIENCE Department of Pathology 1965 To Judy ii ACKNOWLEDGEMENTS The author expresses his gratitude and appreciation to the following peOple for their help in making this thesis possible: To Dr. R. F. Langham, major professor, for his guidance and personal stimulus to me during my graduate work in pathology; especially for his many suggestions and help in fulfilling the thesis requirements for this degree. To Dr. V. L. Sanger‘who, during the short period of our acquaintance, has given unselfishly of his time and knowledge in preparation of this thesis. To Dr. U. V. Mestosky, radiologist in the Machigan State University Department of Veterinary Clinics, for his help and time with the radiographic aspects of this thesis. To the Armed Forces Institute of Technology, Air University, United States Air Force, for the Opportunity to study at Michigan State University. To the National Cancer Institute, National Institutes of Health, U. S. Public Service, particularly to Dr. R. A. TjaLma for providing normal canine papulation data. Finally, appreciation is expressed to the laboratory technicians at Michigan State University for their assistance in preparing the tissues and other research material for my use. 111 TABLE OF CONTENTS INTRODUCTION. . . . . . . . . . . . . . . REVIEW OF THE IJTERATURE. . . . . . . . . MATERIALS AND METHODS . . . . . . . . . . Preparation of Material. . . . . . . Study Procedures . . . . . . . . . . TERMINOLOGY . . . . . . . . . . . . . . . RESUETS . . . . . . . . . . . . . . . . . Occurrence and Gross Characteristics ‘Microscopic Characteristics. . . . . Radiographic Characteristics . . . . DISCUSSION. . . . . . . . . . . . . . . . Occurrence and Gross Characteristics Microscopic Characteristics. . . . . Radiographic Characteristics . . . . SUMMARY . . . . . . . . . . . . . . . . . LIST OF REFERENCES 9 O O C O O O O O O O 0 iv Page 12 12 14 16 18 18 32 58 62 62 67 73 74 76 LIST OF TABLES Table Page 1 Occurrence and gross characteristics of 76 canine hemngiosarcmseeeeeeeoeeoeeeseases19 2 Microscopic characteristics of 76 canine hemangiOm sarcmsoeeoeeeeeoeeeoeeeoeoooso33 3 Comparison of observed and expected cases of heman310° sarcoma among specific breeds and weight classes. . . . 63 Figure 10 11 12 14 LIST OF FIGURES Incidence of hemangiosarcomas in male and female dogs of different weight classes. . . . . . . . . . Age and sex distribution of hemangiosarcomas in dogs 0 O I O O O O O 0 O O O O O O O O O O O O O O O Metastatic hemangiosarcoma in the lungs . . . . . . Metastatic hemangiosarcoma in lungs, spleen, kidney, andliver.'.................... Cut surface of hemangiosarcoma in diaphragm (pri- mry Bite). 0 O O O O O O O O O O O O O O O O O O O Metastatic hemangiosarcoma in small intestine with adjacenthOOdCJ-Otoaeeeeseeeeeeeae Numerous spindle-shaped cells lining vascular spaces of a hemangiosarcoma. Hematoxylin and eosin. x 750 O O O O O O O O O O O O O O O O O O O O O O O 0 Cellular type of hemangiosarcoma with little vessel formation. Hematoxylin and eosin. x 468. . . . . . Cellular type of hemangiosarcoma showing vesicular nucleus and large nucleolus. Hematoxylin and eosin. x 750. O O O O O O O O O O O O O O O O O O 0 Large nucleolus in malignant cell of a hemangioo sarcoma. Hematoxylin and eosin. x 750. . . . . . . Cellular type of hemangiosarcoma containing large abnormal mitotic figure (arrow). Hematoxylin and 8081B.x750..oooco.oo.......... Cellular type of hemangiosarcoma containing tri- polar mitotic figure (arrow). Hematoxylin and eosin. x 750. O O O O O O 0 O O O O O 0 O O O O O 0 Vascular type of hemangiosarcoma. Hematoxylin and eosin. x 75 C O O O O O C I O O O O O O O O O O O 0 Higher magnification of Figure 13. Hematoxylin and eosin. x 187.5. C O O O O O O O O O O O O O O O O 0 vi 28 30 30 31 31 49 49 50 50 52 52 54 54 Figure 15 16 17 18 19 20 21 22 23 Vascular type of hemangiosarcoma with vein-like channels and abundant stroma. Hematoxylin and 'e081n.XJ-87a5.eeoeeeeeoeeeeeeeee Vascular type of hemangiosarcoma with capillaryo like channels. Hematoxylin and eosin. x 187.5. . . Cellular type of hemangiosarcoma with reticular fibers. Reticulum stain. x 187.5 . . . . . . . . . Cellular type of hemangiosarcoma with fine reticular fibers. Note large nucleolus (arrow). Reticulum stain. x 750. b. O O O O O O O I O O O O O O O O O 0 Vascular type of hemangiosarcoma with thick bands of reticular fibers. Reticulum stain. x 750. . . . Hemangiosarcoma embolus in vein. Hematoxylin and eosin. x 187.50 O O O O O O O O O O O O O O O O O O Hemangiosarcoma embolus in artery. Hematoxylin and eosin. x 18705. O O O O O O O O O O O O O O O O O 0 Lateral radiograph of chest to show "snowflake" appearance of metastatic hemangiosarcoma in lungs . Ventro-dorsal radiograph of pelvic area. Note "punched out" appearance of left femoral head caused by hemangiosarcoma . . . . . . . . . . . . . vii O Page 55 55 56 56 57 59 59 60 60 INTRODUCTION "Tumors of animals are not fundamentally different from those of man. Therefore, a much broader view on the comparative pathology should be taken; in accordance with the trends of cancer chemotherapy, tumor pathogenesis, and etiology, natural cases of cancer in animals could be utilized to considerably greater advantage for the study of fundamental proba lems of cancer research than has been the case thus far." GMeier, 1963) This thesis is a retrospective study of 76 canine hemangiosar= comas collected from 1956 to 1965. The tumors were either diagnosed in dogs submitted to the Michigan State University Veterinary Clinics or diagnosed on the basis of specimens referred to the Department of Pathology by private veterinarians. The specific objectives of this study were to learn more about (1) epizoology, (2) histologic characteristics, (3) radiographic characteristics, and (A) diagnostic criteria. REVIEW OF THE LITERATURE Arey (1954) stated that both the blood cells and the blood vessels arose from mesenchyme. The earliest definitive cell with this potential was called the angioblast, while the process of vessel development was known as angiogenesis. Originally, there was a solid mass of these angioblastic cells which eventually developed lumen. In the process, the peripheral cells formed flattened endothelium. For a short time, the arteries and veins were not distinguishable. Ram (1957) pointed out that normal endothelium was a continuous ' lining formed by the junction of l endothelial cell with another. Palade (1953) demonstrated in a detailed ultrastructural study that in addition to the usual cellular elements (nucleus, mitochondria, endoplasmic reticulum, and small granular components) endothelial cells possessed 2 additional characteristic structural elements: (1) intracellular fibrils of 240 2 diameter and (2) a large number of vesicles concentrated immediately under the cell membrane facing both the capillary lumen and intercapillary spaces. These averaged 650 X in diameter. ‘Many of these vesicles appeared to Open at the surface of the cell membrane. In addition, a thin, irregular layer of relatively dense material surrounded the outside surface of the endothelium. In places it appeared as a belt of extraordinarily fine fibrils. This layer corresponded to the basement membrane. In addition, Stout (1943) pointed out that the growth of capillaries in granulation tissue occurred by the sprouting of endothelial cells 2 3 from pre-existing capillaries, forming first a solid cord which secon= darily became canalized. Referring to hemangiosarcoma, he stated that "probably all of the malignant tumors exhibit this type of growth in their infiltrative growth". Joest (1924) cited Leisering (1861) and Siedamgrotsky (1874) as being the first 2 authors to describe tumors of vascular origin in the dog. Leinaux (1899) described 2 dogs with hemangioendotheliomas of the parotid region. However, these appeared to be benign tumors. Ragenbagen (1907) reported 4 hemangiomas of the cavernous type in dogs between 6 and 10 years of age. One of these animals had 2 hemangiomas. Kingman and Newsom (1918) were the first American authors to publish a case report of a malignant tumor of endothelial origin in animals. They described an aged German Shepherd dog which had the habit of opening a screen door that had a stout spring on it; as the dog passed through, the door usually struck hflm on the side. An enlargement in the area occurred over a semanth period, measuring 8 x 14 x 4 inches. The tumor was soft, and upon insertion of a small cannula, blood exuded. At necrOpsy examination, nodules of the same type as on the side of the body were observed throughout the lungs. MicroscOpically, the tumor had large cavernous spaces filled with blood, although in some areas there was little blood. A diagnosis of malignant endothelial blastoma was made. I Feldman (1932) classified this tumor in his textbook on veterie nary oncology as a malignant hemangioendothelioma. He described the tumor as usually localized but possibly metastasizing tO‘widely separated areas of the body. It usually infiltrated tissue, but it 4 was exceptional for a hemangioendothelioma to become severely deo structive. Grossly, the tumor was typically dark blood red to purple and had a soft consistency. 0n pressure blood exuded, to be replaced when pressure was relieved. If punctured, hemorrhage was profuse. MicroscoPically, there was a tendency to form multiple, irregular, and immature capillaries. The capillary spaces or cavities sometimes varied in size. The cell lining tended to push into the lumen and eventually filled most of it. Frequently, solid nests or irregular sheets of cells resembled a sarcomatous structure. Same had small whorls with no lumens. Blood disintegrated due to being removed from circulation. The type of cell also varied. Many cells were oval or polyhedral, or sometimes flattened and almost spindle shaped. The cytOplasm was basophilic and the nuclei, which were fairly large and oval, contained many fine chromatin granules. In some, mitotic figures were numerous.. ‘Mallory (1914) described the tumor as it occurred in man and referred to it as an endothelial blastoma. Shennan (1914) published a case report for the reason that although this tumor was histologio cally benign, it metastasized. There was no sarcomatous transformaa tion in the tumor. Other authors (De Navasquez, 19363 Thiel, 190%; Langhans, 1879; and Geshickter and Keasbey, 1935) stated that while malignant tumors of blood vasculature were rare, they occurred and were usually benign in their appearance microscopically. Magnusson (1934) stated, "From foci with endotheliumolike cells there are all sorts of transitions to tissue resembling fibrosaro coma with a fairly great amount of strama. ... It is impossible to decide which cell of the tumor originates from the endothelium and which comes from the adventitia. 5 ... The tumor consists partly of solid parts, partly of various sized blood-filled lumina lined with endothelium. The lumina are mainly bound by low cells provided with large lancetwshaped nuclei with a well developed chromatin network. The lining cells of some of the smaller lumina, particularly in the more central parts of the tumor, differ from this by exhibiting abundant polymorphism from round to cuboidal forms to lower and flatter cells." Robinson and Castleman (1936) summarized this form of the tumor and gave it a specific title - "benign metastasizing hemangioma". They concluded that "the occurrence of metastases should be the deciding factor and not the histologic features" and that therefore "the pri» mary tumor, in spite of the absence of histologically malignant properties, was not benign". Freilich and Coe (1936) first reviewed the literature on "sngio= sarcomas" from 1918 to 1934. They were able to locate only 29 reports of this tumor in man. Lasserre 3;,gg, (1938), in their studies of cancer in dogs from 1933 to 1937, noted "the majority of blood vessel tumors are localized in the subcutis”. They concluded that the average age of dogs afflicted was 4 to 5 years. Stout (1943), in his detailed and authoritative article, pointed out that, "In spite of the fact that blood vessels are ubiquia tons, and benign vascular tumors exceedingly common, malignant tumors of blood vessels are exceedingly rare. Just how many of them have been recorded is impossible to say, because many tumors have been reo ported as such with insufficient or obviously erroneous data. After reading reports of 118 cases labeled with some name suggestive of a malignant vascular tumor, the writer felt compelled to reject 41, or 35% of them, either because there was an inadequate or no histologic report or because, in his Opinion, the illustrations and text described a tumor of some other type. ... Usually such tumors, even if they appear successively and grow to a large size, are not 6 suspected of being malignant growths. ... There are two features of the growth of hemangioendothelioma which are striking and uniformly present in all true tumors of this kind, although they are sometimes masked and require differential staining clearly to demonstrate them. These are: first, the formation of atypical endothelial cells in greater numbers that are required to line the vessels with a simple endothelial membrane; and, second, the formation of vascular tubes with a delicate framework of reticulin fibers and a marked tendency for their lumens to anastomose. No tumor should be considered a hemangio= endothelioma unless these criteria are both present. The variations found are due to the marked variability in the number, shape, size, and tinctorial peculiarie ties of the malignant endothelia. They may be irreguo larly rounded or polygonal, or they may become heaped up and more or less fill the lumen. Usually, they remain within the vascular tubes but occasionally they grow outside of it and form solid sheets of cells." Stout, later in his article, described the use of the silver reticulum stain in differentiation of hemangioendothelioma from other malignant tumors. The silver reticulum stain will "cause the tubes to stand out in sharp relief, because each one, even the most malignant tumor, has a delie cate fibrous supporting framework and the silver brings out a distinctive pattern. This pattern will be revealed with silver even.when, with other stains, it is entirely obscured by an overgrowth of cells. When ordinary stains make it impossible to decide whether cells are grouped inside or outside of the vessel wall, and so to detenmine whether one is dealing with endothelia or pericytes, silver staining of the vascular reticulin sheath shows the exact relationship of the cells beyond peradventure." He also discredited the entity "benign,metastasizing hemangioma" and "doubts its existence". . Kinkade (1949) published a comprehensive review of 9Uangiosarcoma". It dealt'mainly‘with the literature after Freilich and Coe“s review . of 1918 to 1934. 7 Hemangiosarcomas have been experimentally produced in laboratory animals by the use of chemical carcinogens. Rigdon (1952) applied 20~methylcholanthrene to the skin; Steiner (1942) implanted a pellet of it in the salivary gland; and White and Stewart (1942) added it to the diet. Furth and Furth (1938) injected 1:2abenzpyrene into the spleen. Andervont g;_g;, (1942) injected Ocaminoazotoluene into the axillary region. Edwards g£_g;, (1942) discovered a hemangiosarcoma in the liver of a 03H strain male mouse, which.was transplantable to other mice. Metastasis occurred in 2 of the mice after the implantation of first generation tumor cells, but these metastases were restricted to the liver. The transplants resembled the original tumor in every respect. A few case reports of hemangiosarcoma are in the veterinary literature (Kingman and Newsom, 1918; Lindsay and Gilmore, 1946; Pounden and Sprunger, 1947; Liebenman, 19S5; Farrell and Farrell, 1960; Simpson, 1960; and Jgplid, 1961). Weipers and Jarrett (1954) described a specific type of hemangioma which occurred in the scrotum of dogs. Although the tumor had the tendency to recur histologically there was no evidence of intrinsic malignancy. The reported incidence of canine hemangiosarcoma varied with different studies. In all articles it was reported that the tumor occurred rarely. Mulligan (1961) reported the occurrence of hemangioo sarcoma of the dog to be about 2.6% of all tumors submitted, and 6% of the malignant tumors. This compared with 8.81 of tumors submitted classified as hemangiOpericytoma and 0.91 as benign hemangioma. He found a distinct sex difference in his study; 14 males to 6 females. 8 In metastatic hemangiosarcomas the lung was the most commonly affected organ, followed in frequency by the liver, regional lymph nodes, and heart. The most common sites of primary occurrence were the spleen and the subcutis. Mulligan (1949) was the only author to describe breed predi8p0= sition. He pointed out that "the Gemman Shepherd is definitely pre= disposed to neoplasms of endothelial cells“. In reviewing standard textbooks of pathology, both veterinary and human, the subject of hemangiosarcoma was treated as an insignifi= cant topic in some, while others had exceedingly long discourses describing an individual tumor. Willis (1948) questioned the value of a separate classification for tumors of endothelial cells when he pointed out, "Let us recall once again the intermutability of the various mesenchymal tissues, the close kinship of their tumors, and the fact that vascular tissue is, of all the mesenchymml tissues, the most ubiquitous and one of the most plastic. Then we will be pre= pared to regard true angiomas and angiosarcomas, not as fixed species, but merely as conspicuously vasoo formative variants of the genus mesenchymoma." Herbut (1955) pointed out that hemangiosarcomas in man were seen in all ages but there was a preponderance in childhood and youth. Both sexes were equally affected. Pack and Ariel (1958) noted that these tumors tended to occur in the subcutaneous tissues of the extremities. Robbins (1962) asserted that hemangiosarcomas were often observed in the skin, liver, spleen, lungs, and bones. Most authors stated that although the tumor characteristically produced endothelial-lined spaces containing blood, in many of the less differentiated tumors sheets of cells were laid down with no vascular characteristics. 9 Robbins described the use of a phosphotungstic acid a hematoxylin stain (PTAH) for the differential diagnosis of hemangiosarcoma from fibrosarcoma and leiomyosarcoma. The radiographic characteristics of this tumor are unique and interesting. Golden (1959) in his textbook of radiology described the distribution of metastases of malignant tumors to the lungs. There were 3 types cf metastatic neOplasm.to lungs according to ape pearance. They were: (1) nodular, (2) lymphangitic, and (3)Jpneué ‘monic. Sarcomas, including hemangiosarcomas, usually assumed the nodular form. The metastases were almost always multiple with sharp borders, uniform in density, usually round, and of varying size. They were usually more numerous in the lower portions of the chest° When numerous they were usually of uniform size. Bradley and Coley (1960) pointed out that hemangiosarcomas ina volving bone were difficult to differentiate from osteosarcomas. Reactive bone was laid down adjacent to the tumor in layers and resembled osteomyelitis. Hemangiosarcomas had a tendency to have their primary occurrence in the shaft of a bone rather than at the epiphysis, and when they occurred as a primary bone tumor, they had a tendency to metastasize to other bones and to the lungs. According to De Lorflmier (1954), hemangiosarcomas of the bone 1 \ were exceedingly rare - the rarest of primary bone tumors. They were characterized as bulky, cystic, or telangiectatic tumors and were usually single, although multiplicity could eventuate. The tumor grew rapidly, with progressive destruction and with early dissolution through the cortex but relatively late metastasis. He further stated that the basic radiographic manifestations were 10 (l) conspicuous rarefaction, (2) possibly slight expansion of the cortices, (3) probably dissolution through the cortices and extension to regional soft tissue, (4) little or no osteosclerotic reaction, and (5) moderate response to xmray therapy with demineralization. Important in a differential diagnosis would be a giantmcell tumor or ‘metastases of some other primary tumor such as an adenocarcinoma of the kidney. Lichtenstein (1952) suggested that hemangiosarcomas in bone were likely to be aggressive and so far advanced by the time they were recognized that a grave prognosis was indicated, despite radical surgical procedures. Ritvo (1955) pointed out that the tumor characteristically tended to extend through layers of new bone with the formation of parallel layers of bone above like an onion skin. This lesion resembled proliferative osteomyelitis. The fastest bone growth or reaction was at the edge of the tumor. Radiographically, the tumor usually appeared as a diffuse osteoporosis. As the tumor invaded soft tissues, the shadow was not as distinct as before. Ritvo furthenmore stated, "it is the only malignant tumor which arises from the cortex of bone". Lesser (1955) described an interesting technique for the demon0 stration of the vascularity of tumors in bone. He used a radiOo opaque solution which was injected into the blood vessels that supplied the bone tumor. Radiographs were then taken. He found that the various kinds of tumors in bone have a characteristically different vascular network. The object of the study was to help in the diago nosis of prflmary bone cancers. 11 In veterinary medicine, Carlson (1961), in his textbook of radio ology, described a situation similar to that seen in man. This was especially true with reference to this tumor when it metastasized to the lungs. MMIERIALS AND METHODS Preparation of Material All clinical tissues were fixed in 10% neutral formalin. Each clinical tissue carried a number from the Veterinary Clinics and a number from the Department of Pathology. The clinical numbers were arranged in chronologic order beginning with l and continuing to over 100,000 at the time of completion of this study. Numbers from the Department of Pathology were listed according to alphabetical letter followed by a series of digits in chronologic order. The letter "A" covered the fiscal year 1956-1957, the letter "B" covered the fiscal year 1957-1958, etc., continuing to the last year of dais study, the letter "J“, covering the fiscal year l964°1965. The letter "I" was omitted so as not to be confused with a numeral. After the tissues were submitted, each tissue was trhnmed, ‘ embedded, and sectioned at 6 microns. Each section was stained with hematoxylin and eosin (H & E) (Armed Forces Institute of Pathology, 1960, page 30). Tissues referred by private practitioners were usually fixed in 10% neutral formalin. Selected slides in this study were stained with Gomori's iron reaction stain (Armed Forces Institute of Pathology, 1960, page 151). A second special stain used was a modification of reticulum technique as outlined by Lillie (1954). An outline for the staining technique is as follows: 12 13 l. Deparaffinize sections through 2 changes each of xylene, absolute alcohol, 95% alcohol, and distilled water. 2. Oxidize for 20 minutes in 0.5% periodic acid and rinse in distilled water. 3. Dry thoroughly in dryer (heat). 4. Lay slides face up on glass rods over large pan and deposit on each about 1.5 to 2 cc. of diamine silver hydroxide.* Let stand 3 minutes and decant. 5. Rinse quickly in distilled water. 6. Reduce 2 minutes in 10% neutral formalin and wash 3 minutes in running water. 7. Tone 2 minutes in 0.21 acid gold chloride (by visual examia nation) and rinse in tap water. 8. Fix 2 minutes in 5% sodium thiosulfate (not more) and wash in tap water. 9. Counterstain for 5 minutes with nuclear fast red.** 10. Rinse in distilled water and dehydrate through ascending con- centrations of alcohols. 11. Two changes of xylene. 12. Mount with permount. *Diamine silver hydroxide solution: Place 1 cc. 28% ammonia water in a small flask. Add 10 cc. of 10% silver nitrate, the first 7 to 8 cc. fairly rapidly, the rest cautiously, shaking between each addition to clear the brown clouds of silver oxide until a faint, slightly brown permanent apalescence remains. Add an equal volume of distilled water (solution to be made approxflmately 10 minutes before -staining). In preparation of this stain all glassware must be acid clean. **Nuclear fast red solution: Dissolve 0.1 Gm. of nuclear fast red in 52 solution of aluminum sulfate with aid of heat. Cool, filter, and add a grain of thymol as a preservative. 14 Study Procedures The case files of the Department of Pathology from 1956 to 1965 were reviewed. All case reports of hemangiosarcoma in the dog were retrieved, along with the corresponding slides. If a case originated from the university's mmall annual clinic, the clinical history was obtained. These histories were reviewed and epidemiologic infonmation from each was recorded in table form, similar to Brodey (1960)o Column headings included pathology number, clinic number, breed, sex, age, primary location, size of tumor, tentative clinical diagnosis, rate of development, and metastasis. The H~& E-stained slides were examined by the author and 2 patholoa gists of the Department of Pathology for review and confirmation of the original diagnosis. The microscopic characteristics of each tumor were catalogued according to cell morphology, size and shape of nuclei, size and tinctorial qualities of nucleoli, and tendency for cytoplasm.to fuse. The mitotic figures were classified according to their frequency and types of abnormal forms. The neOplasme were graded for hyperchromatism, invasiveness, lack of organization, and type and size of vessel formation. The section was examined for necrosis, amount ind type of stroma, invasion of preeexisting blood vessels or lymphatics, and mdscellaneous characteristics, such as hemorrhage and inflammation. \ The reticulum-stained sections were evaluated for the amOunt, type, and structure of the reticulum present. The sections stained for iron were evaluated for identification of iron pigment within macrophages which were present in some of the neoplasms. 15 Cases fran the smalljanimal clinic with radiographs were studied with regard to the characteristics of the tumor in or near bone and metases to the lungs. Gross photographs and photomicrographs were taken of selected lesions. All measurements taken through the microscope were obtained by use of an ocular micrometer. The ocular micrometer was calibrated in a routine manner by use of a stage micrometer. TERMINOLOEY This tumor of endothelial cell origin has been given many difa ferent names, most of which are confusing to the student of pathology. For this reason, the following discussion on tenminology is presented. The following terms have been used in describing this tumor: angiosarcoma, hemangioendothelioma, angioendothelioma, endothelioma, hemendothelioma, angioblastoma, intracranial angioblastoma, angioa fibrosarcoma, endothelial sarcoma of blood vessel, fibroangioendotheli= ome, hematoblastoma, hemangioblastoma, hemangiosarcoma, angioblastic seaplane, angioblastic meningioma, malignant aneurysm, capillary angina sarcoma, capillary sarcoma, hemangioendothelialblastoma, hemangio= endothelialsarcoma, angioblastic sarcoma, telangioendothelioma, telangiectatic sarcoma (American Cancer Society, 1953). Mbulton (1961) defined the tumor as "a malignant tumor of endoe thelial cells". The terminology is just as much a problem today as' in the past, since most authors still disagree on the term that should be used. Anderson (1961) used the term.“malignant hemangioendothelioma" or "angiosarcoma”. Herbut (1955) used both "hemangioendothelioma" and "hemangiosarcoma". Boyd (1958) used the term Whemangioendotheliomam’, with the adjective "benign" or "malignant", depending upon the cell type. He pointed out that the tumor‘would more conveniently be called "hemangioendothelial sarcoma". Pack and Ariel (1958) used the term.”angiosarcoma", which is a relatively older tenm. 'Robbins recently (1962) used the term "hemangioendothelioma“ to "represent an 16 17 intergrade between the well differentiated hemangiomas and the frankly anaplastic, totally cellular hemangioendothelialsarcoma". He also stated that "certain writers use the term.hemangioendothelioma to refer to a malignant endothelial tumor. This usage is confusing and inappropriate, since the malignant counterpart should properly be referred to as a sarcoma". 0n the other hand, ”hemangioendothelioma" seems to be a popular term in the veterinary literature (Moulton, 1961). In this study the term."hemangiosarcoma" was selected for the malignant neoplasm for the following reasons: 1. The term hemangioendothelioma was most inadequate and inap= propriate for the reasons stated by Robbins (l962). 2. The term hemangiosarcoma has been recommended by the National Cancer Institute (Public Health Service, U. S. Department of Health, Education and Welfare, 1964). 3. The word hemangiosarcoma connotes both malignancy and histoo .genesis. Hopefully, progress has been made in limiting the tenminology that is used in referring to this tumor. In a recent publication (American Cancer Society, 1965), only 4 terms were recognized for ' this neoplasm: hanangiosarcoma, angiosarcoma, malignant hemangioendoa thelioma, and hemangioendothelialsarcoma. RESULTS Occuggenceggpd Gross Characteristigg TABLE 1 is a compilation of the salient features of each case as outlined in'Materials and Methods. Breed incidence. The breeds which had the highest incidence of hemangioa sarcomas were Cocker Spaniel (14), Boxer (14), German Shepherd Dog (6), and Labrador Retriever (4). The tumor was diagnosed in 22 other breeds. The number of hemangiosarcomas based on the weight of the dog is shown (Figure l). The method for categorizing the dogs by weight was that used by Tjalma (1965) in his study of bone cancer. The largest number of hemangiosarcomas occurred in the large breeds of dogs. Sex distribution. In those animals of known sex, a total of 44 hemangiosarcomas was found in the male and 25 in the female, or a ratio of 1.76 males to 1 female. The influencing factor for the sex difference was the great preponderance of the tumor in male dogs of the large breeds (Figure 1). Excluding this group, the ratio of males to females was approximately equal. Age distribution. The number of hemangiosarcomas according to age is presented (Figure 2). The mean age was 9.0 years. The mean age for males was 8.5 years, while the mean age for females was 10.7 years. 18 19 TABLE 1. Occurrence and gross characteristics of 76 canine hemangioa sarcomas. 'Tath. Clinic _ , Age No. No. Breed___ Sex (yrsg) A325 --~ --~ oe- an: A522 --- can new use A1503 --- Doberman Pinscher M ace A1983 7163 Cooker Spaniel M 6 32412 -=- -== M “4 B3408 32424 Cocker Spaniel M 13 B4928 32908 ‘Weflmaraner F 4 B6134 33282 English Setter M 8 C376 ~-- Cocker Spaniel M 8 C720 --= Cocker Spaniel F 12 C858 33849 Boxer M 8 C3343 --= Mixed F 173 05391 «6- Boston Terrier M 8 65497 7163 Cocker Spaniel M 8% 65550‘ 30359 Boxer M 5 05838 36238 Pointer F 3% C5887 -=° Labrador Retriever F 6 D683 --= Boxer M 9 20 r.- u 'n W Size! Rate of Primagz Location 4(am.)g Dewglgpment Metastasig‘ , mouth ---. as: use I tongue 1.1 x .7 x .5 -s- .9, 3rd premolar area -=- can as- 1. ear “°° --- none free in abdomen mouth femur r. prox. foreleg 1. scapular area 1. cheek.vent. to ear prox. r. femur liver (caudate lobe) muscle between scapulae spleen sheath of penis I. femur and adj. area . upper Jaw 1. front leg 5 die. from acetabulum to mid shaft 7 dis. 6 x 3 x 2 large 6 x 9 l encircled femoral neck; 2 others, 2 dia. 4.5 length 9-0 rapid=3 wk. several mo. 2 mo. 6 mo.=1 yr. 4-wk. rapid=2 mo. numerous masses in abd. cavity mandibular l.n. regional l.n. lungs none lysis of bone none ”C06 numerous nodules in lung, l.n. inguinal l.n. pres viously removed CBC: lysis of bone cervical l.n.. lung. liver, kidney, heart TABLE 1--continued 21 IBath. Clinic Age No. No. Breed Sex (yrs.) D1250 --- Beagle Hound F 11 ' D1349 5820 Dachshund F 12 D3651 _ 38357 Cocker Spaniel M 10 D3752 --- Cooker Spaniel M 12 D4327 -°- Boxer M 11 D4615 38631 Boxer M 7 D5863 37853 Setter M 12 E1019 --° Fox Terrier F 14 E1628 38310 Kerry Blue Terrier M 10 E2374 --3 Boxer M 6 E3701 ~-- Cooker Spaniel oaa 8 E3900 ~-- German Shorthaired F 10 Pointer E4161 757 Cocker Spaniel F 8 E4186 37149 German Shepherd Dog M 7 E4227 --- Dachshund M 9% E4792 --- Boxer F 11 E5197 41651 German Shepherd Dog M 11 E5432 40841 Springer Spaniel M 9 22 _ Size Rate of Prflmarz Location (om.) Development Metastasis mesentery 2 nodules=6x9 rapid liver, spleen dia. spleen 5 dia. =09 liver spleen -~~ 2 wk. none liver 7 x 12 unknown lungs, kidney, intestines -s° °°- -°~ small intestine, lung spleen -e° -°~ mesentery, lung, liver- spleen 11 x 16 === none prox.l/3 radius ==- ace so» ~-- °-- coo numerous l.n., lung -9- === can numerous nodules, lungs S.C., r. ll-l3th 10 dia. rapid=2 mo. nae rib mammary gland 4 dia. can 9-9 elbow 5 x 7 1 yr. coo mesentery «=9 as, kidney, liver, 'mesentery, lung S.C., scrotum 1 dia. rapid none liver, spleen --- slow none lg. muss involving kidney, spleen, omentum, liver see primary location TABLE l-econtinued 23 an"... - n— nun —..) Ffih. Clinic Age No. No. Breed ‘__f Sex (yrs,) E5763 41222 German Shepherd Dog M 11 E5907 -~- Cocker Spaniel M 9 F247 ee- Terrier F 13 F814 --- Terrier F 11 F890 ..9 English Setter M 5 F1494 42828 Dalmatian Coach Dog M 12 F2644 -2- Golden Retriever M 7 F3788 43982 Boxer 'M 5 F3970 --- Boxer F 9 F5137 ee- Scottish Terrier F 13 F6173 -s- German Shepherd Dog M 8 C107 ~== Labrador Retriever eee eee G294 -o- Labrador Retriever M 13 @1013 one Pointer F 14 63833 39715 Boxer eee 4 G4150 --- Poodle FCS) 11 G4373 --- Dachshund M 4 G4446 =-- German Shepherd Dog MI 10 @4752 eee Cocker Spaniel F 10 64792 -~- Boxer eee 11 C5327 47472 Cocker Spaniel M 5 24 sax-Vii um-rl-rq,xrfl 7"".- r'fi m» n Size Rate of Primagx Location _7 (cm.) Development Mbtastgsis S.C. of l. flank 3 dia. slow~4 mo. none spleen 3 dia. rapid none S.C., inguinal 2.2 dia. °== spleen, lung, heart, area jejunum, kidney base of tail nee * eee Skin Baa one out: S.C., r. side 9-- one none post. last rib r. kidney 5 dia. =°= none S.C., to Side Of 3°” 969 can chest r. nares -°~ 5 mo. none lung small, numerous «we none pancreas, duodenum --- «an ce- post. antebrachium 4.5 die. 5 mo. none liver, spleen variable 2 mo. mandibular, mediastie ‘ nal l.n. various 1 x 4 dia. 366 mo. liver, spleen, l.n., heart, lung, kidney spleen 10 dia. rapid none ovary 4 x 3 x 2 eee none spleen .18 dia. sudden none S.C., cervical l x 3 rapid lung, kidney, cereo brum, cerebellum spleen 4.5 dia. °°° none ovary --° -=° spleen, omentum r. oral commissure 2 dia. none TABLE la-cont inued 25 Path . Clinic Age No. No. Sex (yrs ,1 G6848 --- Pointer M 9 H472 48205 Poodle M 7 H1075 48483 Standard Poodle F 795 111202 6..., German Shepherd Dog M 8 112156 "e Boxer M 10 1 H2796 -9- Beagle Hound F «we 114957 a“ Boxer M 11 115628 a..- Boxer M 11 E5657 9... Bassett Hound FCS) 9 H6057 -..- Cooker Spaniel F 13 H6119 101469 Viszla M 5 H6360 31766 102835 (Golden Retriever M 7% .1261]. we Wirehair Terrier F 6 .12652 cm- Miniature Boodle M 8 .12691 103252 Labrador Retriever F 7 J2814 -==- m... “a 12 J2958 103080 Cocker Spaniel F 13 J3526 --~ g... M 10 J4075 103820 Airedale Terrier M 1075 ‘- _ _.. v— ‘r 26 -W n—eLe-fi r-—-l Size Rate of Primary Location Lem.) Development Metastasis --- -=- one lung, omentum S.C., ant. to penis 0.2 x 0.6 slow none spleen 2 x 3 dia. ewe liver, lung 1. hind leg muscle ace one heart, lung, kidney, mesentery S.C., prepuce “°° 1 mo. eee spleen, liver -~= 2°3 wk. spleen, liver bladder one 1=2 mo. lung r. femur e-a use liver, kidney, lung 1. h0¢k(.°ft 1:218.) °°° 000 one 3rd phalanx --- ~69 === S.C., vent. abd. e-e slow numerous l.n., mesene tery, peritoneum, . diaphragm, stomach; nmmole of abd. wall lung none one heart, lung spleen -=° one sea as- eee eee lung, heart diaphragm 7 dis. ace pleura, ileum, perie cardium S.C., inguinal 5 dis. 3 mo. nae area intestine S.C., l. metacarpus lung, spleen prescapular l.n. —-—w- TABLE le-continued Path No. A - July 1, 1956, B - July 1, 1957, C - July 1, 1958, D - July 1, 1959, E - July 1, 1960, to June 30, to June 30, to June 30, to June 30, to June 30, 27 1957 1958 1959 1960 1961 (4530'!!! July 1, July 1, July 1, July 1, 1961, to June 30, 1962 1962, to June 30, 1963 1963, to June 30, 1964 1964, to June 30, 1965 Note: Cases A1983 and C5497 are from the same animal but were counted as 2 cases since 2% years had elapsed between occurrences. Cases H6119 and H6360 are from the same animal and are counted as 1 case because of the short period of time between occurrences. Sex M - male Primagy Location abd.' abdomen r. right prox. proximal 1. left Size dia. diameter F - female vent. adj. S.C. post. Rate of Development ventral adjacent subcutis posterior F(S) - female, spayed ant. 1g. l.n. tis. anterior large lymph node(s) tissue * Present for several months, increasing rapidly last few days. 28 80' Giant 40-80' Large 20-40! Medium - Females 20! Small - Males - Unknown unknown 0 5 10 15 20 25 30 35 40 45 Number of Cases Figure 1. Incidence of hemangiosarcomas in male and female dogs of different weight classes. 10‘{ I O K Total cases 8‘ ' "0. “819. of 6‘ ' Cases ' o . ' 1O 4 // ‘\\ I/p‘\ . Females ’ \ ‘ 2.. p....\( ’D-m-O/ ‘0’ / ‘\ I \O’ 6 7 8 9 10 ll 12 l3 l4 Age in Years Figure 2. Age and sex distribution of hemangiosarcomas in dogs. Primary location. The most common primary sites were the skin, including the subcutis (24), and the spleen (15), followed by bone (6), and liver (Sfi. The primary location'was also classified by the area of body involved. The most commonly affected area was the trunk (39) followed by the extremie ties (16) and head and neck (10). Metastases. Fifty-four histories had information on metastases. Thirtyo two (59%) had metastasized to various organs. The lungs'were the most common site (19), followed in frequency by distant lymph nodes (11), liver (9), kidney (8), mesentery (7), spleen (6), heart (6), and regional lymph nodes (5). Typical numerous metastases are shown (Figures 3 and 4). Rate of develoggent. The rate of development was obtained from the history, given by the owner or the veterinarian. The rate was divided into 3 categories: (I) O to 1 month, (2) 2 to 5 months, and (3) 6 months and over. Six neoplasms were placed in the 0° to lemonth, 11 in the z- to 5mmonth, and 3 in the 6 months and over category. Growth rate of 3 tumors was reported to be slow and 6 rapid. Size and shage of tumor. Tumor size varied from a few millimeters to several centimeters. It was often round, bluish-black, and poorly encapsulated. The metastatic lesions were usually round and less hemorrhagic than the primary lesion. The typical gross appearance of a primary lesion on cut surface is shown (Figure 5). A single, circua lar metastatic lesion‘which has ruptured resulting in an adjacent clot is illustrated (Figure 6). 30 Q "-.-. ‘ . . ‘ «(,- . ‘ ImmmlIlmumnmmxr.um“mus--.._- . A Figure 3. Metastatic hemangiosarcoma in the lungs. Figure 4. Metastatic hemangiosarcoma in lung, spleen, kidney, and liver. IIIITTTTTTTTVYT 1- “NH All Figure 5. Cut surface of hemangiosarcoma in diaphragm (primary site). 5 AA -A v A A v a v -9.“ v o- v V v J 269! L4 SAY 6v 7v-” 9_ IO 7” 1; WWW" - Figure 6. Metastatic hemangiosarcoma in small intestine with adjacent blood clot. 32 Recurrence. In 11 cases where information.was available, 7 were known to recur. The diagnosis was made from necropsy specimens in 32 cases and bioPsy specimens in 26 cases. Tentative diagnosis. Although not recorded in TABLE 1, the tentative diagnosis by the clinician was noted in each case where this infomma= tion.was given. Of the 21 cases in which this information was found, only 6 were diagnosed as hemangiosarcoma. §g5roscopic Characteristics Tissue sections of hemangiosarcomas from 76 dogs were examined microscopically. The tumor characteristics are presented (TABLE 2). Shape of the cell. The cells generally were spindle shaped except that in the more anaplastic tumors the cells were pleomorphic (Figures 7 and 8). Size and shape of nuclei. The shape was elongate, oval, round, or pleomorphic. In general, the round or pleomorphic nuclei were larger, more vesicular, and more anaplastic (Figure 9). The size was variable. The elongate nuclei ranged from 3.x 12 microns to 6 x 16 microns. The oval nuclei, which.were by far the most common type seen, ranged from 6-x 8 microns to 12 x 18 microns, while the round forms were 7 to 18 microns in diameter. Size measurements were usually not determined on nuclei which were pleomorphic. Size and stainigg properties of nucleoli. There was a great variation in the size of the nucleoli, from those which could barely be seen to those as large as 7 to 9 microns in diameter. The larger fonms were usually seen in the large oval and round nuclei (Figure 10). They were invariably round, and there was a correlation between the size andithe staining characteristics. The smaller nucleoli (0.5 to 33 TABLE 2. Microscopic characteristics of 76 canine hemangiosarcomas. Size & Size & Types Hypera Path. Shape of Shape of Color of Mitoses/ of Mia chroma= No. Cell__ Nuclei (P) Nucleoli_1#) hpf toses tigm A325 spindle oval 7-9 dia. 0=1 bizarre + (9313) single & to round multiple (20 dia.) eosinophilic A522 spindle oval 0.5a1 0 was -++ (7x10) basOphilic A1503 pleomorphic round-(ll) 465 2 bizarre -++ to spindle to oval multiple (10312) eosinophilic A1983 spindle oval 102 Gal normal -++ (9x12) single . eosinophilic 32412 elongate oval 0.561 0=l blast ++ to spindle (9x12) single basophilic 33408 spindle oval l 0 «co ++ (7x13) single not many B4928 pleomorphic elongate 1 2-3 blast ++ to spindle (6x16) to few; most round (15) not seen B6134 spindle oval most not seen; 1 blast ++ (8x15) a few up to 3 C376“ elongate elongate 2a3 1 blast & ~++ ' to spindle to oval single bizarre “L to pleo- (4315) to eosinophilic morphic round (12) C720 spindle oval only a few 0 mac ++ (6x8) observed .5-1 single basophilic C858 pleomorphic round (8) none 0 -0- ++ observed 34 ‘fl'n —t---'-l — 1— r 1") a-rim 7—. m ‘fi!"""nm" IL—g‘fi 1 - a Invasion of Degree of Preoexisting Angplgsig Amount of Blood or Miscellaneous_ Cellular Vasculgr Necrosis Stroma Lymph Vessels Infl. Hem. Maggo, W H in erOded + use: + + 0 area * ‘++ ‘+++ in eroded ++ no '+ + 0 area 44- +++ in eroded + oe- + + 0 area + ++ in eroded ++ no + + 0 area -+++ in areas + «an 0 + O of hemore- *very rhage little 444- in eroded + nae +' + 0 area +++ +++~ in areas +++ no 0 + 0 of hemor- rhage +++ none ++ yes 0 0 0 * +++~ +44- in areas ++ yes 0 + 0 of hemor° * ‘ rhage +++ none +- ~o- 0 + 0 * +++ none + no 0 + 0 35 TABLB‘Zn-continued Pin—w» q Size & Size 8 Types Hyper» Path. Shape of Shape of Color of Mdtoses/ of use chroma= No. Cell Nucleigju) Nucleoli (u), hpf toses tism C3343 pleomorphic round 3-4 1 blast & ++ (9t015) single bizarre eosinophilic 05391 spindle to oval l=3 Ool blast -P+ pleomorphic single & multiple eosinOphilic 05497 very pleo~ round (15) 2=3 3=4 blast AF++ morphic single 9 a few multiple eosinOphilic 05550 spindle oval 1=2 263 blast -++ (7x11) single & multiple 05838 spindle elongate l 2w3 nonmal +++ to oval single (5x14) basoPhilic 05887 pleomorphic oval 2 2 normal & +++ (10314) single & blast a few tua multiple mor giant eosinophilic cells D683 spindle oval none 0 an» ++ (6x12) observed D1250 pleomorphic elongate .501 3 blast ++ to oval basOphilic (4212) to round (10) D1349 spindle pleomorphic 4 no blast +++ to round single (9°10) eosinophilic D3651 Spindle oval 2°3 5 nonmal & ++ (9x12) to eosinophilic blast round (10) 36 Invasion of Degree of Preaexisting Ana lasia Amount of Blood or zmygscellaneous_ Cellular Vascular Necrosis Strgmgj Lymph Vesggls In§l$;fieg. Macrggp ++ +++ in areas ++ aa- 0 + 0 of hemoro *poor hage stain +++ in areas ++ -o= 0 +' 0 of hemor'3 rhage ifii- in center + yes 0 + a 0 of cellu°‘*very little lar areas in cellu- lar areas +++- in eroded ++ eon + 0 0 areas *very scanty +++ -++ none ++ one + + + * +++ none + one + 0 0 +++ in areas + -9- 0 + 0 of hemor- rhage -+++ in areas +++ yes + + 0 of hemora * rhage +4+- +44- in center + can + 0 0 of cellu= lar area +++ in areas ++ yes 0 + 0 of hemor- rhage 37 TABLE 2--continued ‘5' m1-_-—J-‘—'-. ‘--_>9 _- —. -~ 2-,...fi_ — ——!_...... m Size & Size & Types Hypere Path. Shape of Shape of Color of Mitoses/ of His chromae No. Cell Nuclei_(u)_ Nucleoli jg)» hpf toses tism D3752 spindle oval (6x8) 2=3 1 normal ++ to round eosinOphilic (12-14) D4327 pleomorphic round (12) .5=l 1-2 normal coo to spindle single basOphilic D4615 spindle elongate .5ol 1 normal & ++ (4x12) single blast ' baBOphilic D5863 elongate elongate 293 2=3 blast & +++ to spindle to oval to single bizarre round basophilic & eosinOphilic E1019 spindle oval .Sel 1 normal & +4» (7x13) single blast a few large (2-3) 81628 spindle to oval l=2 l blast + pleomorphic (7112) single eosinophilic E2374 elongate oval 'most not seen; 1 normal ++ to spindle some round few 05"]. basophilic E3701 spindle oval 2°3 1 blast -+k+ (8x17) to single 1 round eosinophilic E3900 spindle round l°2 2:3 normal & ++ (8-10) single blast eosinOphilic E4161 spindle to oval 1 5°10 blast & +++ pleomorphic (8x10) to single & bizarre pleomorphic multiple 38 Invasion of Degree of Preaexisting Anaplasia Amount of Blood or nflgfiscellanegugmn Cellular Vascular Necrosis Stroma Lymph Vessels Infl. Hem. Maorok 4%4- +++ in areas ++ yes 0 + 0 " of hemor- rhage +++ in areas + yes 0 + 0 of hemor- * rhage +++ none +++- yes 0 0 0 * +++ . in areas + no 0 + 0 of hemor- rhage l-+++ none +++ no 0 0 0 44- +++- none + --- 0 + 0 * +++ in areas ++' no +’ + + of hemor- rhage +++' in areas + one + + 0 of hemora rhage +++ none ++ ace 0 0 0 +4+- ++. in areas + yes + + 0 of hemor- rhage TABLE 2--continued Path. No. E4186 E4227 E4792 E5197 E5432 E5763 E5907 F249 F814 F890 39 Size & Size & Types Hyper- Shape of Shape of Color of 'Mitoses/ of Man chromao Cell Nuclei_(E) Nucleolifgg) ,hpf toses timm pleomorphic 1, oval 2 5 normal & +++ to spindle (12x18) to single & bizarre round (11) multiple pleomorphic round to l 3=4 normal -+++ oval single 9 (7313) few multiple pleomorphic oval .5=1 1 normal +++” to spindle (9x11) single - to round few mul» tiple spindle oval most not 1 normal & +++ (8x15) seen a a blast a few few .5=1 round (18) pleomorphic oval 3~5 5a10 bizarre +++ to spindle (11318) to single & round (16) some mule tiple eosin0philic very pleo- oval 1=3 2a3 normal & +++- morphic (8:15) to eosinOphilic blast ' round (18) pleomorphic oval to 2=3 3 bizarre &-+++ round to blast pleomorphic pleomorphic pleomorphic small (1) 1 normal +++ spindle to round (15- 2 l oae ++ pleomorphic 18) to single & pleomorphiC'multiple spindle to oval to 3=4 le2 normal ++ pleomorphic round single eosinOphilic 40 InvaSion Sf —— -—-——-— mgr-v—ammm‘m gum Degree of Preaexisting Anaplasia Amount of Blood or Miscellaneous Cellular Vascular Necrosis Stromgp Lymph Vessels Infl. Hem. Maogog +++ +++ none ++ no + + + ++’ 444- in areas ++~ no + + + ‘ of hemorn hage 44-, +++ none + no 0 + + very little none + no + + + tissue ++4- ++ in areas ++ yes + + 0 of hemore rhage +++ ++ none 44- yes 0 + 0 +++ ++ none + no 0 + 0 +++ +++ in areas +| no +' + + of hemor- rhage 44- +++ in areas + yes 0 + + of hemor- rhage +++ ++ none + yes 0 0 0 41 TABLE 2-ncontinued --- Size & Size & Types Hyper= Path. Shape of Shape of Color of Mitoses/ of Mia chroma° ‘39. Cell Nuclei (u) Nucleoligju) hpf toses tism F1494 spindle to oval l 2°3 blast ++ pleomorphic (9x11) to single round F2644 pleomorphic round 5 3 blast +++ (13-15) single eosinophilic F3788 spindle oval none 3=4 normal & (9x11) to observed blast -++ round F3970 spindle oval 1 2 normal & -++ (8x10) to single blast round (12) basOphilic F5137 pleomorphic irregular 3-4 5:6 bizarre +++ round to single oval eosinophilic (15x18) F6173 spindle oval to none 0 use +++ elongate (3x12) 0107 pleomorphic round (13) 3 0 one -+++ to oval single (9x16) eosinophilic 0294 pleomorphic round (15) 3=4 7~8 bizarre +++ single & multiple eosinOphilic 01013 pleomorphic oval 3 102 normal +++ (10:14) to single round (13) eosinophilic G3833 pleomorphic oval l 0 can ++ (8x10) single basOphilic 64150 spindle to round (11) 3 l=2 normal & + pleomorphic to oval single & blast (11x15) ‘multiple baSOphilic 42 Invasion of Degree of Pre-existing An lasia Amount of Blood or Miscellaneous Cellular Vascular Necrogis Stroma Lymph Vessels Infl, Hem. Macroz 444- in areas I + yes + + 0 of hemor- rhage '+++ none + yes 0 + 0 ++ +++ none + yes 0 + 0 +++ 44- none. + no 0 0 0 444- ++ in areas + yes + + 0 of hemoro rhage +44— ++ in areas + yes 0 +> + of hemor- rhage -+++ in areas + no + +- + of hemor- rhage +++ in areas l+ no +- + + of hemor° *very little rhage observed 4+4- none + no + 0 0 +++- none +++ yes + + 0 +++- in areas + no 0 + ~+ of hemor- rhage 43 TABLE 200continued Size & Size & Types Hyper- Path. Shape of Shape of Color of Mitoses/ of M10 chroma0 No. Cell fi_Nuclei (H) Nucleoli (g) _hpf toses tism G4373 spindle to oval 1 l 00- -++ pleomorphic (7x12) basOphilic 04446 spindle to round (7) 102 102 normal & '+++ pleomorphic to oval eos inophilic blast (6x11) 0752 spindle round to 203 102 normal &__ ++ oval single & blast multiple eosinOphilic 04792 pleomorphic oval 1 0 00- +++ to round (6x15) eosinophilic some large bizarre 05327 spindle pleomorphic 304 405 bizarre +++ single eosinOphilic 06848 spindle to oval 102 0 000 + pleomorphic (7x11) single & a few multiple eosinophilic H472 Spindle pleomorphic 102 0 000 ++ (9x11) basOphilic H1075 spindle to pleomorphic l 102 blast ++¥ pleomorphic (10x12), a single & few round multiple (15) basOphilic H1202 spindle oval 104 102 blast ~++ (9x12) to single & round (11) multiple basOphilic & eosinOphilic H2156 spindle oval most not 203 normal -++ (8x13) seen Invas ion of Degree of ~ Pre-existing Anaplasia Amount of Blood or > Miscellaneous Cellular Vascular Necrosis Stroma lygph Vessels Infl, Hem. Macro. +++ ++- none + no 0 0 0 +++ in center ++ no 0 + 0 of tumor * 444- +++ in areas ‘+ no + + 0 of hemor- * rhage 44- +++ in areas ‘ + no + + 0 of hemor- rhage +++ +++ in eroded + yes + + 0 areas <+++ in areas + no + + 0 of tumor very little necrosis ++ no + 0 0 tissue at surface +++ ++- in areas + yes 0 + 0 of hemor- rhage +++ in areas +- no 0 + + of hemor- rhage +++ in eroded ++ no + + 0 areas *atypical 45 TABLB'2--continued Size & Size & Types Hyper0 Path. Shape of Shape of Color of Mitoses/ of M10 chroma0 No. Cell Nuclei (g). Nucleoligfip) hpf toses tism H2796 spindle oval 2 0 000 + (9x15) single & multiple eosinOphilic H4957 pleomorphic oval 203 l blast + (9x14) to single & round (12) multiple eosinOphilic H5628 spindle elongate l 304 bizarre +++ (3x12) to single oval basOphilic (6x18) H5657 pleomorphic oval 2 3 normal & ++ (10x16) single & blast multiple eosinophilic H6057 spindle pleomorphic 102 405 normal & + to round single blast (10) eosinophilic H6119 spindle elongate l 0 000 ++ H6360 to oval basOphilic (8110) 31766 spindle oval 103 5 nonmal & ++ (6x14) to basophilic & bizarre round (15) eosinOphilic J26ll spindle round (6) 102 0 000 «+++ to oval single & (5°83 multiple ' basophilic J2652 spindle oval 1 no normal 5. ++ (7x12) basOphilic bizarre J269l pleomorphic pleomorphic 102 102 normal ++ basOphilic blast bizarre 46 Invasion of Degree of ‘ Pre0existing “15113212122...— AmO‘mt of Blood 01' mMiscellsneaus gullular Vasculggg NecrosLs, Stroma Lymph Vessels lpfl. Hem, Macro; +++ in areas + yes + + 0 of hemor0 rhage 1444- in areas +++ no 0 +‘ + of hemor- * rhage +++ ++ in center ++ no + + + of cellu0 lar areas +++ none + no + + 0 ++4- 44' in areas + no 0 + 0 of hemor0 rhage 40- -PF+ in center ++ yes «+ -+ + of cellu0 * lar areas +++ +++ in center ++ yes +' + + of cellu0 lar areas “‘1' none “i" also 0 0 + * +‘H" in areas 4+ yes 0 + + of tumor +++ +++ in areas +' yes + + + of hemor- rhage 47 TABLE 20-continued Size & Size & Types Hyper0 Path. Shape of Shape of Color of Mitoses/ of M10 chroma0 No. Cell Nucleigjp) Nucleoli (p) hpf toses tism J28l4 pleomorphic oval 103 102 000 ++ (10x14) single eosinOphilic J2958 spindle round 102 203 normal -+++ (9010) to multiple oval eosinophilic (8x11) J3526 spindle to oval 104 405 blast & ++ pleomorphic (10x14) to ‘multiple bizarre round (13) basoPhilic & eosinophilic J4075 spindle pleomorphic most not seen 0 000 +++ stain. (10) m“ *Fine meshlike network of reticulin as determined with a special ‘+ 0 scanty ++--‘moderate +++.- marked Infl. 0 inflammation Hem. - hemorrhage ‘Macro. 0 pigment01aden.macrophages dis. 0 diameter 48 Invasion of “rlvw—Ihw‘ ~ n—“Jn .u‘ l~ fl-) Degree of Pre-existing Ana lasia Amount of Blood or Miscellaneousflm 93M Vascular Necggsis Stroma Lymph Vessels Infl. H%°.“M.§££9n +++ -+++ in areas ++ no 0 -+ 0 of tumor 44- +++ in areas +- yes 0 0 0 of hemor- rhage ++4- in center ++- yes + + + of cellu- lar areas +++ none ++ yes 0 + + Figure 7. Numerous spindle0shaped cells lining vascular spaces of a hemangiosarcoma. Hematoxylin and eosin. x 750. Figure 8. Cellular type of hemangiosarcoma with little vessel formation. Hematoxylin and eosin. x 468. 50 Figure 9. Cellular type of hemangiosarcoma showing vesicular nucleus and large nucleolus (arrow). Hematoxylin and eosin. x 750. Figure 10. large nucleolus in.malignant cell of a hemangiosarcoma (arrow). Hematoxylin and eosin. x 750. 51 1.5 microns diameter) were characteristically basOphilic, while the larger nucleoli (2 to 5 microns diameter) were usually eosinophilic. If nuclei were large and pleomorphic, they almost always contained multiple nucleoli . Tendency of the +cytoplasm to fuse. In every instance the cytoplasm appeared continuous with that of an adjacent cell, so that no distinct cell membrane was observed. This was especially noticeable in the more anaplastic forms. Nmber .of. mitoses hi: An average of 5 to 6 hpf was examined to determine the number of—mitoses. This figure varied from 0 to lO/hpf. There were usually 2 to 3 mitoses per hpf. Type of mitoses. This characteristic differed from tumor to tumor. Most of the mitoses were normal, but abnormal forms were present in some. One type that wasnumerous was the s00called "blast" form. This {was a form in which the chromatin strands appeared to be exploding from the center toward the periphery in an unorganized pattern. Other abnormal types included the s00called "X form", ”T form”, and shrunken constricted form. These latter 3 forms were all'classi’fied as bizarre (Figures 11 and 12). Hyperchfmlatism. There was sme correlation between the degree of hyperchromatism and the degree of anaplasia. In those tumors which characteristically had large vascular spaces filled with blood, there was only slight to moderate hyperchromatism. There was a greater degree of hyperchromatism in the more anaplastic hemangiosarcmnas. 52 Figure 11. Cellular type of hemangiosarcma cone taining large abnormal mitotic figure (arrow). Renae toxylin and eosin. x 750. Figure 12. Cellular type of hemangiosarcoma con9 taining mitotic figure (arrow). Hematoxylin and eosin. x 7500 53 Degree of anaplasia. The term anaplasia in this paper refers to the degree of differentiation. The well differentiated tumors were classic fied as vascular and less differentiated as cellular. It should be noted that even in those that werewclassified as cellular, a certain amount of vascularitwaas seen; otherwise, the diagnosis of hemangio= sarcoma would not have been made. The tenm "cellular" meant, in most cases, whorl or clump formation of tumor cells or lack of organization (Figures 8, ll, and 12), as compared to those which had well defined vascular channels (Figures 13, 14, 15, and 16). The classification of vascular was used, not only in those areas which contained blood, but also in those areas which were forming nonfunctional neocapillaries. Necrosis. Little necrosis was observed in most of the neoplastic tissues; however, liquefaction necrosis was constant in old areas of hemorrhage. In a few of the more anaplastic forms there was cessation necrosis deep in the cell structure of the tumor. ggount of stroma. A variable amount of stroma was observed. There was little stroma in the more anaplastic forms, while more stroma was present in those hemangiosarcomas that were well differentiated (Figure 15). The stroma in these tumors consisted of wide bands of collagenous fibers. Few normal blood vessels were found in the tumors. Fine meshlike networks were seen in most of the reticulumostained sections. This was present even in the most anaplastic tumors (Figo urea 17, 18, and 19). Invasion_g§flpreeexisting;blood vessels and lymphatics. Among the 59 hemangiosarcomas that had pre-existing blood vessels and/or lymphatics, 54 Vascular type of hemangiosarcoma. Figure 13. Hematoxylin and eosin. x 75. Higher magnification of Figure 13. Hematoxylin and eosin. x 187. Figure 1'. 55 vascular type of hemangiosarcoma Hema- Pisu. 15 s with vein-like channels and abundant stroma. toxylin and eosin. x 187.5. Hematoxylin and eosin. Vascular type of hemangiosarcoma Figure 16. with capillary-like channels. x 187.5. Figure 17. Cellular type of hemangiosarcoma with reticular fibers. Reticulum: stain. x 187.5 Figure 18 . Cellular type of hemangiosarcoma with fine reticular fibers. Note large nucleolus (arrow). Reticulum stain. x 750. 57 a" . ‘I v” f .g - -_; {:5 ' Figure 19. vascular type of hemangiosarcoma with thick bands of reticular fibers. Reticulum stain. x 750. 58 malignant cells had invaded these structures in 29. In several inn stances tumor emboli were found within veins (Figure 20) or arteries (Figure 21). .This was particularly true in tissues which represented a metastasis from a primary tumor. gaggellgneous. Characteristics included under this heading‘were inflammation, hemorrhage, and pigmentaladen macrophages. There was hemorrhage in a number of the tissues which varied from.small to large amounts; in fact, many of the tissues superficially resembled . large blood clots. Infiltration with inflammatory cells was seen in numerous sections. The most common inflammatory cell was the neutro= phil, followed by lymphocytes and plasma cells. In those tumors which had an eroded surface, there was more likely to be inflammation of the purulent type than in those tumors which were deeper. It was not unusual to find many'macrOphages containing an ironabearing pigment resembling hemosiderin in hemangiosarcomas which contained a great deal of hemorrhage. Radio;razhisr¢haraetsristics While only 12 dogs in this study had radiographs taken during the course of examination, some definite characteristics were observed. Metastasis to the lung produced small miliaryatype lesions throughout. All lobes of the lung were equally affected, and the spread seemed to be hematogenous in origin. This gave the lung a scacalled "snowa flake" appearance (Figure 22). Metastases were usually more radioo opaque than the original tissue. Osteolysis was observed (Figure 23) in those dogs in which the shamangiosarcoma occurred in or near bone. This varied from a soocalled 59 Figure 20. Renangiosarcona embolus in vein. Hematoxylin and eosin. x 187.5. Figure 21. Hemangiosarcona embolus in artery. Hematoxylin and eosin. x 187.5. 60 .r l . 9 . u . ‘A‘O: ov‘ ':ngfrt.~"5!m9“l' ~ ‘ u l H96 ‘4 3 "I Figure 22. Lateral radiograph of chest to show "snowflake" appearance of metastatic hemangiosarcoma in lungs. Figure 23. Ventro=d9rsal radiograph of pelvic area. Note "punched out" appearance of left femoral head caused by hemangiosarcoma. 61 "punched out" appearance to a more diffuse reaction with bone proliferam tion at the edges of the lesion. This diffuse reaction resembled osteomyelitis radiographically. If the tumor occurred in the soft tissue near bone, it presented a homogeneous, somewhat radiolucent appearance, similar to other neoplasms of soft tissue; it was characteristically circumscribed and blended with the adjacent tissue. DISCUSSION Hemangiosarcomas are comparatively rare neoplasms. While 76 neOplasms over a 9-year period seemed like a great many, it is a relatively small number compared with the total number of tumors. During thisftime approximately 16,000 tumors were reviewed by the Department of Pathology. While no exact count was taken, it is reasonable to say that of these at least l4,000*were in the dog. Feldman (1932, p. 60) stated, "statistical data pertaining to the occurrence of tumors of lower animals are meager and fragmentary". While a great deal of improvement has taken place since that time, very little information of this type“was found with regard to hemane giosarcoma in the dog. For this reason, the following discussion is undertaken. Breed incidence. The number of hemangiosarcomas by breeds to the number of expected cases is compared (TABLE 3), just as Tjalma (1965) did in his study of canine bone cancer. Mulligan (1959) reported a predisposition for hemangiosarcoma in the German Shepherd Dog. In this study, the highest rate of incidence occurred in the Boxer, which had approximately 9 times the expected rate. Also of interest was the large number which occurred in the Cocker Spaniel. The distribution of the tumor among the various weight classes of the canine population and the expected incidence in each of these 62 63 TABLE 3. Comparison of observed and expected cases of hemangiosarcoma among specific breeds and weight classes. 7: TotSfEBE-Mm‘ ' naggiagm“ Range of posite Popu- Hemangiosarcoma Deficiency Occurrence* lation Same __ Cases of Observed to ggeed J1 421st;Mr 9128 erved Em acted Exaceteéfiassa Boxer 2.12 - 3.43 '2.40 14 1.61 Excess Cocker Spaniel 7.21 - 13.79 9.43 14 6.32 Excess G. Shep. 0.99 - 7.15 4.21 6 2.82 Excess Lab. Ret. 0.26 - 2.62 1.12 4 0.75 Excess Collie 2.76 - 9.63 5.60 0 3.75 Deficiency Beagle 6.80 - 13.72 9.89 2 6.63 Deficiency Dachshund. 1.76 - . 9.13 4.10 3 2.74 Excess F. Terr. 0.57 - 3.55 2.05 l 1.37 Deficiency Other --- 61.20 23 41.00 Deficiency TOTAL 100.00 67 66.99 Weight Class Giant 80# 0.00 - 1.61 0.40 0 0.28 Deficiency Large 40-80# 18.85 - 38.58 28.03 41 19.90 Excess Medium 20-40# 23.78 - 37.72 30.77 17 21.84 Deficiency Small 20# 11.30 - 30.40 16.98 10 12.05 Deficiency Unknown 15.66 - 31.17 23.82 3 16.92 Deficiency TOTAL 100.00 71 70.99 *Among field population samples. **Base population of 123,123 dogs. 64 populations is also compared (TABLE 3). As can readily be seen, a definite excess over expected cases was found in the large breeds, while no definite statement can be made about other papulations. This information compared favorably with Tjahma's work with canine bone cancer, in which he also found an increased incidence in the large breeds. This poses the question as to whether other types of cancer in the dog occur at a higher frequency in large breeds. Sex distribution. In Tjalma's (1965) study of the normal canine population, which was based on a population of 99,088, he found the sex distribution_to be 521'ma1es and 48% females. In a 30bmonth study of dogs admitted to the small andmal clinic at Michigan State University from July 1960 to January 1963, he found the distribution of cancer to be 226 males and 210 females, with 1 unknown. This is a ratio of 52% in the male and 481 in the female, which corresponded to the universal population. Therefore, a ratio of 1.76 males to 1 female, as presented by this study, would tend to support the findings of Mulligan (1961) and Meier (1963), in which each described a higher incidence of hemangiosarcomas in the male. It should be pointed out that the influencing factor on the sex difference was the great preponderance of hemangiosarcomas in males of the large (40 to 80 pounds) breeds (Figure 1). Excluding this group, the ratio of males to females was approximately equal. Age distrégution. According to Hulligan (1959), the beginning of "cancer age" was 6 years, with most cases of cancer in the dog occur- ring between 7 and 14 years. Cotchin (1959) substantiated this figure by pointing out that 56.2% of the tumorobearing dogs in his study 65 were between the ages of 6 and 10 years. Meier (1963) found hemangio= sarcomas usually occurred in dogs past 6 years of age. In this study, the mean age for the tumor was 9 years, a little earlier for males and a little later for females, and therefore would fall into the category of "cancer age" as described above. Primagy location. The most common primary sites for hemangiosarcoma were the subcutis and the spleen. These locations were substantiated bothin man and in the dog (Lombard, 1935; Cotchin, 1954; Herbut, 1955; Boyd, 1958; Moulton, 1961; Mulligan, 1961; and Robbins, 1962). T'Cotchin (1959), in his study of 4,187 tumors of dogs, found the skin to be the most commonly affected area (37.52). Pack and Ariel (1958) observed that the tumor most often occurred in the extremities in man. In this study in the dog, the tumor*was ‘most commonly found in the trunk. metastasis. Mulligan (1961) noted metastasis in 8 of 20 hemangio° sarcomas in dogs, with 7 of these 8 to the lungs.‘ This finding was substantiated by our study, where the most common metastatic site was the lung. Other organs affected by metastasis would suggest that the usual route of metastasis by this tumor is hematogenous. Metastasis by the lymphatic route was apparently minimal, since only 5 of 32 tumors had metastasized to the regional lymph nodes. In many instances, as shown (Figures 3 and 4), metastasis apparently occurred in a relatively short period of time, because they all were of uniform size. This was especially true in the lungs, which appeared to be completely infiltrated with the same-sized nodules. 66 gage of development. While it appeared from the data that the tumor occurred suddenly, it should be remembered that this conclusion was usually based on the owners' observations. Since cancer is such a subtle disease, many are not recognized by the owner until they are in the late stages. Because of this no accurate statement can be made as to the rate of development. Recurrence. No definite statement can be made as to the frequency of recurrence, since no follow-up data were available on most of the biopsy specimens submitted to the Department of Pathology for diaga nosis. Information concerning this point was available in only 11 cases. Of these, 7 hemangiosarcomas were found to recur. Tentative diagnosis. As pointed out in the results, a tentative diago nosis of hemangiosarcoma was made by the clinician in only 6 of 21 tumors. In surveying other diagnoses made by the clinicians, nothing specific was found as being characteristic; rather, a wide variety of diagnoses was made. The difficulty in clinical diagnosis is due to wide variability of this tumor both grossly and microscopically. No negative information was obtained in this matter, but it can be postulated that there were probably mmny tumors clinically diagnosed as hemangiosarcoma which later proved to be some other lesion on histo= pathologic examination. Just as many of these tumors were not as vascular as one would expect them to be, there were also highly vascular tumors which were not hemangiosarcomas. In bone, the tumor was often clinically diagnosed as an osteolytic osteosarcoma by the clinician. 67 If the tumor had metastasized to the lung, and if radiographs were taken of this area, a correct diagnosis was often made by the radiologist. gige of tumor. This study supported the conclusions of Bell (1956) and Boyd (1958), who pointed out that these tumors are often round, variable in size, and hemorrhagic in appearance. The metastatic lesions were often less hemorrhagic than the primary lesion. Free quently, blood clots were associated with the tumors where they evidently had ruptured (Figure 6). This complication often was given as the cause of death because, while the neoplastic process had not advanced to the point of involvement throughout the body, one of the tumors had ruptured and resulted in exsanguination. This was a common finding substantiated by Moulton (1961), who pointed out that "hemorrhage and necrosis are almost constant features", and by Mulligan (1949), who stated that "rupture of metastases has resulted in sudden fatal hemorrhage". gerasaeniefiheracteriasaigg As Watson and McCarthy (1940) pointed out, "An angioma is a true neoplastic process involving vascular or lymphatic tissue. In this respect, an angioma differs distinctly from simple selfolimiting vessel hypertrophy such as occurs in granulation tissue, and it bears no relation to the ordinary dilatation of previously formed vessels such as occurs in varices." This seemed to be the key point in the microscopic characteristics of this tumor. Some areas of the tumors had benign or even normal appearing endothelial structure, but the characteristics of other areas established the true malignancy of this tumor CTABLE 2). It 68 is haped that the following discussion will point out the charaC° teristics which will help establish the differences between malignant and benign endothelial tumors. §hgpgupf the cell. The general outline of the cell, as described in Results, was spindle shaped. Pack and Ariel (1962) pointed out that, in the more anaplastic tumors the endothelial cells may grow in.whorls or undifferentiated sheets. This feature was also found in the study, in that the shape of the cell was less elongated in the more anaplastic forms of the tumor. This coincided with the overoall failure to form.vascular channels in these same anaplastic tumors. Sizeéggd_shape of nuclei. The general shape of the nucleus was oval. The more anaplastic the hemangiosarcoma, the greater the tendency for the nucleus to round up or to attain pleomorphic characteristics. This characteristic was shmilar to that of other tumors of mesenchymal origin, in that the more anaplastic they are, the less they resemble the tissue or cell of origin. Also, the more nearly round nuclei had a tendency to be vesicular. Size and staining prgperties of nucleoli. There was no reference to this point in the literature, but in our study the larger the nucleo= 11, the greater the tendency for them to be eosinophilic. It may be because of the optics involved that small nucleoli (0.5 to 1.5 microns) were darker and appeared basOphilic. .Therefore, the significance of’ this finding might be debatable. Large and multiple nucleoli were considered as factors in establishing a diagnosis of malignancy. 69 Number of mitoses. While this figure was variable, usually 2 to 3 mitoses were found per high power field. A greater number was usually observed in the more anaplastic tumors. This was considered as moderate in number compared to other forms of cancer. It must be remembered that this subjective observation would vary from investigator to investigator. However, this tended to support the early work by Feldman (1932) and the more recent work by Mulligan (1961) and Moulton (1961), who noted a moderate number of mitoses in hemangioo sarcomas 0 Types of mitoses. The appearance of bizarre mitotic figures may depend upon the angle the cell was cut. This isznost likely true during prophase. However, when a tripolar or 9“blast" mitotic figure was observed it was described as abnormal. Mulligan (1961) was the only author to state that abnormal mitotic figures occurred in hemangio= sarcomas in the dog. Robbins (1962) found most of the mitotic figures to be of normal type. Hyperchromggigg, There was no definite relationship between hyper: chromatism and the degree of malignancy. In discussinghyperchroma° tism, various factors must be considered, such as the thickness of the section, the amount of hmmatoxylin, and even the type of filter in the illuminating system. The nucleus of the normal endothelial cell is dark in appearance due to the small size of the nucleus and condeno sation of the chromatin. In fastogrowing granulation tissue, the nucleus stains somewhat lighter. A greater degree of hyperchromatism was usually seen in the more cellular areas of the hemangiosarcomas. 70 However, there seemed to be a point of diminishing returns. In those tumors which were highly anaplastic, with vesicular type nuclei, the reverse was found to be the rule. Degree of anaplasia. The highly vascular hemangiosarcomas were con0 sidered well differentiated and thus less anaplastic, whereas the highly cellular tumors (less vascular) were thought to be more anao plastic. The hemangiosarcomas varied considerably in these respects. The vascular structure in many of the tumors appeared quite benign, even though they had metastasized.‘ No effort was made to categorize the tumors as t0‘whether they were forming veins or capillaries, since there could be areas of capillary formation and other areas in the same section where the formation tended to be more veinalike. It was considered that if a tumor were cellular in appearance this represented a more anaplastic form of the tumor, An interesting experiment that could be undertaken with this tumor would be the use of Gomori°s technique for demonstrating alkaline phosphatase (Pierce, 1960). The rationale for such was best explained by Urbach and Graham.(l962), who stated that alkaline phosphatase. activity appeared to be most intense only in the endarterial or capillary endothelium, while the wall of the venules or larger ar= terioles had little or no enzymatic activity. An experiment such as this might reveal the actual point of origin of the malignant cells. Kapf (1957) and Ellis g£,g$. (1958) described this technique and used it as an indication of the proliferative activity of the blood vessels in a particular area. In our study this technique was not possible, .since it required either frozen sections or tissues fixed in 2 to 3 71 changes of absolute alcohol before the staining technique could be applied. Necrosis. A majority of the tumors investigated was necrotic to some degree, usually in areas of old hemorrhage. This was probably the result of rupture of l of the delicate endothelial structures. In a few, however, a caseous type of necrosis was found deep in the anaplastic areas. The question arose as to the reason for necrosis in a tumor which essentially was amass of blood vessels, where it ‘would seem that the cells were undoubtedly well oxygenated. An explana= tion for this was found in the work of Urbach and @raham (1962), who stated, "The presence of large numbers of capillaries in tumor tissue does not necessarily hmply an abundant supply of nutrient materials as indicated by tissue tension.measurements, which suggests that the rich vascular bed is functionally inefficient." Proof of this was obtained by Urbach and Noell (1958), in a study in which the oxygen tension was measured in normal tissue and tumor tissue by use of a polarographic current. There was a local state of hypoxia in tumor tissues, and it was suggested that this was the . primary reason for necrosis deep in tumor tissue. Again, an alkaline phosphatase stain would be valuable because the presence of this enzyme is a direct function of the metabolic activity of the endo» thelial cells. Amount of stroma. There was variation in the amount of stroma. A large amount of stroma'was observed in highly vascular hemangio= sarcomas, while the more cellular tumors had less stroma. In some of the more anaplastic forms, the strome was particularly scanty. The 72 reticulum stain revealed the interesting fact that the endothelial cells, even in the most anaplastic forms of the tumor, tended to lay down a fine delicate reticular meshwork. This special stain was applied to a vascular fibrosarcoma and leiomyosarcoma. In these tumors, the reticulin was associated only with endothelial cells and not with the tumor cells. For this reason, and also for the reason that the hemangiosarcomas appeared to have a characteristic meshlike network, it is suggested that this stain is of value in the differentiation of hemangiosarcomas from vascular tumors of other types. Stout (1943) also emphasized the value of this stain in diagnosing tumors of vascular origin. Ipyagipn of_pre-existipg_blgod vessel§_and lymphatigg, In several ‘ instances tumor emboli were found in veins and arteries, while few were seen in lymphatics, suggesting that the former played a greater role in the metastasis from the primary site to secondary areas of the body. This point was suggested in an earlier section of this discussion. Occasionally, these tumor emboli retained their vascular form.(Figure 20). It was not surprising that these tumor emboli were found in the blood vessels, since they originate from the vascular endothelium. aggcellggpous. As pointed out in the previous areas of this discuso sion, hemorrhage seemed to be an almost uniform finding. Mulligan (1961) emphasized the fact that these tumor cells are delicate and have a tendency to break, resulting in small to vast areas of hemore rhage. The large number of macrophages laden with a hemosiderinotype pigment would support the observation that hemorrhage had been present 73 for long periods of time. The significance of inflammation in these tumors is questionable. Although many of the tumors had large numbers of inflammatory cells, this could be explained by the tissue breako down resulting in a leukotaxic effect. Another explanation would be that in many of these tumors which had eroded surfaces, bacterial infection could easily have produced the influx of inflammatory cells. Radiographic Characteristics The fact that only 12 dogs with hemangiosarcoma had radiographs limited the conclusions that could be made. Two significant observao tions were made. First was the characteristic small miliaryetype ("snowflake") distribution of metastases to the lung (Figure 22). The distribution of these can be explained by the hematogenous spread of the metastases. According to Mbstosky (1965), "this characteristic distribution of metastatic lesions is pathognomonic of the hemangiosarcoma". This finding could be of great value‘to' the small animal clinician who uses radiographs as part of his diago nostic procedure. - Second, osteolysis occurred adjacent to the tumor (Figure 23). This supports the findings of numerous authors (De Lorflmier g£,g;,, 1954; Ritvo, 1955, and Bradley and Coley, 1960), who described hemangio= sarcomas in bone as resembling osteomyelitis or osteosarcomas in ' appearance. Again, Mbstosky pointed out that the radiographic appearance of tumors in bone was fairly characteristic and, along ‘with the diagnoses mentioned above, hemangiosarcoma must he considered when viewing a radiograph which has the appearance shown (Figure 23). SUMMARY This thesis is a retrospective study of 76 canine hemangiosaro comas diagnosed in the Department of Pathology from 1956 to 1965. The study was divided into 3 basic parts: (I) occurrence and gross characteristics, (2) microscopic characteristics, and (3) radian graphic characteristics of the tumor. Hemangiosarcomas occurred with greater incidence in the large breeds of dogs, particularly the Boxer. There was a sex ratio of approximately 2 males to 1 female. The mean age was 9 years, a little earlier in males (8.5 years) and a little later in females (10.7 years). The most common primary sites were the subcutis and the spleen. If metastasis occurred the most common site was the lung. The tumor was difficult to diagnose without histologic examination. Hemorrhage was an ahmost constant finding. The neoplastic endothelial cell was generally spindle shaped, resembling a short fibroblast. The nucleus was generally oval and of moderate size (9 x 12 microns), often containing large eosino= philic single and multiple nucleoli. There was a tendency for the cytOplasm of 1 cell to fuse with the next, presenting a syncytial type of arrangement rather than individualized cells. There were both normal and abnormal mitoses. MOst of the nuclei had a moderate amount of hyperchromatism. The cellular arrangement varied from normal_appearing vascular cells to undifferentiated cellular sheets arranged in whorls or clumps with little vessel formation. The use 74 75 of a reticulum stain revealed a fine reticular meshlike network, even in the most anaplastic areas of the tumor. There was a variable amount of necrosis, especially associated with hemorrhages. The amount of stroma varied from large amounts in the vascular forms to scanty in the more anaplastic cellular forms. Invasion of blood vessels occurred in 29 of 59 hemangiosarcomas. There were inflammae tory cells in many of the tumors, particularly those which had an eroded surface. Two radiographic characteristics were observed. First, the metastatic lesions in the lungs presented a fisnowflake" pattern. Secondly, osteolysis occurred in those’neoplasms that were in or near bone. LIST OF REFERENCES American Cancer Society. Manual of Tumor Nomenclature and Coding, corrected ed. 1953. Subcommittee of the Statistics Committee, Am. Cancer Soc., New York: 53°54. American Cancer Society. Systematized Nomenclature of Pathology, lst ed. 1965. Coll. of Amer. Path., Chicago, 111.: 121. Anderson, W. A. D. Pathology, 4th ed. 1961. C. V. Moshy C0,, St. Louis, Mo.: 576. Andervont, H. B., Grady, H. 0., and Edwards, J. E. 1942. Induction of hepatic lesions, hepatomas, pulmonary tumors, and hemangio= endotheliomas in mice with Obaminoeazotoluene. J. Nat. Cancer Inst., 3: 131~l53. Arey, L. B. 1954. Deve10pmenta1 Anatomy, 6th ed. W. B. Saunders Co., Philadelphia, Pa.:' 3409349. Armed Forces Institute of Pathology. Manual of Histologic and Special Staining Technics, 2nd ed. 1960. McCrawaHill Book 00., Inc., New York: 30, 151. Bell, E. T. A Textbodk of Pathology, 8th ed. 1956. Lea and Febiger, Philadelphia, Pa.: 363. Boyd, W. A Textbook of Pathology, 6th ed. 1958. Lee and Fehiger, Philadelphia, Pa.: 259°261. Bradley, L., and Coley, H; D. Neoplasms of Bone, 2nd ed. 1960. Paul B. Hoeber, Inc., Medical Div. Harper and Brothers, New York: 341-363. Brodey, R. S. 1960. A clinical and pathologic study of 130 neOplasms of the mouth and pharynx in the dog. Am. J. Vet. Res., 21: _ 787-812. Carlson, W. D. Veterinary Radiology. 1961. Lea and Febiger, Philaa delphia, Pa.: 2349239. - Cotchin, E. 1954. Further observations on neoplasms in dogs, with particular reference to site of origin and malignancy. 1. Cutaneous, female genital and alimentary systems. Brit. Vet. J., 110: 218-230. - 76 77 Cotchin, E. 1954. Further observations on neOplasms in dogs, with particular reference to site of origin and malignancy. II. Male genital, skeletal, lymphatic and other systems. Brit. Vet. J., 110: 2749286. Cotchin, E. 1959. Some tumors of dogs and cats of comparative veteri9 nary and human interest. Vet. Rec., 71: 104091050. De Lorimier, A. A., Mbehring, H. G., and Hannan, J. R. 1954. DeveIOpa mental and systemic conditions and local lesions in the extremi9 ties, in Clinical Roentgenology, Vol. I. Charles C. Thomas, Springfield, 111.: 3359337. De Navasquez, S. 1936. Angioblastoma of the spleen with metastasis in the liver. J. Path. Bact., 42: 6519656. Ellis, R. A., MOntagna, W., and Fanger, H. 1958. Histology and cyt09 chemistry of human skin. XIV. The blood supply of the cutaneous glands. J. Invest. Dermat., 30: 1379146. Edwards, J. E., Andervont, H. B., and Dalton, A. J. 1942. A transa plantable malignant hemangio-endothelioma of the liver in the mouse. J. Nat. Cancer Inst., 2: 4799490. Farrell, J. M., Jr., and Farrell, D. 1960. Hemangiosarcoma - A case history. Southwest Vet., 14: 61. Peldman, W. H. 1932. Neoplasms of Domesticated Animals. W. B. Saunders Co., Philadelphia, Pa.: 60, 1789188. Freilich, E. B., and Gas, C. C. 1936. Angiosarcoma: Case report and review of the literature. Am. J. Cancer, 26: 2699275. Furth, J., and Furth, 0. B. 1938. ‘Monocytic leukemia and other ne09 plastic diseases occurring in mice following intrasplenic injection of 1:29benzpyrene. A. J. Cancer, 34: 1699183. Geshickter, C. F., and Keasbey, L. A. 1935. Tumors of blood vessels. Am. J. Cancer, 23: 5689591. Golden, R. 1959. Diagnostic Radiology, Vol. 1. Williams and Wilkins Co., Baltimore, Md.: l90L9l90N. Ham, A. W. 1957. The circulatory system, in Histology, 3rd ed. J. P. Lippincott Co., Philadelphia, Pa.: 4809507. Herbut, P. A. 1955. Pathology. Lea and Febiger, Philadelphia, Pa.: 2569258. Jgplid, Burtil. 1961: Haemangioendotheliomas in poultry. J. Comp. Path., 71: 370-376. 78 Joest, E. 1924. Spezielle Pathologische Anatomie Der Haustiere, Vol. III. Richard Schoetz, Berlin: 4989509. Kapf, A. W. 1957. Alkaline phosphatase in skin. Arch. Dermatol., 75: 1-29. . ‘ Kingman, H. E., and Newsom, I. E. 1918. An interesting tumor. Case report. J. Am. Vet. med. Assoc., 5: 7039705. Kinkade, J. M. 1949. Angiosarcoma: a review of the literature. Ann. Otol. Rhin. & Laryng., 58: 1599167. . Langhans, T. 1879. Arch. Path. Anat., 75: 273. Cited by De Navasquez. Lesser, E. C., and Von Schowingen, R. S. 1955. Arteriography in bone tumors. N. Y. State J. Med., 55: 342593430. lasserre, R., lombard, Ch., and Labatut, R. 1938. Recherches sur le cancer des anflmaux domestiques. Rev. Med. Vet., 15, 8: 425-4510 ’ Leisering. 1861. Gefabgeschwulst b. e. hunde. Ber. Vet.9Wes. Sachsen 6. Cited by Joest. Lichtenstein, L. 1952. Bone Tumors. c. v. Mosby Co., St. Louis, M0,: 1259l28. Lieberman, L. L. 1955. Malignant hemangi09endothelioma of the canine heart. J. Am. Vet. Med. Assoc., 126: 296. Lienaux, E. 1899. Deux cas d"endotheliomes de la region parotidienne chez 1e chien. Ann. de Med. Vet., 48: 502. Cited by Feldman. Lillie, R. D. 1954. Histopathologic Technic and Practical Histo- chemistry. Blackiston Co., New York. Lindsay, 8., and GiLmore, J. W. 1946. Prflmary splenic endothelial sarcoma in a dog. J. Am. Vet. Med. Assoc., 109: 1949197. Lombard, C. 1964. Vegetating sublingual haemangiO9endothelioma in cats. Bull. Acad. Vet., Fr., 37: 1639167. Magnusson, R. 1934. Sarcoma of the mmall intestine in connection with a case of hemangiosarcoma of the jejunum. Acta. Chir. Scandinav., 73: 5769590. , Mallory, F. B. 1914. The Principles of Pathologic Histology. W. B. Saunders Co., Philadelphia, Pa.: 3099326. Meier, H. 1963. Epizootiology of cancer in animals. Ann. N. Y. 79 Mostosky, U. V. 1965. Small Animal Clinic, College of Veterinary Medicine, Michigan State University, East Lansing, Mich. (personal communication). Moulton, J. E. 1961. Tumors of Domestic Animals. Univ. of Calif. Press, Berkeley and Los Angeles, Calif.: 77979. Mulligan, R. M. 1949. Neoplasms of the Dog. The Williams and Wilkins Co., Baltimore, Md.: 73976. Mhlligan, R. M. 1961. Mesehcymal and neurilemmal tumors in the dog. Arch. Path., 71: 5129531. Pack, G. T., and Ariel, I. M. 1958. Tumors of the Soft Somatic Tissues. Paul B. Hoeber, Inc., Harper & Brothers, New York: 4429450.. Palade, G. E., 1953. Fine structure of blood capillaries. J. Appl. PhYSiCS, 243 14240 Pearse, H. G. E. 1960. Histochemistry 9 Theoretical and Applied, 2nd ed. J. & A. Churchill, Ltd., London: 8689869. Pounden,‘W. D., and Sprunger, E. 1947. Malignant hemangioendothelioma of a canine spleen. N. Am. Vet., 28: 461. Public Health Service, U. S. Department of Health, Education and Welfare. 1964. A Standard of Veterinary Diseases and Opera9 tions, lst ed. Bethesda, Md.: 110. Regenbagen. 1907. Uber das vorkommen der hgmangiome. Monatschr. f. prakt. Thierh., 18: 2939298. Cited by Feldman. Rigdon, R. H. 1952. Tumors produced by methylcholanthrene in the duck. Arch. Path., 54: 3689377. Ritvo, M. 1955. Bone and Joint X9ray Diagnosis. Lea and Febiger, Philadelphia, Pa.: 4719472. Robbins, S. L. 1962. Textbook of Pathology, 2nd ed. W. B. Saunders Co., Philadelphia, Pa.: 4719472. ' Robinson, J. M., and Castleman, B. 1936. Benign metastasizing hemangioma. Ann. Surg., 104: 453. Shennan, T. 1914. Histologically nonmalignant angioma, with numerous metastases. J. Path. Bact., 19: 1399154. Siedamgrotzky. 1874. Angiom e. hundes. Ber. Vet.9Wes. Sachsen 19. Cited by Joest. Simpson, L. 1960. Hemangioendotheliomatosis in a dog. Golden Jubilee Commemorative Vol., Coll. of Vet. Med., Univ. Phillipp., Dilflman, Rizal, Phillippines: 82993. 80 Steiner, P. E. 1942. Comparative pathology of induced tumors of the salivary glands. Arch. Path., 34: 6139624. Stout, A. P. 1943. Hemangio-endothelioma: a tumor of blood vessels featuring vascular endothelial cells. Ann. Surg., 118: 445. Thiele. 1904. Arch. f. path. Anat., 178: 296. Cited by De Navasquez. Tjalma, R. A. 1965. National Cancer Institute, National Institutes of Health, U. S. Public Health Service, Bethesda, Md. (per9 sonal communication). Urbach, F., and Graham, J. F. 1962. Anatomy of human skin tumour 5 capillaries. Nature, Lond., 194: 6529654. Urbach, E., and Noell, W. K. 1958. Effects of oxygen breathing on tumor oxygen measured polarographically. J. Appl. Physiol., 13: 61965. Watson, W. L., and McCarthy, W. D. 1940. Blood and lymph vessel tumors; report of 1,056 cases. Surg. Cynec. & 0bstet., 71: 5699588. Weipers, W. L., and Jarrett, W. F. H. 1954. Haemangioms of the scrotum of dogs. Vet. Rec., 66: 1069107. White, J., and Stewart, H. L. 1942. Intestinal adenocarcinoma and intra-abdominal hemangiO9endothelioma in mice ingesting methylcholanthrene. J. Nat. Cancer Inst., 3: 3319347. Willis, R. A. 1948. Pathology of Tumors. Butterworth and Co., Ltd., London, C. V. Mosby Co., St. Louis, Mb.: 713. M'TITI'I‘RLWILEMIMTHMJIIEIITLHI'IMTLIMTIH