Document rx3wqYpep1YbBbnK6wpKx46d7

R&S 109728 BIOMEDICAL RESEARCH DOCUMENT DESCRIPTION FORM Duplicate in all cards:-- 63 / f . J 68 69 .76 00(13354 -j year as-1961- File number [Right justify [Numeric only] Author (s), as Last Name FS (No Punctuation) and coden for journal as JAMA preceeded-by. one blank space 20 "21 $ 'tkl s l ft j 7~e-//& s AJ(L ------ ,= 40 41 `fZff_________ 77 78 - Sub-Index Code 60 61 62 11 12 13 Title of Report: end with space-hyphen-hyphenyspace. Follow with Index Terms. separated from each other with -comma-space. Avoid other punctuation; do'not abbreviate. 12 4 61 62 21 L-'-- c . ^ \ * ifc : Of j 7 ... p ,'j 22 23 24 ource (Journal, Vol., Number, Pages, Date) 12 ............... ... r /?7j- ZTC /Atz 61 62. 31 32 Erief Summary 61 62 V -'M R&S 109729 "303254 Development of Hepatic Angiosarcoma in Man Induced by Vinyl Chloride, Thorotrast, and Arsenic Comparison With Cases of Unknown Etiology Hans Popper, MD, PhD, Louis B. Thomas, MD, Norman C. Telles, MD, Henry Falk, MD, and Irving J. Selikoff, MD Examples of human angiosarcoma following exposure to vinyl chloride, Thorotrast, or arsenic (medicinal and industrial) and cases, including children, of unknown etiology were studied to establish diagnostic criteria and to study their evolution. The uniform evolution suggests an environmental factor also in the cases of unknown etiology, which may be established by epidemiologic studies. A precursor stage is characterized by areas of combined hyperplasia of hepatocytes and a variety of sinusoidal and perisinusoidal cells associated with excess of reticulin and with sinusoidal dilatation. The diagnostically useful picture in silver impregnations indicates reticulum formation by the perisinusoidal cells, presumably the lipocytes. The hepatocytic proliferation suggests a hepatocarcinogenic but usually not fully expressed potential. The mixed hyperplasia of the various sinusoidal cells proceeds to an overgrowth of angiosarcoma cells, presumably derived from endothelial cells. In early stages they are usually in contact with hepatocytes (intralobular growth). A trabecular arrangement results from loosening of the lobular plate arrangement by dilatation of sinusoids, leading to primary peliosis. With disappearance of the hepatocytes, various growth patterns develop, terminating in nodular, solid angiosarcoma composed of either spindle-shaped or polyhedral cells which undergo necrosis or hemorrhage (secondary peliosis). The interaction between hepatocytes and sinusoidal cells requires elucidation. (Am J Pathol 92:349-376, I97S) The recognition of the association between exposure to gase ous vinyl chloride during its polymerization to the common plastic poly vinyl chloride and the appearance of hepatic angiosarcoma1 has increased the interest in this tumor. This concern was accentuated by its production in rodents exposed experimentally to vinyl chloride before the human lesion was recognized.3 Until then, hepatic angiosarcoma had been con sidered rare in man, although it has been known that exposure to thorium dioxide (Thorotrast)3 and inorganic arsenicals * may be followed by de- From the Stratton Laboratory for Liver Disease, Mount Sinai School of Medicine of the City University of New York, New York, New York; the Laboratory of Pathology, National Cancer institute. NIH. Bethesda, Maryland; the Office of Medical Affairs, Bureau of Radiological Health. FDA, Rockville, Maryland; the Chronic Diseases Division, Bureau of Epidemiology, Center for Disease Control. USPHS. Atlanta. Ceorgia; and the Environmental Sciences Laboratory, Mount Sinai School of Medicine of the City University of New York, New York, New York. Supported in part bv National Institute of Occupational Safety and Health (NIOSH) Contract 210-75-0044. Accepted for publication March 30. 1978. Address reprint requests to Louis B. Thomas, MD, Chief, Laboratory of Pathology, National Cancer Institute, NIH. Building 10, Room 2A27, Bethesda, MD 20014, 0002-9440/78/0810-0349S01.00 349 350 POPPER ET AL American Journal of Pathology velopment of the same tumor. By contrast, in domestic and experimental animals hepatic angiosarcoma is more common. In humans exposed to vinyl chloride, i,# arsenicals,7 and Thorotrast,8 a peculiar hepatic fibrosis often associated with portal hypertension has been observed; transition of this antecedent lesion to angiosarcoma has been postulated. It therefore appeared important to study available cases of angiosarcoma associated with exposure to vinyl chloride, Thorotrast, or inorganic arsenicals and cases without known etiologic factors in an attempt to a) ascertain differ ences in pattern and evolution associated with various etiologic factors; b) describe a focal hyperplastic precursor lesion; c) trace the evolution of the precursor lesion and of the angiosarcoma; d) stress the connection of sinusoidal cell hyperplasia with the proliferation of the hepatocytes, since hepatocellular carcinoma has also been documented to result from ex posure to Thorotrast9 and inorganic arsenicals 4 and has been found in rodents exposed to vinyl chloride 10 while in humans this association with vinyl chloride exposure is only suggested by anecdotal observations;11' e) study the relation of a mixed mesenchymal cell reaction to intralobular fibrosis; and f) provide diagnostic criteria for both the precursor stage and angiosarcoma useful in epidemiologic surveys. There were relatively few reports of hepatic angiosarcoma of unknown etiology 12-18 before the vinyl chloride experience raised interest in the tumor. They are now supplemented by several others,18'18 but even largescale surveys19,10 emphasized the rarity of the hepatic tumor. This tumor has been reported in all age groups, including children, and has no established sex predilection. These lesions have to be differentiated from benign but actively proliferating vascular tumors such as capillary heman gioendothelioma,21 particularly its infantile variety,22 as well as from malignant mesenchymal tumors with conspicuous participation of capil laries such as malignant mesenchymoma,22 hemangiopericytoma,25 and Kaposi's sarcoma, all of which have been found occasionally in the liver. Highly vascularized hepatocellular carcinomas sometimes also present diagnostic problems. The proper designation of the tumor under discussion has been argued in cases of unknown etiology and in those associated with arsenic and Thorotrast exposure; terms such as "hemangioendothelial sarcoma,"22 "Kupffer cell sarcoma,"24 "malignant hemangioendothelioma," and "hemangiosareoma" have been used. In recent years, however, it has become the custom to use "angiosarcoma" as the most descriptive term although the recent nomenclature proposal for hepatic lesions28 lists most of these terms with equal emphasis. For simplicity, the term "angiosarcoma" is used here. Vol. 92. No. 2 August 1978 HUMAN HEPATIC ANGIOSARCOMA 351 Review of Available Data The first association of hepatic angiosarcoma with vinyl chloride was reported in the work force of one of the oldest polymerization plants in the United States1 and additional cases have subsequently been reported and well-described histologically.1*'28 Since then more instances have been reported in the United States * as well as in Canada,2 England,30,31 Germany,11,32'33 and France,3*'38 The fibrotic antecedent lesion, initially recognized in Bonn, Germany,3 was well documented by laparoscopic and scintigraphic observations 33,39`41 and subsequently reported in this coun try * and in England 42 in workers exposed to gaseous vinyl chloride. The German investigators were stimulated by the recognition of splenomegaly in workers with acro-osteolysis following exposure to high concentrations of vinyl chloride.43,44 Characteristic changes in the spleen associated with the antecedent fibrotic lesion were described as resembling the changes initially associated with the Banti syndrome.* These observations con firmed previous reports of nonspecific hepatic lesions in vinyl chloride workers.43,4* Rare instances of cirrhosis were also reported.42 Idiopathic portal hypertension following the long-term exposure to inorganic arsenicals in Fowler's solution used in the treatment of psoriasis has been reported,2,44`30 a treatment which also has been rarely followed by hepatic angiosarcoma.31,52 The largest series of angiosarcomas following exposure to arsenicals in pesticides, which also cause hepatocellular carci noma and cirrhosis, although less frequently, comes from vineyard work ers in the Rhineland.4,33,34 Hepatic angiosarcomas and carcinomas and, apparently, the fibrotic antecedent lesion following Thorotrast exposure have been well-described in Portugal.33,3* The number of angiosarcomas following Thorotrast exposure has increased in recent years,32"*1 and the lesion has been produced experimentally in rabbits.32 Finally, inorganic copper has also been incriminated.33 The initial recognition that vinyl chloride has a carcinogenic potential in rodents34 stimulated the extensive investigations of Maltoni,2 and changes in the livers of workers exposed to vinyl chloride have been compared with those in rodents,*3 Angiosarcoma is more frequent in animals, both domestic and experimental, than in humans. It occurs spontaneously in dogs ** and in some strains of inbred mice,*7 and it has been observed in rodents and primates exposed to carcinogens, particu larly those which also produce hepatocellular carcinoma M"71 and after infection of hamsters by polyomavirus.22 Materials and Methods This study is based on review of material which became available primarily for consulta- R&S 109731 R&S 109732 352 POPPER ET AL American Journal of Pathology tion at the Laboratory of Pathology of the National Cancer Institute and at the Mount Sinai School of Medicine. Most of the material was submitted as a result of an extensive epidemiologic investigation of angiosarcoma in the United States from 1964 to 1974 by the Center for Disease Control, Atlanta. Georgia. Additional material came from the Envi ronmental Sciences Laboratory of the Mount Sinai School of Medicine. Some material of vinvl-chloride~related cases was provided for study directly by pathologists in Louisville, Kentucky; other material came from many pathologists from this country and Europe. Eight of the Thorotrast cases were obtained from the Armed Forces Institute of Path ology." In addition, cases were pulled from the files at the National Cancer Institute and at Mount Sinai Hospital. In all instances, histologic slides were studied, and, in some, blocks or wet tissue were also available. Where possible, additional special stains were performed such as Mallory's trichrome, chromatrop aniline blue, silver for reticulum (mainly collagen Type III)" without toning to leave hard collagen (Type I) yellow-brown (otherwise stained by.conventional connective tissue stains of the aniline blue variety), PAS reaction after removal of glycogen by diastase to identify increased phagocytosis, iron reaction, and Shikata stain to demonstrate elastic tissue and the s component of hepatitis B antigen. The material studied as well as sex and age distribution are listed in Table 1, In the vinyl-chloride-associated material is a series of cases in which multiple biopsies or biopsies and subsequent autopsies could be studied. The fibrotic precursor lesion was identified in biopsies and occasionally in autopsy specimens in areas sufficiently removed from the angiosarcoma to reasonably exclude mechanical or other effects. These were male patients, and the estimated exposure to vinyl chloride varied from 12 years to 28 years; the average exposure was 19.6 years. The latent period between first exposure and diagnosis of angiosarcoma ranged from 12 to 38 years; the average latent period was 22.3 years. The arsenic-associated lesions were the result of exposure to Fowler's solution for many years for psoriasis or asthma; in the 4 cases for which detailed information is available, the duration of exposure ranged from 10 to 17 years, with a mean of 14 years. Thorotrast had been given for radiologic visualization of, primarily, brain lesions many years ago. When Thorotrast administration could be verified, records indicated that it had been given 16 to 40 years before the patient sought medical care for a hepatic disease." The patient material designated "unknown" represents the histologically completed part of a survey of cases of hepatic angiosarcoma of unknown etiology retrospectively reviewed after death certificate, autopsy, or surgical biopsy recorded the above diagnosis. Among 117 cases, 4 were children. In this survey the cases not conforming to this diagnosis were eliminated, and in the confirmed cases a search for etiologic factors was initiated but is not complete.7* Among the factors under consideration are previous liver disease, other chemicals, and various sources of exposure to known etiologic agents such as arsenic. Results Features Distinguishing Etiologic Factors The same pattern and evolution were encountered in all the cases studied, independent of the etiology, with hardly any exceptions. The most prominent feature of the Thorotrast-associated lesions was deposi tion of large amounts of Thorotrast in dense connective tissue in the capsule and in the portal tracts but often also within the parenchyma. Proliferation of bile ductules was absent or only rarely seen in the Thoro trast-associated cases, in contrast to the vinyl chloride and arsenical as well as cryptogenic cases, in which marked bile duct proliferation was com mon. Hematopoietic foci were found in cases of each etiology. Therefore, Vol. 92. No. 2 August 1978 HUMAN HEPATIC ANGIOSARCOMA 353 in the following presentation of the observations, little reference will be made to etiology. The same holds true with few exceptions for the legends of the illustrations. Precursor Stage This stage was characterized by hepatocytic proliferation associated to varying degrees with sinusoidal lining cell proliferation and with focal sinusoidal dilatation. The hepatocytic hyperplasia was observed in two forms which varied in size and degree of demarcation, uniformity of hepatocellular hyperplasia, and extent of associated sinusoidal lining cell activation and fibrosis. In the first form, multiple, poorly circumscribed foci of hepatocytes exhibited variations from the surrounding parenchyma. The cells varied in size and formed two-cell-thick plates with their nuclei adjacent to the sinusoidal border. Binucleated and even multinucleated hepatocytes were seen (Figure 1). The bile canaliculi were sometimes dilated but devoid of bile plugs. Lipofuscin pigment was often increased. The sinusoidal lining cells in these foci were increased in number and varied in morphologic appearance. They included normal endothelial cells, plump cells with spindle-shaped nuclei and with diffuse PAS-positive diastase-resistant reaction of the cytoplasm, a few macrophages, and many lipocytes, identi fied by small fat droplets. The silver-impregnated reticulum framework was barely increased. The second form was observed when transformation to angiosarcoma was present or subsequently established. A uniform, conspicuous hyper plasia and hypertrophy of hepatocytes in contiguous, almost nodular areas, was associated with an even more conspicuous increase in various sinusoidal cells (Figure 2). The sinusoidal cell proliferation included nor mal and enlarged endothelial cells, macrophages with PAS-positive dias tase-resistant granules, lymphocytes, and occasionally plasma cells or segmented leukocytes (Figure 3). This sinusoidal cell reaction differed from that seen in hepatitis in that it was not associated with degeneration and necrosis of hepatocytes. The reticulum framework was increased in silver stains and showed an excess of both longitudinal and cross fibers. The irregular areas with increased reticulin were recognized under lowpower magnification (Figure 4). The areas of hepatocytic and sinusoidal cell hyperplasia and hypertrophy distorted the lobular architecture. The areas usually involved only part of the lobule but occasionally extended into a neighboring one. They were usually closer to the portal tract than to the central zone and were not round but garland-shaped. The surround ing parenchyma, also curved, was either normal or compressed (Figure 5). The hepatocytes in the compressed areas often revealed degeneration, R&S 109733 (/) 354 POPPER ET AL American Journal ] of Pathology: and PAS-positive macrophages accumulated while the reticulin frame-; work was condensed. Sometimes these compressed zones were hyperemkr or hemorrhagic. After Thorotrast exposure, Thorotrast granules were: found in the macrophages of the compressed areas; they were absent or; infrequent in the hyperplastic areas. Bile plugs were noted on the border between the hyperplastic and the compressed areas. The mixed hyperplastic areas differed from nodules in multiple nodular hyperplasia of humans and from the hyperplastic areas and neoplastic nodules of rodents by a) proliferation of various sinusoidal cells, b) increase of the reticulin framework, and c) garland-shaped outline. Focal sinusoidal dilation observed frequently in the precursor stage involved in part portions of the mixed hyperplastic areas and in part the surrounding parenchyma, without relation to lobular topography. The cells lining the irregularly dilated sinusoids were increased in number (Figure 6) and the surrounding reticulin framework was thickened. Nonspecific alterations in the surrounding parenchyma included occa sional focal steatosis and clumping of hepatocytic cytoplasm. The portal tracts showed varying degrees of fibrosis which sometimes disrupted the limiting membrane and extended into the periportal parenchyma. Portal fibrosis was occasionally associated with proliferation of bile ductules and, in a few instances, with accumulation of lymphoid cells around bile ducts (Figure 7). This pericholangitis was usually accompanied by diffuse cana licular cholestasis. Where it was possible to study the capsule, it showed focal capsular and subcapsular fibrosis in vinyl-chloride-associated cases. In the Thorotrast cases, these fibrotic areas contained Thorotrast deposits. Inflammatory cells were not noted in the areas of capsular fibrosis. Transition to Angiosarcoma Five processes participated to varying degrees in the transition to angiosarcoma: 1) proliferation with increasing anaplasia of intralobular endothelial cells; 2) initial hyperplasia of hepatocytes followed by atrophy and disappearance; 3) increasing fibrosis in perisinusoidal spaces; 4) pro gression of sinusoidal dilatation to peliosis; and 5) sarcomatous transfor mation of lining cells of sinusoids and of portal capillaries. The combina tion of the processes resulted in three pathways to angiosarcoma: 1. Intralobular growth with fibroplasia. The number of sinusoidal cells in the hyperplastic areas increased further so that they were in places arranged in more than one layer, although flat endothelial cells lined part of the sinusoidal spaces. The conspicuous hepatocytic hyperplasia might result in hepatocytic trabeculae lined by increased sinusoidal cells (Figure VoI. 92, No. 2 August 1978 HUMAN HEPATIC ANGIOSARCOMA 355 R&S 109735 S). The Disse space was widened and filled with many reticulum fibers in layers of variable thickness since cross fibers were also irregularlv in creased. It contained fibroblasts and many lipocytes. Some sinusoidal lining cells had a bulky cytoplasm and hyperchromatic nuclei; similar cells were found also in the Disse spaces and indented adjacent hepatocvtic plates. The sinusoids were either narrowed or somewhat dilated in places and contained an increased number of inflammatory cells. Even tually, spindle-shaped lining cells predominated, while lymphocytes, macrophages, and other inflammatory cells disappeared. Atypical cells formed multiple uniform layers lining the sinusoids and, because of increasing atypia and anaplasia, were considered to be angiosarcoma cells. Further progression was reflected in several features which occurred coincidentally with the previously described changes: a) Hard connective tissue (Type I collagen) associated with elastic fibers increased con spicuously around angiosarcoma cells within and around vascular spaces and compressed adjacent hepatic plates which atrophied and resembled bile ductules (Figure 9) and finally also disappeared. Fibrotic areas often became hyalinized and replaced many lobules, b) The angiosarcoma cells filled and extended the sinusoidal spaces to interfere with microcirculation, and the hepatic parenchyma showed anoxic necrosis (Figure 10). c) The angiosarcoma cells formed large clusters and, occasionally, solid nodules (Figure 11) composed mostly of spindle-shaped and some times of polyhedral cells. 2. Intralobular growth with sinusoidal dilatation. This pattern was an accentuation of the sinusoidal dilatation which was noted mainly in the mixed hyperplastic areas. It was characterized by further proliferation of sinusoidal and perisinusoidal cells. With increasing dilatation, adjacent hepatic plates were disrupted so that small cystic spaces formed by confluence of dilated sinusoids. These irregular spaces were separated by plates of hyperplastic hepatocvtes (Figure 12) and the surrounding frame work was remarkably thickened (Figure 13). Often, hematopoietic cells accumulated in the dilated sinusoids. Into larger spaces, blindly ending spurs of hepatocvtes extended, covered by atypical sinusoidal or angiosar coma cells. Eventually, the diffuse irregular loosening of the parenchymal architecture by some uniform and some irregular sinusoidal dilatation produced a plexiform, trabecular arrangement of the hepatocvtes, with great variations in the quantitative relation between hepatocvtes and sinusoidal cells. Initially the hepatocvtes appeared hyperplastic and ar ranged in two and more cell layers around bile canaliculi which often contained bulky bile thrombi (Figure 14). When the hyperplastic hepatocytes predominated, the picture resembled the trabecular type of he- 356 POPPER ET AL American Journal of Pathology R&S 109736 patocellular carcinoma although the hepatocytes were not anaplastic. The considerably widened Disse spaces around the trabeculae contained a variety of mesenchymal cells, including segmented leukocytes, macro phages, lipocytes, and lymphocytes, as well as spindle-shaped angiosar coma cells (Figure 15) and a significantly increased, irregularly arranged reticulum framework. The angiosarcoma cells, often in two layers, did not contain PAS-positive glycogen-resistant granules indicative of phago cytosis, nor iron granules or Thorotrast, even when the latter was found in other locations. The destruction of the lobular architecture by the de scribed loosening need not involve the portal and the central canals. They remained as beam-like trabecular structures containing vessels and bile ducts and traversed the enlarging blood spaces. They were covered by layers of angiosarcoma cells, which also infiltrated these beams which underwent hyalinizing fibrosis (Figure 16). Bile ductules were similarly initially preserved and appeared also as trabeculae. The blood spaces eventually became grossly visible blood cysts in which fibrin clots formed. Another variant of the trabecular form of angiosarcoma was gradual fibrosis in the hepatocytic cords while inflammatory cells disappeared (Figure 17). Hard, Type I collagen, intermixed with elastic fibers, formed in the reticulin matrix and predominated in later stages. A third variant was relatively thin fibrotic stalks lined by spindle-shaped angiosarcoma cells, presenting a papillary arrangement (Figure 18). In all three variants, angiosarcoma cells also formed solid clusters and, eventually, nodules. 3. Portal growth. Mainly spindle-shaped angiosarcoma cells lined capil laries or lymphatics of the portal tracts or appeared single or in small nests in their connective tissue (Figure 19). This was associated, even in small lesions, with involvement of the periportal parenchyma. Only in few instances was angiosarcoma restricted to the portal tracts. Thus, portal involvement appeared to be caused by spread of intralobular angiosar coma. This growth pattern was accompanied by proliferation of bile ductules surrounded by inflammatory cells intermixed with angiosarcoma cells. Considerable fibroplasia, mainly of reticulin character, but, sub sequently, also hvalinized collagen, produced masses of fibrous tissue, with varying amounts of either ductules or angiosarcoma cells together with occasional solid nests of predominantly spindle-shaped angiosarcoma cells. Portal angiosarcoma was accompanied in a few instances by connective tissue septums containing proliferated bile ductules, some venules and arterioles, and angiosarcoma cells intermixed to various degrees with inflammatory cells. These septums formed bridges connecting portal Vol. 92. No. 2 August 1978 HUMAN HEPATIC ANGIOSARCOMA 357 tracts with each other or with central canals. They subdivided the lobules to create the picture of a cirrhosis (Figure 20). In the thus formed parenchymal nodules the hepatocvtic plates were rearranged, usually in layers two and more cells thick, but, in contrast to common cirrhosis, they had an increase in reticulin and in sinusoidal cells surrounding the hepatocytes. Variations in Angiosarcoma Pattern Clusters and nodules were composed c, igiosarcoma cells of two types. One type was spindle-shaped and had large nuclei and distinct cytoplasmic extensions (Figure 11). The nuclei were often hyperchromatic and sometimes multiple, but nucleoli were small. Few vascular spaces were lined by angiosarcoma cells but, occasionally, by flat endothelial cells. The other type of cell was large and had abundant cytoplasm. These polyhedral cells were often multinuclear and shosved greater tendency for anaplasia than the spindle-shaped cells (Figure 21). The polyhedral cell nodules had few vascular spaces, while necrosis and hemorrhage were common. Transitions between and admixtures of both types of cells were frequent. The polyhedral cells were not intermixed with inflammatory cells except following necrosis. Sometimes they formed multiple layers lining blood-filled cystic cavities (Figure 22). Both types of nodules, which also were grossly visible, compressed the surrounding parenchyma and had often a connective tissue pseudo capsule. They were frequently associated with invasion of veins by tumor cells, mainly of branches of the portal vein. Extensive necrosis was com mon with or without vein invasion. With silver stains, some areas of necrosis showed persisting hepatic trabeculae and portal tracts, while others failed to reveal remnants of hepatic structure. Specific developmental patterns were distinguished to trace the evolu tion of the multicentric angiosarcomas to sinusoidal, trabecular, papillary, portal, and nodular end stages." Combinations, however, of several pat terns were usually seen in different parts of the same liver. For instance, fibrosing areas around hepatocytes, presumably derived from intralobular lesions, merged with others around proliferated ductules originating in portal tracts. Two cases of unknown etiology had a combination of both angiosarcoma and hepatocellular carcinoma. In one of them, nodules of these two types of malignant growth were separated by a fibrous pseudo capsule (Figure 23). Additional Features Foci of hematopoietic cells, mainly nucleated red cells but also mega- R&S 109737 u & (/> 358 POPPER ET AL American Journal of Pathology karvocytes, were frequent in all forms of angiosarcoma (Figure 24). The hematopoietic cells were also in the perisinusoidal space but more fre quently in the spaces lined by the angiosarcoma cells. The portal tracts showed variable amounts of lymphocytic infiltration. Bile ducts and duc tules showed not only proliferation of their epithelial cells, associated with hvperchromasia and formation of several layers, but also excess mucus formation, irregular dilatation, and out-pouching of epithelium into the surrounding, often hyalinized, connective tissue. Extrahepatic metastases (although multicentric origin is not excluded) of either spindle-shaped or polyhedral cell character were reported in 69% of the cases (Table 1). There was a lower percentage of metastases in those of known etiology, notably Thorotrast, than in those of unknown etiology, but the differences are probably not significant. The autopsy records of the entire series list involvement of other organs as follows: lung and pleura (35%), spleen (28%), lymph nodes (25%), bone (23%), adrenal glands (14%), diaphragm (9%), heart (8%), and brain (8%). Many other organs were involved rarely. Spleen involvement was recorded far more frequently (36%) in the group of unknown etiology than in the vinyl chloride (13%) or Thorotrast group (11%). Diagnosis Diagnostic criteria are particularly important in the evaluation of biopsy specimens. The following features proved useful in the survey of the material. The trabecular pattern with sinusoidal dilatation was most diagnostic when plates of hyperplastic hepatocytes were covered by single or multiple layers of angiosarcoma cells and hepatocytic spurs extended into blood-filled spaces. A second diagnostic feature was anaplasia of sinusoidal lining cells with or without an associated inflammatory reac tion, especially when accompanied by sinusoidal dilatation and peri sinusoidal fibrosis. The diagnosis of angiosarcoma was more difficult when only a few anaplastic cells were observed in the sinusoids since lymphoma, myeloma, or single metastatic carcinoma cells had to be excluded. The association with inflammation is characteristic for angiosarcoma. A third feature, more readily overlooked especially in small biopsy specimens, is angiosarcoma cells mixed with inflammatory cells and proliferated bile ductules in portal spaces. Solid tumor nodules offered the greatest differ ential diagnostic problem; particularly, polyhedral cells are difficult to distinguish from hepatocellular carcinoma and other types of sarcoma. Connective tissue stains, including silver stains, assist in the distinction from hepatocellular carcinoma. Frequent misdiagnoses were encountered in the material available for study because lymphoma, heman- 6i.60L SSH Table 1--Reviewed Hepatic Angiosarcoma Material Age Sex at death (yrs> Race Etiology No. M F Average Range W B Vinyl chloride Thorotrast Fowler's solution Unknown Children Females Mates Total 19 26 5 4 28 85 167t IS IS 4 1 -- 85 128 ___ 52 36-67 19 7 54 1 49 32-73 41-58 25 5 4 mo 33 13 28 57 -- 61 39 -- _ 18-82 18-89 -- 4 23 81 157 1 3 3 7 * Information about metastases not available for 1 vinyl chloride and 2 male cases, t Includes 2 orientals and 1 whose race was unknown. Specimens Biopsy and Biopsy Autopsy autopsy 13 17 11 9 19 2 2 52 No. of cases with metastases 10/16* 8/19 3/5 4 44 13 18 3 22 70 7 63 133 29 3/4 16/18 50/68* 90/130 Vol. 92. No. 2 August 1978 359 HUMAN HEPATIC ANGIOSARCOMA > 360 POPPER ET AL intralobular as with fibroplasia American Journal of Pathology ^SECONDARY ` PE LI OSlS Text-figure 1--Proposed schema of evolution of hepatic angiosarcoma (AS). giopericytoma, mixed mesenchymoma, hemorrhagic hepatocellular carci noma, or various types of metastatic tumor were diagnosed as angiosar coma. The mixed hyperplastic areas are not readily diagnosed in needle biopsv specimens. Silver stains are then most helpful. Discussion Morphologic study of a large number of cases of angiosarcoma and their developmental stages revealed few, if any, differences between the cases induced by agents such as vinyl chloride, arsenic, Thorotrast and those in which no etiologic agent has been established. Intensive study of a far larger number of cases did not confirm our early impression that vinylchloride-related angiosarcoma differs morphologically from cryptogenic cases.4 This suggests that in cryptogenic cases, environmental factors are responsible, the nature of which is being sought by intensive epidemio logic studies.74 The available surveys 19'!0 indicate a recognized env iron mental etiology in only a small percentage of cases of hepatic angiosar coma. The observations provided information of potential assistance in the diagnosis of the lesion, including the precursor stage. This information should be helpful in the diagnosis of industrial hazards. The enforcement of strict hygienic regulations has reduced the levels of vinyl chloride in the polymerization plants or other places where polyvinyl chloride is proc- A - 1 h i . Vol. 92, No. 2 August 1978 HUMAN HEPATIC ANGIOSARCOMA 361 liu g o * essed.41,79 Presumably no additional instances of initiation will occur. However, in view of the long period of promotion of tumors, addi tional cases of angiosarcoma, unfortunately, have to be expected in the future. The differential diagnostic criteria max' have practical imporjj tance if other environmental and industrial hazards produce the same So morphologic sequences. The diagnostic significance of the characteristic precursor stage of mixed hyperplasia and hypertrophy of hepatocytes and sinusoidal cells, frequently associated with sinusoidal dilatation, may be helpful in screen ing, although its specificity requires further study. It is best recognized in silver stains even with low-power examination. The second purpose of this paper is an attempt to better understand the development of hepatic angiosarcoma (Text-figure 1). Characteristically it entails, at least in early stages, a proliferation and hyperplasia of both hepatocytes and sinusoidal cells. The metabolic factors accounting for this interplay remain to be established. The appearance points to a prolifera tive, potentially carcinogenic stimulus for the hepatocytes. This is sup ported by occasional primary hepatocellular carcinomas seen after arsenic and Thorotrast and, in rare cases, after vinyl chloride exposure. In young rats with less-developed microsomal biotransformation, including lesser metabolite degrading activity,77 vinyl chloride exposure leads more fre quently to hepatocellular carcinoma than in adult rats.10 This observation points to the possibility that variations in enzymatic biotransformation may determine the type of tumor that develops. In experimental animals, both angiosarcoma and hepatocellular carcinoma have been produced, sometimes simultaneously.78 In human hepatocellular carcinoma, varying degrees of sinusoidal cell activity are noted, also expressed in variations of the reticulum framework around the trabecular form of this tumor. This suggests variable stimulation of sinusoidal cells even in frank hepatocellu lar carcinoma. Cases of both tumors in humans have been seen very rarely. Thus, the interaction between carcinogenic effects on sinusoi dal cells and hepatocytes is a promising area for further study. Peculiarly, in Thorotrast-induced epithelial tumors, bile duct carcinoma may be more frequent than hepatocellular carcinoma.78 The described alterations of the bile ducts in the angiosarcomas studied in this series suggest a carcino genic stimulus on bile duct cells, but the possibility that this is only a secondary reaction to the presence of a tumor in the liver cannot be excluded. Reports of intensive investigations of the metabolism of vinyl chloride are available.41,80"83 It is transformed by hepatic microsomal action to mutagenic84"88 and potentially oncogenic87 metabolites which covalently bind to DNA 88,88 and are injurious to hepatocytes in animals exposed to t? R&S 109742 362 POPPER ET AL American Journal of Pathology very large doses of vinyl chloride.90 These metabolites are then rendered inactive by enzymatic or nonenzymatic action or by binding to gluta thione.91 In the precursor stage, significant degenerative alterations of the hepatocytes, such as steatosis, focal necrosis, and lipofuscin deposits which are not age-dependent, have only been recognized histologically after short intervals between exposure and biopsy; these features subsequently re gress.91 This explains why conventional liver function tests are of little diagnostic value in this stage. Only with progression of the process to tumor formation appear measurable alterations of hepatic function. An giographic 92 and scintigraphic99 demonstration of the lesion has greater diagnostic value and the lesions are also reflected in hemodynamic altera tions.94 A predominant early alteration is the sinusoidal cell proliferation associ ated with sinusoidal dilatation and fibrosis. While in such stages these cells are intermixed with various inflammatory cells, a transformation to angiosarcoma cells is associated with a disappearance of the other mesenchymal cells in all growth patterns. Hypothetically, the inflamma tion may have a suppressing effect from which the angiosarcoma cells seem to escape. Azoxymethane-induced angiosarcoma growth is stimu lated by antilymphocytic serum which has no effect on the hepatocellular carcinoma produced by this agent.99 An immune complex disorder has been claimed as the basis of the vinyl chloride injury.9* This could explain the inflammatory reaction in both parenchyma and portal tracts. Transition to angiosarcoma cells from the sinusoidal lining cells could be traced. They appeared to be derived from endothelial cells, not only because they were devoid of any phagocytic activity but also because they were clearly distinguished from phagocytic macrophages. Their origin from endothelial cells was also confirmed by electron microscopic study.97 However, better confirmation by use of histologic markers of endothelial cells, for instance, Weibel-Palade bodies,99 or of catalase and per oxidases 99 is desirable. The angiosarcoma cells exhibited various forms of arrangement to include a) lining or tectorial, b) enveloping hepatocytes or bile ductules, c) vasoforming, or d) solid. They were either spindle-shaped with relatively little cytoplasm or polyhedral. The latter sometimes resem bled epithelial cells and thus created differentia] diagnostic problems. The solid or nodular forms developed from almost any growth pattern, includ ing the intralobular, the portal, and the trabecular. Angiosarcoma may develop in the lobular parenchyma and in the portal tracts. The primary growth in the parenchyma appears to be far more frequent, with the invasion of the portal tracts being secondary. In the , n g fc .h. ^ I- I t h i' i h iI m iu i i ii. .! ` .> ,, .^ Vol. 92, No. 2 August 1973 HUMAN HEPATIC ANGIOSARCOMA 363 fibrotic stages particularly, both growth patterns are combined, reflected in the presence of both hepatocytes and ductules. The sinusoidal dilatation found in the majority of instances is not explained by passive congestion because of its irregular location in the lobule. Similar forms of sinusoidal dilatation without angiosarcoma have been reported in women taking oral contraceptive drugs100 and in men and women receiving large doses of anabolic steroids.101 In the latter cases, the peliosis may be sufficiently advanced to produce hepatic failure. In these instances, it has been explained by injury to the sinusoidal lining 102 and this was also demonstrated in vinyl-chloride-intoxicated mice." The observations here are in keeping with this assumption of a sinusoidocidal effect. In contrast to this, primary peliosis, the progression of which to large bloody cysts is associated with persistence of hvalinized connective tissue beams, best seen in silver stains, is a secondary peliosis occurring within angiosarcoma nodules undergoing extensive central ne crosis and hemorrhage. In these instances no septums remain in the silver stains. Necrosis with hemorrhage is the predominant gross manifestation of these tumors. The relation of the sinusoidal cell hyperplasia to fibroplasia represents an interesting problem. The excess of perisinusbidal and sinusoidal cells of several types, accompanied at least initially by hepatocytic hyperplasia, was regularly associated with excess reticulum fibers. This reticulin in crease was also found in the presence of sinusoidal dilatation and was a useful diagnostic criterion. The morphologic appearance suggests a fibro blastic activity of some of the perisinusoidal cells.10* Increased and en larged Ito cellsIM or lipocytes,10* small fat droplets containing peri sinusoidal cells, were observed in light and in electron microscopic studies,BliB7,l0B,1<n particularly in the fibrotic precursor stage. They are considered precursors of fibroblasts.108-10B Their transformation to fibro blasts is associated with formation of collagen Type III, mainly reticu lum.10 Also, the splenic enlargement after vinyl chloride exposure results partially from proliferation of fibroblastic cells.110 Subsequent stages of the hepatic lesion are associated to a varying degree with desmoplasia, either in the form of a creeping fibrosis around hepatocytes, often in pseudoductular arrangement or around bile ductules, or as massive hvalinized fibrosis. Then hard, collagen Type I, a different gene product than colla gen Type III, develops within the reticulum and eventually replaces it. This implies a significant desmoplastic potential of the tumor, although the role of the fibroblasts in the later stages has not been established. A peculiar form is the septal fibrosis which leads to an almost cirrhotic picture. R&S 109743 </) 364 POPPER ET AL American Journal of Pathology The characteristic abundance of hematopoietic cells, including mega karyocytes, in the developing angiosarcoma, both intrasinusoidal and perisinusoidal, may reflect a local formation of these cells in the activated mesenchyme 103 or bone marrow irritation. The great variability in appearance of the tumor is accounted for by the variable growth potential of hepatocytes and sinusoidal cells and by the irregular tendency for fibrosis and sinusoidal dilatation. However, in virtually all cases a multicentric angiosarcoma of variable type results. References 1. Creech JL Jr, Johnson MN: Angiosarcoma of liver in the manufacture of polyvinyl chloride. J Occup Med 16:150-151, 1974 2. Maltoni C, Lefemine C: Carcinogenicity bioassays of vinyl chloride. 1. Research plan and early results. Environ Res 7:387--105, 1974 3. MacMahon HE, Murphy AS, Bates MI: Endothelial-cell sarcoma of liver follow ing thorotrast injections. Am J Pathol 23:585-611, 1947 4. Roth F; Arsen-Leber-Tumoren (Haemangioendotheliom). Zschr Krebsforsch 61:468-503, 1957 5. Marsteller HJ, Lelbach WK, Muller R, Juhe S, Lange CE, Rohner HG, Veltman C: Chronisch-toxische Leberschaden bei Arbeitern in der PVC-Produktion. Dtsch Med Wochenschr 98:2311-2314, 1973 6. Thomas LB, Popper H, Berk PD, Selikoff I, Falk H: Vinyl-chloride-induced liver disease; From idiopathic portal hypertension (Banti's syndrome) to angiosarcomas. N Engl J Med 292:17-22, 1975 7. Morris JS. Schmid M, Newman S, Scheuer PJ, Sherlock S: Arsenic and noncirrhotic portal hypertension. Gastroenterology 66:86-94, 1974 8. da Silva Horta J: Late effects of thorotrast on the liver and spleen, and their efferent lymph nodes. Ann NY Acad Sci 145:676-699, 1967 9. Smoron GL, Battifora HA: Thorotrast-induced hepatoma. Cancer 30:1252-1259, 1972 10. Maltoni C: Predictive value of carcinogenesis bioassavs. Ann NY Acad Sci 271:431--447, 1976 11. Gokel JM, Liebezeit E, Eder M; Hemangiosarcoma and hepatocellular carcinoma of the liver following vinyl chloride exposure: A report of two cases. Virchows Arch [Pathol Anat) 372:195-203, 1976 12. Adam YG, Huvos AG, Hajdu SI: Malignant vascular tumors of liver. Ann Surg 175:375-383, 1972 13. Pollard SM, Millward-Sadler GH: Malignant haemangioendothelioma involving the liver. J Clin Pathol 27:214-221, 1974 14. MacSween RNM, Vetters JM, Ross SK, Ferguson J. Johnstone JM, Sandison AT: Haemangio-endothelial sarcoma of the liver. J Pathol 109:39-44, 1973 15. Sugahara K, Shirakura T, Kawano N. Kino I: Primary malignant heman gioendothelioma of the liver; Report of a successful case of resection and review of the Japanese literature. Am J Gastroenterol 62:240-244. 1974 16. Fukuhara M, Tsujii T, Morita T, Ishikavva H: An autopsy case of primarv cardiac hemangiosarcoma and analysis of cases in the literature. Jpn J Med 15:235-241, 1976 17. Fiechtner JJ, Reyes CN Jr: Angiosarcoma of the liver in a rural population: Four cases diagnosed in a 29-month period. JAMA 236:1704-1706. 1976 18. Chowdhury AR, Black M. LorberSH, Chev WY: Hemangioendotheliomatosis of the liver: A 12-\ear follow-up. Gastroenterology 72:157-160, 1977 Vol. 92. No. 2 August 1978 HUMAN HEPATIC ANGIOSARCOMA 365 19. Baxter PJ, Anthony PP, MacSween RN'M, Schcuer PJ: Angiosarcoma of the hu-r in Great Britain. 1963-73. Br Med J 2:919-921. 19i7 20. Bradv J. Liheratore F. Harper P, Greenwald P. Burnett tV, Davies JNP, Bishop M, Polan A. \'ianna N: Angiosarcoma of the liser: An epidemiologic survcs. J N'atl Cancer Inst 59:1383-1335. 1977 21. Kurrein F: Systemic angioendotheliomatosis with metastases, J Clin Pathol 29:347-353. 1976 22. Edmondson HA: Tumors of the liver and intrahcpatic bile duels Atlas of Tumor Pathology. Section 7. Fascicle 25. Armed Forces Institute of Pathology, Washing ton, DC, 1958 13, Enzinger FM. Smith BH; Hemangiopericytoma. An analvsis of 106 cases. Hum Pathol 7:61-82. 1976 24, Baker H de C, Paget GE, Davson J: Haemangio-endothelioma (KupfFer-cell sar coma) of the liver. J Pathol 72:173-182, 1956 25. Leevy CM, Popper H, Sherlock S (editors): Diseases of the Liver and Biliary Tract: Standardization of Nomenclature, Diagnostic Criteria, and Diagnostic Meth odology. Fogartv International Center Proceedings No. 22, DHEW Publication No. (NTH) 76-725. 1976 26. Falk H, Creech JL, Heath CW Jr, Johnson MN, Key MM: Hepatic disease among workers at a vinyl chloride polymerization plant. JAMA 230:59-63, 1974 27, M4kk L, Creech JL. Whelan JG, Johnson MN: Liver damage and angiosarcoma in vinyl chloride workers. A systematic detection program. JAMA 230:64-68, 1974 28. Ishak KG: Mesenchymal tumors of the liver. Hepatocellular Carcinoma. Edited by K Okuda, RL Peters. New York, John Wiley & Sons, 1976, pp 2S6-298 29, Delorme F. Mikk L: Angiosarcomes du foie chez des ouvriers ayant ete en contact prolonge avec le chlorure de vinvle: Description morphologique des lesions. Union Med Can 104:1836-1844. 1975 ' .30. Lee FI, Harrv DS: Angiosarcoma of the liver in a vinvl-chloride worker. Lancet 1.1316-1318, 1974 31. Smith PM, Williams DM J, Evans DM D: Hepatic angiosarcoma in a vinyl chloride worker. Bull NY Acad Med 52:447-452, 1976 32. Lange C-E, Jiihe S, Veltman G: Ueber das Auftreten von Angiosarkomen der Leber bei zwei Arbeitern der PVC-herstellenden Industrie. Dtsch Med Wochenschr 99:1598-1599, 1974 33. Zimmermann H, Eck H: Zur pathologischen Anatomie der Vinylchlorid-Krankheit. Virchows Arch [Pathol Anat) 368:51-59, 1975 34. Wallnfifer H, Zinnagl N: Haemangiosarkomatose nach Polvvinvlchlorid-exposition. Med Klin 72:410-413, 1977 35. Muller R, Bechtelscheimer H, Gedigk P, Marsteller HJ, Lelbach WK: Das morphologische Bild der Lebersehaedigung nach chronischer Yinylehlorid-Exposition. Leber Magen Darm 5:204-208, 1975 36. Ravier E, Diter JM, Pialat J: Un cas d'angiosarcoma hepatique chez un ouvrier expose au chlorure de vinvle monomere. Arch Mai Prof 36.T71-177, 1975 37. Rtv J, Sauvage T, Habrard A: Le 3 cas frangais d'angiosarcome hepatique chez un ouvrier du chlorure de vinvle. Arch MaJ Prof 37:551-556, 1976 38. Bonneton G, Champetier J, Sotto JJ, Guidicelli H, Letoublon C. Pahn M: Un cas de rupture spontanee d'angiosarcome hepatique chez un travailleur expose au chlorure de vinvle. Chirurgie 101:936-942, 1975 39. Biersack HJ, Lange CE, Ebinger H, Marsteller HJ, Lelbach WK, Veltman C, Winkler C: Sequenzszintigraphische Untersuchungen von Leber und Milz bei Patienten mit Vinylchlorid-Krankheit. Dtsch Med Wochenschr 100:615-617, 1975 40. Marsteller HJ, Lelbach WK, MOller R, Gedigk P, Lange CE: Klinische und laparoskopische Aspekte der Leberschaeden bei Chemiearbeitem in der Vinylchlorid-Polymeri5ation. Leber Magen Darm 5:196-202, 1S75 366 POPPER ET AL American Journal Of Pathology 41. Gutacker HW, Lelbach WK (editors): Leberschaden durch Vinylchlorid-Krankheit. Baden-Baden, Verlag Gerhard Wjtzstrock, 1977 42. Smith PM, Crossley 1R, Williams DMJ: Portal hypertension in vinyl-chloride production workers. Lancet 2:602-604, 1976 43. Wilson RH, McCormick WE, Tatum CF. Creech JL: Occupational acroosteolysis: Report of 31 cases. JAMA 201:577, 1967 44. Lange CE, Juhe S, Stein G, Veltman G: Die sogenannte Vinylchlorid-Krankheit: Eine berufsbedingte Systemsklerose? Int Arch Occup Environ Health 32:1-32, 1974 45. Suciu I, Drejman I, Valaskai M: Etude des maladies dues au chlorure de vinyle. Med Lav 58:261-271, 1967 46. Pushin GA: 0 porashenii petscheni i sheltschnyich putei u rabotschich zanjatyich w proizwodstwe nekatorivich widow plastmass. Sov Med 23:132, 1965 47. Smith PM, Williams D.\ij: Vinyl chloride and cirrhosis. Digestion 10:321-322, 1974 48. Huet P-M, Guillaume E, Cdtd J, Legar6 A, Lavoie P, Viallet A: Noncirrhotic presinusoidal portal hypertension associated with chronic arsenical intoxication. Gastroenterology 63:1270-1277, 1975 49. Knolle J. FOrster E, Roessner A, Themann H, Hiihn P, Meyer zum BOschenfelde K-H: Die nicht-zirrhotische portale Fibrose (hepatoportale Sklerose) nach chronischer Arsenvergiftung: Klinische und pathologisch-anatomische Befunde. Dtsch Med Wochenschr 99:903-908, 1974 Datta DV: Arsenic and non-cirrhotic portal hypertension. Lancet 1:433, 1976 Regelson W, Kim U, Ospina J, Holland JF: Hemangioendothelial sarcoma of liver from chronic arsenic intoxication by Fowler's solution. Cancer 21:514-522. I96S 52. Lander JJ, Stanley RJ, Sumner HW, Boswell DC, Aach RD: Angiosarcoma of the liver associated with Fowler's solution (potassium arsenite). Gastroenterology 68:1582-1586, 1975 53. Denk R, Holzmann H. Lange HJ, Greve D: Ueber Aisenspaetschaden bei obduzierten Moselwinzern, Med Welt 20:557-567, 1969 54. Luechtrath H: Cirrhosis of the liver in chronic arsenical poisoning of \ intners. Ger Med Mon 2:127-123. 1972 oo. da Silva Horta J: Late lesions in man caused by colloidal thorium dioxide (thoro- trast): A new case of sarcoma of the liver twentv-two vears after the injection. Arch Pathol 62:403-418. 1956 56. da Silva Horta J, Cayolla da Motta L: Follow-up study of thorium dioxide patients in Portugal. Ann NY Acad Sci 145:830-842, 1967 Dahlgren S: Thorotrast tumours: A review of the literature and report of two cases. Acta Pathol Microbiol Scand 53:147-161. 1961 58. Visfeldt J, Poulsen H: On the histopathologv of liver and liver tumours in tho rium-dioxide patients, Acta Pathol Microbiol Scand [A] S0-97-10S. 1972 59. Tesluk H, Nordin W.A: Hemangioendothelioma of lixer following thorium di oxide administration. Arch Pathol 60:493-501, 1955 60. Nettleship A, Fink WJ: Neoplasms of the liver following injection of thorotrast. Am J Clin Pathol 35:422-426, 1961 61. Telles NC, Thomas LB, Popper H. Ishak K, Falk H: Evolution of thnrotrastinduced hepatic angiosarcomas. Environ Res (In press) 62. Swarm RL, Miller F,, Michelitch HJ: Malignant snscular tumors in rabbits in jected intravenously with colloidal thorium dioxide. Pathol Microbiol t Basel) 25:2744. 1962 63. Pimentel JC, Menezes P: Liver disease in vinevard spravers. Gastroentcrologs 72:275-283, 1977 64. Viola PL, Bigotti A, Caputo A: Oncogenic response of rat skin, lungs, and bones to vinyl chloride. Cancer Res 31:316-522, 1971 R&S 109746 R&S 109747 Vol- 92. No. 2 August 1973 HUMAN HEPATIC ANGIOSARCOMA 367 65. Popper H. Maltoni C. Selikoff IJ, Squire RA, Thomas LB: Comparison of neo plastic hepatic lesions in man and experimental animals Cold Spring Harbor Conferences on Cel! Proliferation, Yol 4: Origins of Human Cancer, Cold Spring Harbor Laboratory, 1977, pp 1359-1382 66 Waller T, Rtibarth S: Haemangioendothelioma in domestic animals Acta Yet Scand 8:234-261. 1967 67 Deringer MK; Response of strain HR/De mice to painting with urethan. J Natl Cancer Inst 29:1107-1121, 1962 68. Stewart HL: Hemangiosarcoma (malignant hemangioendothelioma). Cancer Ed ited by FF Becker. New York, Plenum Publishing Co.. 1975, vol 4. pp 303-374 69. Herrold KM: Histogenesis of malignant liver tumors induced by dimcth> Initrosamine: An experimental study in Syrian hamsters. J Natl Cancer Inst 39:1099-- I 111. 1967 70. Andervont HB: Inductiori of hemangio-endothcliomas and sarcomas in mice "ith 0-aminoazotoluene. J Natl Cancer Inst 10:927-941, 1950 71. Narisawa T, Wong C-Q. Weisburger JH: Azoxymethane-induced liver heman- giosarcomas in inbred strain-2 guinea pigs. J Natl Cancer Inst 56:653-654, 1976 72. Stanton MF: Transplantability, morphology, and behavior of polyoma virus-in duced hepatic hemangiomas of hamsters. J Natl Cancer Inst 35:201-213. 1965 73. Falk H, Telles NC. Ishak KC, Thomas LB, Popper H: Epidemiology of thorotrast induced hepatic angiosarcoma in the United States. Environ Res (In press) 74. Miller EJ. Matukas YJ: Biosvnthesis of collagen: The biochemist's vie". Fed Proc 33:1197-1204, 1974 75. Falk H. Thomas LB, Popper H, Ishak KG: Investigation of hepatic angiosarcoma in the U.S. (In preparation) 76. Nicholson WJ: Cancer following occupational exposure to asbestos and vinyl chloride. Cancer 39:1792-1 SOI, 1977 77. Stoming TA, Bresnick E: Hepatic epoxide hvdrase in neonatal and partially hepatectomized rats. Cancer Res 34:2SI0-2813, 1974 78. Adamson RH: Long-term administration of carcinogenic agents to primates. Medical Primatology. Edited by J Moor-Jankowski. Basel, S. Karger AC. 1972, pp 216-255 79. Suckow EE, Henegar GC, Baserga R: Tumors of the liver following administra tion of thorotrast. Am J Pathol 38:663-677, 1961 SO. Barbin A. Bresil H, Croisv A, Jacquignon P, Malaveille C, Montesano R. Bartsch H: Liver-microsome-mediated formation of alkylating agents from vin\ I bromide and vinyl chloride. Biochem Biophys Res Commun 67:596-603, 1975 SI. Watanabe PG, Hefner RE Jr, Gehring PJ: Vinyl chloride-induced depression of hepatic non-protein, sulfhydryl content and effects on bromosulphalein (BSP) clearance in rats. Toxicology 6:1-8, 1976 82. Watanabe PG, McGowan GR, Gehring PJ: Fate of ["CJvinyl chloride after single oral administration in rats. Toxicol Appl Pharmacol 36:339-352, 1976 83. Bolt HM, Laib RJ, Kappus H, Buchter A: Pharmocokinetics of vinyl chloride in the rat. Toxicology 7:179-188, 1977 84. Creim H, Bonse G, Radwan Z, Reichert D, Henschler D: Mutagenicity in litro and potential carcinogenicity of chlorinated ethylenes as a function of metabolic oxirane formation. Biochem Pharmacol 24:2013-2017, 1975 S5. McCann J, Simmon V, Streitwieser D, Ames BN: Mutagenicity of chloroacetaldehvde. a possible metabolic product of 1,2-dichloroethane (ethylene dichloride), chloroethanol (ethvlene chlorohydrin), vinyl chloride, and cyclophosphamide. Proc Natl Acad Sci USA 72:3190-3193, 1975 86. Loprieno N. Barale R, Baroncelli S, Bartsch H. Bronzetti G, Cammellini A, Corsi C, Frezza D, Nieri R. Leporini C, Rosellini D, Rossi AM: Induction of gene muta- R&S 109748 3G8 POPPER ET AL American Journal of Pathology 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 1066.. tions and gene conversions by vinyl chloride metabolites in yeast. Cancer Res 37:251-257, 1977 Green T, Hathsvay DE: The biological fate in rats of vinyl chloride in relation to its oncogenicity. Chem Biol Interact 11:545-562, 1975 Kappus H, Bolt HM. Buchter A, Bolt W; Rat liver microsomes catalyse covalent binding of "C-vinyl chloride to macromolecules. Nature 257:134-135, 1975 Osterman-Golkar S, Hultmark D, SegerbSck D. Calleman CJ, G5the R, Ehrenberg L, Wachtmeister CA: Alkylation of DNA and proteins in mice exposed to vinyl chloride. Biochem Biophvs Res Commun 76:259-266, 1977 Reynolds ES, Treinen Moslen MT, Szabo S, Jaeger RJ, Murphy SD: Hepatotoxicitv of vinyl chloride and l,l-dichloroethy!ene: Role of mixed function oxidase system. Am J Pathol 81:219-236, 1975 Schattenberg P-J, Totovic V, Cedigk P, Marsteller HJ: Die Ultrastruktur der Leber schiidigung bei der chronischen Vinylchlorid-Intoxikation. Virchows Arch [Pathol Anat] 373:233-247, 1977 Whelan JG, Creech JL, Tamburro CH: Angiographic and radionuclide character istics of hepatic angiosarcoma found in vinyl chloride workers. Radiology 118:549557, 1976 Biersack HJ, San Luis T Jr, Lange CE, Thelen M, Veltman G, Winkler C: Scintigraphy of liver and spleen in vinyl chloride workers. Acta Hepatogastroenterol (Stuttg) 24:357-361, 1977 Blendis. LM, Smith PM, Lawrie BW, Stephens MR, Evans WD: Haemodvnamic studies in vinyl chloride monomer workers with portal hvpertension. Gut 18:400401, 1977 Kroes R, Berkvens JM, Weisburger JH: Immunosuppression in primary liver and colon tumor induction with N-hydroxy-N-2-fluorenylacetamide and azoxymethane. Cancer Res 35:2651-2656, 1975 Ward AM, Udnoon S. Watkins J, Walker AE, Darke CS: Immunological mecha nisms in the pathogenesis of vinyl chloride disease. Br Med J 1:936-938, 1976 Schaffner F, Popper H, Selikoff IJ: Initial features of vinyl chloride (VC) hepatic injury. Gastroenterology 71:A35. 1976 Weibel ER, Palade GE: New cvtoplasmic components in arterial endothelia. J Cell Biol 23.101-112, 1964 Fahimi HD, Cray BA, Herzog VK: Cstochemical localization of catalase and peroxidase in sinusoidal cells of rat liver. Lab Invest 34:192-201, 1976 Winkler K, Poulsen H: Liver disease with periportal sinusoidal dilatation. A possible complication to contraceptive steroids. Scand J Castrnenteml 10:699-704, 1975 Bagheri SA, Bover JL: Peliosis hepatis associated with androgenic-anabolic steroid therapy. Arm Intern Med 81:610-618, 1974 N'adell J, Hosek J: Peliosis hepatis. Twelve cases associated with oral androgen therapy. Arch Pathol Lab Med 101:405-410, 1977 Popper H. Thomas LB, Schaffner F, Maltoni C, Selikoff IJ: Interaction between sinusoidal cells and hepatocytes in human and experimental angiosarcoma induced by environmental factors. Kupffer Cells and Other Sinusoidal Cells. Edited by E Wisse, DL Lnook. Amsterdam, Elsevier North-Holland Biomedical Press, 1977, pp 173-181 [to T- Recent ads ances in the study on the fine structure of the hepatic sinusoidal wall: A res iew. Cunma Rep Med Sci 6:119-163, 1973 Bronfenmajer S, Schaffner F. Popper H: Fat-storing cells (lipoevtes) in human liser. Arch Pathol Lab Med S2:447-453. 1966 Triche T, Nanba K, Ishak K, Wolkoff A, Berk PD: Hepatic ultrastructural changes in vinyl-chloride tVC) svorkers. Clin Res 23:259A, 1975 I Vol. 92. No. 2 August 1978 HUMAN HEPATIC ANGIOSARCOMA 369 10T. Knrokawa S. lnagaki T. Oknsama S: Electron microscopic observation of tin- lt\or in portal hypertension following chronic exposure to viml chloride monomer. Jpn Soe Gastroenterol 12 6-1-69, 1977 108. Schnacli H, Stockinyer L, U'ewalka F: Adventitious connective tissue cr !s in the space of Disse and their relation to fiber formation Rev lot Hepat 17:855-860. 1907 109. Kent G. Gay S, lnouye T, Balm R. Minick OT, Popper H: Vitamin A-eontaimng lipocytes and formation of l\pe III collagen in liver injurv. Proc Natl Acad Sei L5A 73:3719-3722, 1976 110. Meusermann U, Stutte HJ: Enzyinhistiochemische, histometrische und ultra.strnkttirelle Untersuehungen von Milzen bci der Vins Ichlorid-Krankheit. \'irchosvs Arch [Pathol Anat] 375:303-317, 1977 Acknowledgments Particular gratitude is expressed to Dr. John Herbert, Center fur DP Control, for Ins enthusias tic support in the conduet of this study. The authors also wish to thank . , ue Hostler for typing the manuscript and \lr. Ralph Isenluirg for photographie services. R&S 109749 -p 1r - Km> !! ' ; - *- -f.V .^V* `V* -V, R&S 109751 T . .> > 9** . ** V7**. 9, i_.-;., -Vv ^ 0 >- 4 o >.ip. *. ,> ,..;o ae fl\ ** ' Of e: c 'O'/' SO-'5' ` 9 *-V -*V:-P- r y *;--. -VAO'.fr ; #*,; ' . ,`a;5- ..0 V* #*\ >/. *> " o. ' pr -* *> `0. ;* n ' jv *.: & r> r?)V . *< a p w *5 I--Focal variation in size of hepatocytes within lobular parenchyma (between arrows). Note small and large cells <*oe*droplet steatosis in surrounding parenchyma. Liver biopsy (S74-41E) from worker exposed to vinyl chloride. **4L * 100) Figure 2--Hepatocytes in more than one-cefl-thick plates surrounded by increased sinusoidal cells " ** *!most nodular area which exerts pressure on the surrounding parenchyma. Autopsy (L7374) specimen from *posed to vinyl chloride. (H&E, X40) Figure 3--Mixed hyperplastic area. Hepatocytes, usually in two-cell- "**J**ta*. indicated by sinusoidal position of nuclei and excess of sinusoidal cells of various types. Autopsy (L72Q6) V*-Ml4i' from worker exposed to vinyl chloride. (H&E, x250) Figure 4--Increased reticulin in mixed hyperplastic ^ Autopsy (A74-163), (Silver Staint X60) T' .* /.-'h-'ii *: - ;** ' .`.r: A * l "N * * > f' 'v.a v> * *. > ;ia*'. * V *v' >. -wr. Crvr:J? 4' l!5. '<*** **2 ':& *. `S&li $3??^ CtfSi** bf . ' V .*. *** 4. 1* * t V A * CV * v/ * ` */L * * * f .*.** * * * .* ^'Jev <*-fV~. *. J *" ^7 ,, 4V - *a &*> ns?.1 tv*.* A\` ,*!st'* . *, I -^/ / V< ** .* - . f. ' *Ht *3v ."p 2** ^ '''-v,; `'*'.*V**-*'&:'iV>\*:'Itas. > >; '?'0,ox4 -- ( yO'; 'J avi. 'f*l*.v Sj -*v^` a* A M i v * *(f ` ^ ~ '"A\ (**"s *hj-e**,t}P\| .**<fcj,*0**o 'S**{' m *9 t tt . . V Q *J . l. Figure 5--Mixed hyperplastic area in leU lower comer characterized by hepatocytes m two-cell-thick plates sn^ increased sinusoidal cells. The surrounding parenchyma reveals one-cell-thick plates and appears compressed. Wllfl the hepatocytic plates bent. Autopsy (A75-75). (H&E xiOO) Figure 6--Irregular dilatation of sinusoids, the lih'tjj ceils of which are increased. Note hepatocytic nuclei near sinusoidal border as indication of two-cell or more-thi<jJ plates. Autopsy (A74-281). (H&E. xiOO) Figure 7--Portal fibrosis m part extending into periportal parenchymavriB1 loss of hepatocytes m the limiting plate. Note accumulation of lymphocytes in right lower aspect. Autopsy (A7*'t^r (H&E. x 100) Figure 8--Hepatocytes in two-cell or more-thick plates, sometimes containing bile plugs m dilated t>w canaliculi |straight arrow). Sinusoidal and perisinusoidal spaces contain a variety of cells, some showing beginning anaplasia and forming multiple layers (curved arrows). Autopsy (A74-31). (H&E. xiOO) E-re 9 *r a `.p.noie'*lls lar ' nusou "yPerpi --* ** " r.-'v > ;'* ' VjTJ' ' ' * rk r- ><*# $&*&$& -... r * ., j . !, ,._ * * . . .. J- R&S 109753 '*** *~Hepatocytes in pseudoductular arrangement surrounded by excess fibrous tissue and vascular spaces lined ' C8|,S' often in multiP|e layers {arrows). Accumulation of lymphocytes on lower border. Autopsy *** IH4E. xi 00) Figure 10--Angiosarcoma cells filling sinusoids, in upper aspect between hyperplastic and in ~mmr. *so*cf between necrotic hepatocytes. Autopsy (A77-301). (H&E, X100) Figure 11--Nodule composed of angiosarcoma cells in places lining vascular spaces which are also lined by nontumorous endothelial 77 '"'*** Biopsy (S74-816). (H&E. xi60) Figure 12--Loosening of lobular architecture by irregular dilatation of . containing angiosarcoma cells intermixed with sinusoidal lining cells of other types. Note conspicuous and hypertrophy of hepatocytes in upper aspect. Autopsy (A74-287). (H&E, X100) R&S 109754 <r A iV t; Figure 13--Increased reticulum framework around hyperplastic hepatocytes in places associated with sinusoid* dilatation. Note bundles of hard collagen (arrow). Autopsy (A74-31). (Silver stain, X250) Figure 14--Trabecul* form of angiosarcoma. Note hyperplastic hepatocytes in upper aspect, in part around thick bile plugs (arrow) s'* fibrosis of Disse space compressing hepatocytes. The Disse spaces are widened and contain a variety of cells. 1** sinusoidal spaces are lined by angiosarcoma cells. Autopsy (A74-31). (H&E, xi60) Figure 15--Trabecular form* angiosarcoma with spindle-shaped sarcoma cells covering hepatic trabeculae. Note hyperplastic hepatocytes and* variety of cells m sinusoidal spaces, some of them fat storing (arrows). Biopsy (S74-3014). (H&E. X250) Figure 16" Empty blood spaces (primary peliosis) traversed by fibrosed and hyalinized beams representing persisting port* tracts and central canals. Fibrosing, trabecular angiosarcoma in right upper and left lower corners. Autopsy (L748S1 (Silver stain, X40). i i*tn 17- ' `--267 20--Con v^rounc o-c;ure o  tr. ,..v ft ' V. s. V. ;> R&S 109755 rar-re* ' ce 1 icf: *** ty--Pfogressive fibrosis of trabecular angiosarcoma with compressed, atrophic hepatocytes (arrows). Autopsy " v<vV |H4E, xi00) Figure 18--Papillary angiosarcoma. Note tumor cells around thin connective tissue stalks **** stalks (arrows). Autopsy (A74-89). (H&E, X160) Figure 19--Angiosarcoma filling sinusoids of lobular r"^wyma and in and around vascular spaces in portal tracts (arrows). Autopsy (A74-31). (H&E. X1Q0) Figure .TrTT*^6 tissue septums linking portal with portal and central canals, containing angiosarcoma cells and w~>v'c"v nodules consisting of hyperplastic hepatocytes. in part separated by angiosarcoma, to produce the cirrhosis. Autopsy (A74-133). (H&E, X40) V" * *<fr*:: . - a. A .*l Po ' -. \j. " '; 4. ','> ^ ^ `.0/ \e *?S .i * c _s >*s^ r's-/V*/*t:-<.Ww5tw?> . '<T. Ng$';'A s&f&>:SSS,E_>i#> 5SC- V9 *SSFAr& *T ,>i / * `-w/jriy/!i(V}, . . ' -\*3$ ** w/*K. _ */*. EV-sHtii . V- ii'-AX; k'Z ,^ VX^ ' vv.*-vV*v**^'0'>7U; >WH! /\;VV..- m# v; >j vi\i .*:! :, V- . - *.-K .r**. iV Vv-v:1^ '."A Cs-''\i'\^n^;W>V-.-r'i?- * \^-r*' . Vr/ -v.- ' r*-*.5 cV'-*vy; ; Vi V .. Wr 'mmmm; *v R&S 109756 ,. ,? * '- - *4, *V `i_O * '' - - v---, - '>' _. V-' - :\?~7>< -, - *..->. v?>v; r .I/hV *./.'*''.! J, , / "-^-~ ,, ..*>" ? * %*> -, *' ``VV, `^--,s***/ -- p. .** <*%;cr*r:v . * '^Vj* r V jph4 . a.. ,. > c *j 4 ^ -V /a4c&w>- ,_ , O - *rv .*- :-^ 'W.ui 4 % V * % S> 'IV* ' s ^* v, ^ o'i , / cs'j; ^ 0 -> . .. . > a.f -j 4" 8c fc . ei . VM T* ^ ?\t ` ** / 5/ ^ *ci * x--< /y JS ^ O '* . 4\ . ^ # 6^ *, --- ---- ------ >* ?'. 't* ^?nrn 21--^ultipje layers of polyhedral angiosarcoma cells surrounding hepatocytes. Autopsy (A74-31). (H&. x400) Figure 22 Nodules comppsed of polyhedral cells with necrosis and hemorrhage {secondary peliosis)Autopsy (A74-78). (H&E, x40) Figure 23--Solid angiosarcoma above and trabecular hepatocellular carcinoma below, separated by a pseudocapsule, in autopsy specimen of case without established etiology. Autopsy (A75-17S). (H&E, v 25) Figure 24--Accumulation of hematopoietic cells within sinusoids in angiosarcoma (megakaryocytes. straight arrow, nucleated red cells, curved arrows). Autopsy (A74-281). (H&E. X250)