Hepatocholangiocarcinoma: The Great Masquerader
The second presentation, by Dr. Rodney Miller, was based on the collaborative work of Drs. Rodney Miller, Rebecca Moore, and Gabrielle Willson and was titled “The Great Masquerader.” Dr. Miller’s discussion emphasized the various histomorphology patterns of hepatocholangiocarcinoma (HCCC) and thus the potential for diagnostic challenges.
After presentation of numerous images demonstrating the highly variable morphological patterns seen in the primary and metastatic lesions, a vote was taken, given the following diagnostic choices: (1) epithelioid hemangioendothelioma, (2) mesothelioma, (3) HCCC, (4) hepatocellular carcinoma, (5) atriocaval mesothelioma, and (6) hepatoblastoma. Hepatocholangiocarcinoma was the correct response and recorded a 57% vote. The other votes recorded were: epithelioid hemangioendothelioma (7%), mesothelioma (18%), hepatocellular carcinoma (2%), atriocaval mesothelioma (4%), and hepatoblastoma (12%). After the vote, a detailed presentation of HCCCs in the NTP database was given.
Hepatocholangiocarcinoma is a rare tumor with an overall incidence of less than 1% in male and female mice in the current NTP database. Hepatocholangiocarcinoma is defined as a combination of neoplastic hepatocytes and neoplastic biliary epithelial cells. In addition to the proliferative hepatocellular and biliary epithelial components, there may be a poorly differentiated cellular component that has a sarcomatous, anaplastic, and undifferentiated appearance. The variable histological morphology of the primary HCCC and the metastatic lesions may pose a diagnostic challenge. The purpose of this research and presentation was to describe the morphological features, incidence, and behavior of HCCC in B6C3F1 mice. The NTP historical database from 1982 to 2008 was reviewed for a diagnosis of HCCC in B6C3F1 mice (retrieved August 27, 2009, from the National Toxicology Program Web site: http://ntp-server.niehs.nih.gov). A total of 164 HCCCs (ninety-five male and sixty-nine female mice) from 74 two-year carcinogenicity studies were identified and reviewed.
Hepatocholangiocarcinoma occurred more commonly in males (58%) than in females (42%). Metastases were evident in 138 animals (84%) and often occurred in multiple sites. The sites most frequently observed to have metastatic lesions included lung (125/138, or 91%), mediastinum (104/138, or 75%), mesentery (80/138, or 58%), lymph nodes (71/138, or 51%), skeletal muscle (55/138, or 40%), kidney (47/138, or 34%), heart (46/ 138, or 33%), and pancreas (10/138, or 7%). Hepatocholangiocarcinomas were not considered to be treatment related in any study; however, in twenty-four of the seventy-four studies (32%), there was a concomitant treatment-related increase in hepatocellular tumors in one or both sexes. Although most of the NTP mouse control groups from this period did not have a diagnosis of HCCC, the highest incidence of spontaneous HCCC in any control group was as high as two of forty-nine for female controls and five of fifty for male controls.
All HCCCs reviewed in this study had a malignant hepatocellular component, as required by definition. The vast majority of HCCCs had small to large areas of necrosis and/or small to large areas of cystic spaces. The cystic areas had ill-defined epithelial linings or were partially lined by a flattened to cuboidal epithelium. Foci of proliferating epithelial cells forming ductal structures, representing the malignant biliary component, were often evident in close proximity to these areas of necrosis, or the cystic areas. Not infrequently, these foci were small and difficult to appreciate upon cursory examination (Figure 2A). Occasionally (about 16%), there were focal areas of proliferation of undifferentiated round to spindle cells in a loose fibrillar matrix within or adjacent to the malignant hepatocellular component (Figure 2B). It was not unusual to have transformation of neoplastic hepatocytes in the same proliferating hepatic cord into varieties of columnar epithelial cells to form ducts (Figure 2C).
The diagnostic challenge often began with the metastatic lesions. The metastatic lesions were often grossly visible, multifocal, and tan and occurred within the parenchyma of the tissue bearing the metastasis. The metastatic lesions also commonly extended to and occurred on serosal surfaces, where they stimulated serosal lining cells to react. If classical malignant hepatocellular components were present in the metastatic lesions along with the neoplastic duct-forming cells or the undifferentiated or spindle cells (Figures 2D and 2E), it stimulated the pathologist to return to the liver to confirm the presence of HCCC.
Commonly, the metastatic lesions contained only epithelial cells of the malignant biliary component (as was often the case in mesenteric metastases), an undifferentiated component, or sometimes a combination of both. Less commonly, the metastatic lesions contained the malignant hepatocellular component. The undifferentiated component was composed of nondescript round cells or spindle cells resembling a sarcoma. When this sometimes puzzling, but virtually pathognomonic, pattern of metastasis is seen, it often leaves the pathologist wondering whether the primary tumor is a carcinoma, a sarcoma, a mesothelioma, or a neoplasm of unknown histogenesis. The pathologist should then be stimulated to review the primary liver mass and evaluate it closely, looking for the diagnostic trilogy that leads to a diagnosis of HCCC: the malignant hepatocellular component, areas of necrosis or cystic degeneration, and the ducts of the malignant biliary component. The presence of a focal and undifferentiated or spindle cell proliferation in the liver mass adds even more strength to the argument and helps confirm that the primary tumor should be called an HCCC.
In summary, HCCC is a rare tumor in mice and is an aggressive tumor that metastasizes readily. The HCCC metastatic rate of 84% is much higher than that of hepatocellular carcinomas (about 23%) or hepatoblastomas (about 23%) (Turusov et al. 2002). Hepatocholangiocarcinoma was more commonly observed in male mice than in female mice in these NTP studies. Although it has been reported to be induced by chemicals such as benzidine dihydrochloride and N-2-acetoaminofluorene (Frith, Ward, and Turusov 1994), treatment-related HCCCs were not observed in the NTP dataset. Hepatocholangiocarcinoma can have a highly variable histological morphology, and careful evaluation of all features of the primary and metastatic lesions must be taken into account to render the correct diagnosis. Careful consideration must be given to hepatocellular proliferation, biliary epithelial proliferation, undifferentiated cell proliferation, cystic areas, and necrotic areas in the primary liver tumor and put into context with any combination of hepatocellular, biliary or undifferentiated cell proliferation in the metastatic lesions.