FINAL DIAGNOSIS: CLEAR CELL PANCREATIC ADENOCARCINOMA
Adenocarcinoma of ductal origin compromises 85-90% of all pancreatic tumors (1,2). The WHO recognizes histomorphologic variants of ductal adenocarcinoma including mucinous non-cystic carcinoma, signet-ring carcinoma, adenosquamous carcinoma, undifferentiated (anaplastic) carcinoma, undifferentiated carcinoma with osteoclast-like giant cells, and mixed ductal-endocrine carcinoma (1). These variants are in addition to the classical tubuloglandular ductal adenocarcinoma. Investigators have alluded to both well and moderately differentiated classical ductal adenocarcinomas as occasionally having pale or even clear cytoplasm (1,2,3). Currently, the WHO recognizes clear cell carcinoma as a 'miscellaneous' carcinoma of the pancreas, which characteristically is rich in glycogen and poor in mucin although adenocarcinomatous components, ductal phenotype, and presence of mutations in the k-ras oncogene are recognized (1). Unfortunately, little data is available in these monographs and in their references regarding the incidence and extent of clear cell morphology, growth pattern, and immunohistochemical profile of these clear cell tumors. A handful of cases have been reported as 'clear cell carcinomas of the pancreas' in the setting of anaplastic carcinoma, pleomorphic giant cell carcinoma, solid pseudopapillary tumor, and in a single case of primary intraductal carcinoma with invasive clear cell component (4,5,6,7). Therefore, although well recognized elsewhere in the gastrointestinal tract, clear cell variant of classical ductal adenocarcinoma remains an understated diagnosis in the pancreas (8). We report a unique case of infiltrating pancreatic ductal adenocarcinoma with predominantly clear cell morphology (more than 95% clear cells).
The differential diagnosis of clear cell lesions in the pancreas includes pancreatic neuroendocrine tumors, solid-pseudopapillary tumor, PEComa (sugar tumor), mixed ductal-endocrine carcinoma, ductal adenocarcinoma with clear cell features, solid serous cystadenoma, and metastatic renal cell carcinoma (RCC)(1-3,9-11). The possibility of a neuroendocrine tumor and a mixed ductal-endocrine carcinoma is excluded in this case by the absence of neuroendocrine markers in the neoplastic cells. The reported case does have a few features common to solid-pseudopapillary tumor in that it occurred in the tail of the pancreas, was associated with abundant alpha-1-antitrypsin production, and showed focal NSE positivity (10). However, the absence of solid/cystic pseudopapillary growth pattern, vimentin and CD10 expression, and expression of CEA in the neoplastic cells are more in keeping with ductal adenocarcinoma. PEComa (sugar tumor) occurs rarely in the pancreas, has a solid growth pattern, and is immunohistochemically characterized by positivity for HMB-45, which is negative in our case. Likewise, the tubuloglandular growth pattern and mucin production combined with the absence of vimentin and CD10 expression excludes metastatic RCC.
The clear cells of the reported lesion express epithelial markers (CK7, CK20, and CAM5.2) in the absence of positivity for neuroendocrine markers (synaptophysin and chromogranin) and vimentin. This immunoprofile along with strong positivity for monoclonal CEA in the tumor cells strongly supports a ductal origin for this tumor. Furthermore, the tumor arises in a background of extensive pancreatic intraepithelial neoplasia (PanIn). In multiple foci of PanIn III a distinct transition from eosinophilic to clear cell morphology is seen within a single ductular unit and demonstrates that the in situ component of this lesion is partially composed of clear cells. Finally, molecular evidence of ductal origin is confirmed by sequencing the k-ras oncogene from microdissected neoplastic glands. We therefore feel that the lesion reported here is best described as a moderately differentiated clear cell ductal adenocarcinoma.
An additional outstanding feature of this case is the presence of multiple dense hyaline globules found both intraluminally and intracytoplasmically. These globules were histochemically positive for PASD and immunohistochemically positive for alpha-1-antitrypsin. Alpha-1-antitrypsin production is well documented in pancreatic tumors such as solid-pseudopapillary tumor as well as in extrapancreatic tumors, most notably hepatocellular carcinoma (1-2,10,12-13). Thus, while the expression of alpha-1-antitrypsin is in no way specific for ductal adenocarcinoma, the dense hyaline globules seen in this case add to its unique histomorphology.
Clear cell morphology in ductal pancreatic adenocarcinoma has been reported in the past (1-3). In their discussion of ductal adenocarcinoma of the pancreas the WHO recognize clear cell changes in both well and moderately differentiated pancreatic adenocarcinomas (1). Likewise, in their discussion of well-differentiated pancreatic adenocarcinomas Solcia et al recognize variants with clear cytoplasm (2). However, in both of the previously mentioned monographs data from specific cases regarding the extent and incidence of clear cell change, overall growth pattern, and immunohistochemical features are not provided.
Although clear cell features have been described in the context of anaplastic carcinoma, pleomorphic giant cell carcinoma, solid pseudopapillary tumor, and in a single case of primary intraductal carcinoma with invasive clear cell component, there is a lack of reported cases in the literature for a conventional ductal adenocarcinoma composed of predominantly clear cells (1-8,14). Also, due to the paucity of data, it is difficult to make predictions about the prognostic significance of clear cell differentiation in ductal adenocarcinomas.
In conclusion, we present a unique case of moderately differentiated clear cell ductal adenocarcinoma of the pancreas. Abundant intracytoplasmic and intraluminal mucin production along with features such as dense desmoplasia and multifocal perineural invasion closely mimic the growth and spread of a conventional moderately differentiated ductal adenocarcinoma. An additional outstanding feature of this case is the presence of dense alpha-1-antitrypsin globules located both intracytoplasmically and intraluminally within neoplastic glands and focal NSE positivity that more commonly occurs in solid-pseudopapillary tumors of pancreas. In the proper morphologic context supported by immunohistochemistry clear cell carcinoma may be regarded as a rare variant of ductal adenocarcinoma. Further investigation and follow up is required to determine the implications, if any, of clear cell differentiation in these tumors.
Contributed by Kenneth Clark, MD and Sourav Ray, MD