Morphologic and immunohistochemical characterization of small and large duct cholangiocarcinomas in a Western cohort: A panel of pCEA, CRP, N-cadherin, and albumin in situ hybridization aids in subclassification.

Two morphologic subtypes of intrahepatic cholangiocarcinoma (iCCA), small duct and large duct, are now recognized, and importantly, these subtypes are associated with distinct molecular pathways and therapeutic options. Initial studies demonstrated the feasibility of morphologic subclassification and helped characterize the immunoprofile of the subtypes. However, few studies have been undertaken in Western countries where incidence of the subtypes is likely distinct from that in the East. To address this, 48 tumors from a North American cohort, consisting of 29 iCCAs, 18 extrahepatic CCAs (eCCAs), and 1 tumor of unclear origin (liver vs. gallbladder growing into liver) were classified by morphologic criteria as large duct (19 tumors), small duct (13 tumors), or indeterminate (13 tumors; all iCCAs). Notably, only 3 iCCAs were classified as large duct. Additionally, we evaluated the utility of common biomarkers to aid in subclassification, given that a significant portion of iCCAs were challenging to classify (e.g., indeterminate morphology). Tumors were screened for expression of mucicarmine, epithelial membrane antigen (EMA), monoclonal (mCEA), polyclonal CEA (pCEA), N-cadherin, CD56, and albumin by in situ hybridization (ALB-ISH). Of these, pCEA, CRP, N-cadherin, and ALB-ISH showed statistically significant differences between large and small duct types (P < 0.0028), with high specificity (≥88 %) and at least moderate sensitivity (≥60 %). Eleven of the 13 morphologically indeterminate tumors could be classified based on their expression of these 4 markers. Four additional large duct iCCAs were subsequently obtained from a second North American institution and assessed for pCEA, N-cadherin, and albumin expression. Combining these data with the initial cohort of large duct iCCAs (total of 7 large duct iCCAs) showed similar biomarker associations. In conclusion, in this Western cohort, 55 % of iCCAs (16 of 29) could be subclassified as large or small duct type based on morphology alone. With the aid of the 4-marker panel, 93 % of iCCAs (27 of 29) could be classified. Unlike in East Asian cohorts, the vast majority of iCCAs (88 %) was small duct type, and hepatolithiasis was not observed. CRP, N-cadherin, and ALB-ISH were found to be specific for small duct iCCA, whereas diffuse, strong expression of pCEA showed specificity for large duct tumors. This is the first report to highlight the utility of pCEA to subclassify iCCAs. Additionally, in cases in which the primary site (within the biliary tract) was unclear, CRP, ALB-ISH, N-cadherin, and pCEA were helpful in distinguishing iCCA from eCCA.