Can pathologists reliably establish ampulla of Vater carcinoma histologic subtype on H&E alone? Concordance of subtype and comparison to immunohistochemistry-established subtype

Abstract

AbstractHistologic subtyping of ampulla of Vater carcinoma (ampullary carcinoma, AC) may guide choice of adjuvant chemotherapy for advanced disease, as intestinal (INT) subtype may have a better treatment response and survival as compared to the pancreatobiliary (PB) and mixed (MIX) subtypes. However, ambiguous morphologic features can result in inconsistent subtyping among pathologists. Recently immunohistochemistry (IHC) for cytokeratin (CK)7, CK20, CDX2, and the Mucin (MUC) family proteins had been used to aid in the subtyping. It is important to determine if accurate subtyping can be achieved based on hematoxylin and eosin (H&E) evaluation alone, or if IHC is required. In this study, 5 gastrointestinal pathologists evaluated 34 cases of AC for histologic subtype (PB, INT, or MIX) based on H&E stain alone. IHC studies (CK7, CK20, and CDX2) were used to determine a gold standard subtyping for comparison (strong CK7 with weak CK20 and CDX2 = PB; strong CK20 or CDX2 and weak CK7 = INT; strong CK7 and CK20 or CDX2 = MIX). Interobserver concordance (kappa) was calculated. Percent correctness of subtyping was calculated in relation to IHC-established subtype gold standard. Interobserver concordance for subtype was fair (kappa 0.31). Percent correctness for subtype against IHC-established subtype was 50% for a majority (3 or more) of pathologists. In all 11 cases with 4 or 5 pathologists being concordant, the subtype was also correctly identified. In addition, data on chemotherapy regimen and overall survival was gathered from the electronic medical record. Patients who were assigned a subtype at time of diagnosis tended to receive gemcitabine for PB subtype, and FOLFOX-based therapy for INT subtype, with some cases diagnosed as “not-otherwise specified” typically also receiving gemcitabine. Survival data were limited by loss to follow-up. In conclusion, despite 11 cases with strong concordance and correctness, overall subtyping concordance was only fair, and a majority of the 5 pathologists correctly identified the subtype in only 50% of cases. Thus, while some AC cases yield clear-cut subtyping, significant discrepancy among pathologists remains. It appears that utilization of IHC to identify subtype in AC cases is necessary to provide accurate, reliable prognostic and therapeutic information.