Interobserver Agreement in Endometrial Carcinoma Histotype Diagnosis Varies Depending on The Cancer Genome Atlas (TCGA)-based Molecular Subgroup.Am J Surg Pathol. 2017 Feb; 41(2):245-252.AJ
The Cancer Genome Atlas recently identified a genomic-based molecular classification of endometrial carcinomas, with 4 molecular categories: (1) ultramutated (polymerase epsilon [POLE] mutated), (2) hypermutated (microsatellite instability), (3) copy number abnormalities-low, and (4) copy number abnormalities-high. Two studies have since proposed models to classify endometrial carcinomas into 4 molecular subgroups, modeled after The Cancer Genome Atlas, using simplified and more clinically applicable surrogate methodologies. In our study, 151 endometrial carcinomas were molecularly categorized using sequencing for the exonuclease domain mutations (EDM) of POLE, and immunohistochemistry for p53 and mismatch repair (MMR) proteins. This separated cases into 1 of 4 groups: (1) POLE EDM, (2) MMR-D, (3) p53 wildtype (p53 wt), or (4) p53 abnormal (p53 abn). Seven gynecologic pathologists were asked to assign each case to one of the following categories: grade 1 to 2 endometrioid carcinoma (EC), grade 3 EC, mucinous, serous carcinoma (SC), clear cell, dedifferentiated, carcinosarcoma, mixed, and other. Consensus diagnosis among all 7 pathologists was highest in the p53 wt group (37/41, 90%), lowest in the p53 abn group (14/36, 39%), and intermediate in the POLE EDM (22/34, 65%) and MMR-D groups (23/40, 58%). Although the majority of p53 wt endometrial carcinomas are grade 1 to 2 EC (sensitivity: 90%), fewer than half of grade 1 to 2 EC fell into the p53 wt category (positive predictive value: 42%). Pure SC almost always resided in the p53 abn group (positive predictive value: 96%), but it was insensitive as a marker of p53 abn (sensitivity 64%) and the reproducibility of diagnosing SC was suboptimal. The limitations in the precise histologic classification of endometrial carcinomas highlights the importance of an ancillary molecular-based classification scheme.