Microsatellite pathologic score does not efficiently identify high microsatellite instability in colorectal serrated adenocarcinoma

Abstract

The microsatellite pathologic score has been proposed as a valuable tool to estimate the probability of a colorectal cancer having high microsatellite instability; however, this score has not been tested in serrated adenocarcinoma. Our aim was to evaluate microsatellite pathologic score in serrated adenocarcinoma, conventional carcinoma, and colorectal cancer with high microsatellite instability histologic features. Eighty-nine serrated adenocarcinoma and 81 matched conventional carcinomas were tested with microsatellite pathologic score, and the results were compared with those of 24 high microsatellite instability histologic features. Validation was performed by microsatellite instability analysis. Although all colorectal cancers with high microsatellite instability histologic features rendered a more than 5.5 score, the microsatellite pathologic score performance was of lower rank in high microsatellite instability serrated adenocarcinoma because none of the cases scored above 5.5 (>77% probability of being high microsatellite instability). High microsatellite instability serrated adenocarcinoma shows pathologic features different from those observed in high microsatellite instability histologic features such as adverse prognostic histologic features at the invasive front. We describe a serrated adenocarcinoma subtype showing high microsatellite instability and some, but not all, high microsatellite instability histologic features that would not be detected if the microsatellite pathologic score cutoff is set at the highest rank. To increase microsatellite pathologic score sensitivity in serrated adenocarcinoma, we propose to set up a 2.1 cutoff score when faced by a right-sided colorectal cancer with serrated features.