Natural Language Processing for the Accurate Identification of Colorectal Cancer Mismatch Repair Status in Lynch Syndrome Screening

Abstract

Lynch syndrome (LS) is the most common type of hereditary colorectal cancer (CRC) syndrome caused by pathogenic variants in mismatch repair (MMR) genes.¹ Current multisociety guidelines recommend screening all CRC tumors for LS.^2,3 The most widely adopted screening method is MMR immunohistochemistry (IHC) followed by germline analysis if indicated.^2,3 However, the text-based nature of pathology and IHC reports used for LS screening results impedes creation of an efficient tracking system for identifying affected patients and screening outcomes.⁴ In this study, we developed and validated a natural language processing (NLP) tool for extracting MMR IHC results in LS screening in a large, diverse, multicenter, community-based setting.⁵.