Real-world utility of Histotype Px colorectal as a prognostic and potentially predictive biomarker.

215 Background: The decision to offer adjuvant chemotherapy (ACT) for patients with high-risk stage II and stage III colon cancer is fraught with challenges as most patients do not benefit from potentially toxic therapy. Histotype Px Colorectal stratifies patients into distinct risk groups by combining staging parameters with the DoMore v1 marker, a novel artificial intelligence-based digital biomarker that analyzes routine H&E-stained FFPE whole-slide images. The purpose of this study was to externally validate its prognostic performance and investigate its potential to predict adjuvant chemotherapy (ACT) benefit. Methods: This was a retrospective analysis of patients diagnosed with pathological stage II/III colon adenocarcinoma treated at Ohio State University in 2011-2024. Clinical parameters and outcome data were extracted from patient records. Blinded to clinical outcomes, anonymized digitized slides were analyzed by Histotype Px. Logrank test and Cox proportional hazards regression were used to analyze cancer-specific death (CSD). The multivariable models included age, sex, pN stage, pT stage, number of lymph nodes sampled, tumor perforation, lymphovascular invasion, perineural invasion, MSI status, and receipt of ACT. Competing risk analysis was used to calculate CSD rates. Results: Baseline characteristics of the 503 eligible patients included median age of 63 years (22-97), 54% male, 55% stage III, 61% right-sided tumors and 23% MSI, with 51% of patients receiving ACT and a median follow-up of 44 months. The DoMore v1 marker was statistically significant in both univariable (p