Advances in colon cancer prognostics reveal that integrating machine learning with traditional TNM staging significantly improves survival predictions. In a study of over 382,000 diagnosed cases, models utilizing additional clinical variables demonstrated enhanced accuracy, with a 4-year Brier score drop from 0.19 to 0.14, and a Harrell concordance index increase from 0.73 to 0.83. These findings indicate that machine learning can refine individualized patient outcomes far beyond existing staging methods.
Journal Article by Cotler JH, Janczewski LM (…) Boughey JC et 6 al. in BMC Surg
Copyright © 2025 Elsevier Inc. All rights reserved.