Machine learning model effectively predicts postoperative recurrence in duodenal adenocarcinoma

A machine learning-based model has been developed to predict postoperative recurrence in patients with duodenal adenocarcinoma, utilizing data from 1,830 patients who underwent radical surgery. Among 100 predictive models tested, the penalized regression + accelerated oblique random survival forest model (pam) delivered the highest performance with a concordance index (c-index) of 0.882 in the training cohort. This tool aims to assist clinicians in evaluating disease severity and planning follow-up and adjuvant treatments.