Prediction of Rectal Cancer Complete Response to Chemoradiotherapy by Multiscale Imaging and Deep Learning Analyses (PRECORMIDEL) [poster]

Abstract

THE CLINICAL PROBLEM •Rectal cancer (RC) is a major contributor to colorectal cancer burden. For patients with locally advanced RC, neoadjuvant chemoradiotherapy (nCRT) is standard, yet only ~15% achieve a sustained clinical response [1]. •Response to nCRT varies greatly between individuals, biomarkers have shown limited reliability due to tumor heterogeneity [2], and predictions still rely mainly on stage and grade. •Accurate response prediction is essential to safely implement watch-and-wait strategies and reduce the risk of undertreatment and tumor regrowth [3].