HECTOR - Multimodal Deep Learning Model for Predicting Recurrence Risk in Endometrial Cancer

Explore a groundbreaking 58-minute conference talk by Sarah Volinsky from Stanford University on HECTOR, a multimodal deep learning model for predicting recurrence risk in endometrial cancer. Delve into the development of this innovative tool that utilizes hematoxylin and eosin-stained whole-slide images and tumor stage data from over 2,000 patients across eight endometrial cancer cohorts. Learn how HECTOR outperforms current gold standards in prognostic accuracy, demonstrating impressive C-indices in internal and external test sets. Discover how this model can potentially revolutionize personalized treatment approaches by identifying patients with significantly different outcome probabilities. Gain insights from Volinsky's expertise in computational pathology and her ongoing PhD research at Leiden University Medical Center, where she leads the AIRMEC team in developing deep learning models for endometrial cancer prediction. Understand the broader context of AI applications in medicine through this MedAI Group Exchange Session, part of a weekly series fostering critical examination and discussion of key topics at the intersection of artificial intelligence and healthcare.