Foundation models enable weakly supervised Nancy Index scoring for ulcerative colitis

Researchers have developed a weakly supervised multiple instance learning approach for automated scoring of ulcerative colitis activity using foundation models. This method leverages case- and slide-level labels to predict the five-grade Nancy histological index, addressing the time-consuming and variable nature of manual grading. The study evaluated various foundation models on a multicenter dataset, finding that Virchow2 performed well and that ensembling improved prediction accuracy. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Automates histology scoring for ulcerative colitis, potentially improving clinical trial efficiency and diagnostic consistency.

RANK_REASON Academic paper published on arXiv detailing a new weakly supervised method for medical image analysis.

Read on arXiv cs.CV →

• Nancy Index
• Ulcerative Colitis
• Foundation Models
• Virchow2
• H&E
• arXiv

• paper
• other