Adaptive reasoning improves medical image diagnosis for inflammatory bowel disease

ARISE: Adaptive Agentic Reasoning with Image-grounded Self-Evaluation for Interpretable IBD Assessment

Computer Vision and Pattern RecognitionArtificial Intelligence

Summary

Inflammatory bowel disease is a condition requiring frequent image tests that can be hard to interpret without expert doctors. The authors created a new step-by-step approach called ARISE that helps computer models better understand medical images by thinking through the problem in stages and checking their own reasoning. They tested ARISE on two types of images from patients and found it made more accurate and understandable diagnoses than existing methods. This approach could help doctors make better decisions by showing how the computer reached its conclusions.

What this means in practice

  • For hospital radiology teams: Use ARISE to enhance diagnostic accuracy and transparency in interpreting capsule endoscopy and ultrasound images for inflammatory bowel disease.
  • For medical imaging software developers: Integrate ARISE's multi-stage reasoning framework to improve interpretability and reliability of AI tools analyzing complex gastrointestinal images.$Commercial implications: Enables development of clinical-grade diagnostic software offering detailed reasoning steps for IBD assessment, appealing to healthcare providers.

Authors

Pronoma Banerjee, Anuva Shah, Jason Wu, Md. Masudur Rahman, Sanjay Mohanty, Satya Kurada, Juan P. Wachs

Abstract

Inflammatory bowel disease (IBD) requires frequent imaging-based assessment, yet interpretation of modalities such as wireless capsule endoscopy (WCE) and intestinal ultrasound remains heavily dependent on specialist expertise. Vision-Language Models (VLMs) demonstrate significant potential in multimodal medical image analysis, but their clinical adoption is hindered by their insufficient domain-specific reasoning, susceptibility to hallucination, scarcity of high quality training data in fine-grained diagnostics and limited interpretability. We introduce ARISE (Adaptive Agentic Reasoning with Image-grounded Self-Evaluation), an autonomous planning framework that models few-shot medical image understanding as a sequential agentic workflow. ARISE structures agent execution into a transparent 5-stage workflow: hypothesis generation, image-grounded evidence summarization, evidence-conditioned refinement, symbolic verification, and final diagnosis. We apply ARISE to IBD assessment across two independent patient cohorts: wireless capsule endoscopy (WCE) images for Crohn's disease and B-mode ultrasound data for ulcerative colitis. ARISE consistently improves diagnostic performance over baseline VLMs while exposing where reasoning succeeds or fails, providing a more interpretable basis for clinical decision support and realistic deployment.