Colon-X: Advancing Intelligent Colonoscopy toward Clinical Reasoning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ji, Ge-Peng, Liu, Jingyi, Fan, Deng-Ping, Fu, Huazhu, Barnes, Nick
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915860839399424
author Ji, Ge-Peng
Liu, Jingyi
Fan, Deng-Ping
Fu, Huazhu
Barnes, Nick
author_facet Ji, Ge-Peng
Liu, Jingyi
Fan, Deng-Ping
Fu, Huazhu
Barnes, Nick
contents In this study, we present Colon-X, an open initiative aimed at advancing multimodal intelligence in colonoscopy. We begin by constructing ColonVQA, the most comprehensive multimodal dataset ever built for colonoscopy, featuring over 1.1M+ visual question answering entries across 76 clinical findings and 18 multimodal tasks. Beyond serving as a community-wide data foundation, we further investigate a critical yet underexplored transition in colonoscopy - evolving from multimodal understanding to clinical reasoning: (a) To capture the current landscape of multimodal understanding behaviors, we systematically assess the generalizability of 22 multimodal large language models and examine their reliability under human-induced perturbations. The results reveal that clinical outputs from leading MLLMs remain far from robust and trustworthy. (b) To narrow this gap, we further explore reasoning-centric intelligence tailored for colonoscopy. Specifically, we curate ColonReason, a clinically grounded reasoning dataset annotated through a multi-agent debating pipeline, and develop ColonR1, the first R1-styled model that mitigates reward information collapse through task-adaptive rewards and gradient-stable policy optimization. Under data-scarce conditions, our ColonR1 achieves 56.61% overall accuracy, outperforming supervised fine-tuning by 25.22%, and sets a new reasoning-enabled baseline for multimodal colonoscopy analysis. All data and model resources are publicly available at https://github.com/ai4colonoscopy/Colon-X.
format Preprint
id arxiv_https___arxiv_org_abs_2512_03667
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Colon-X: Advancing Intelligent Colonoscopy toward Clinical Reasoning
Ji, Ge-Peng
Liu, Jingyi
Fan, Deng-Ping
Fu, Huazhu
Barnes, Nick
Computer Vision and Pattern Recognition
In this study, we present Colon-X, an open initiative aimed at advancing multimodal intelligence in colonoscopy. We begin by constructing ColonVQA, the most comprehensive multimodal dataset ever built for colonoscopy, featuring over 1.1M+ visual question answering entries across 76 clinical findings and 18 multimodal tasks. Beyond serving as a community-wide data foundation, we further investigate a critical yet underexplored transition in colonoscopy - evolving from multimodal understanding to clinical reasoning: (a) To capture the current landscape of multimodal understanding behaviors, we systematically assess the generalizability of 22 multimodal large language models and examine their reliability under human-induced perturbations. The results reveal that clinical outputs from leading MLLMs remain far from robust and trustworthy. (b) To narrow this gap, we further explore reasoning-centric intelligence tailored for colonoscopy. Specifically, we curate ColonReason, a clinically grounded reasoning dataset annotated through a multi-agent debating pipeline, and develop ColonR1, the first R1-styled model that mitigates reward information collapse through task-adaptive rewards and gradient-stable policy optimization. Under data-scarce conditions, our ColonR1 achieves 56.61% overall accuracy, outperforming supervised fine-tuning by 25.22%, and sets a new reasoning-enabled baseline for multimodal colonoscopy analysis. All data and model resources are publicly available at https://github.com/ai4colonoscopy/Colon-X.
title Colon-X: Advancing Intelligent Colonoscopy toward Clinical Reasoning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.03667