frontiersin.org
Artificial Intelligence Assisted Capsule endoscopy for Detection Lesions of Crohn's disease
AI-assisted capsule endoscopy may improve detection of Crohn’s disease lesions and reduce missed findings, which could affect diagnosis and monitoring. People with IBD may see faster or more consistent reads of capsule studies if AI proves reliable in larger clinical trials.
Clinicians who use capsule endoscopy, researchers developing AI tools for endoscopy, and patients interested in diagnostic advances for Crohn’s disease.
What To Know
What to know This article is a systematic review and meta-analysis evaluating how well artificial intelligence (AI), mainly deep-learning models, detects Crohn’s disease lesions on capsule endoscopy images.
The pooled results reported high diagnostic accuracy across included studies and models, with authors noting heterogeneity related to algorithm type, sample size, and study design.
The review suggests AI could help reduce missed lesions and shorten capsule reading time, potentially supporting less-experienced readers and increasing the clinical utility of capsule endoscopy for small-bowel Crohn’s disease.
However, the authors call for more high-quality, larger studies before AI tools can be widely adopted and optimized in routine clinical practice. This paper is a research synthesis (meta-analysis) rather than a clinical guideline or new device approval; it summarizes existing studies rather than reporting a single new commercial product.
This is a meta-analysis of primarily research studies and AI models up to May 2024; pooled accuracy estimates may reflect study-level heterogeneity. The review highlights promising results but notes limitations and calls for larger, higher-quality prospective evaluations before routine clinical adoption.
