Capsule Endoscopy Review Times Cut With AI
Capsule Endoscopy Review Times Cut With AI
Publish Date: 2026-05-23 09:01:00
Source Domain: www.emjreviews.com
- Improved Lesion Detection: An open-source AI model named SEE-AI significantly improved lesion detection during capsule endoscopy, achieving higher sensitivity compared to conventional reading methods.
- Reduced Interpretation Times: Use of SEE-AI decreased the mean reading time from 17.9 minutes to 13.7 minutes during the retrospective analysis of 249 examinations across six hospitals.
- Enhanced Sensitivity: The per-lesion sensitivity increased from 86.4% with standard interpretation to 98.8% with SEE-AI-assisted reading, and per-patient sensitivity improved from 80.3% to 99.1%.
- Benefits for Specific Patient Groups: SEE-AI particularly improved detection and reduced reading times in patients with suspected small-bowel bleeding, enhancing sensitivity to 98.2% per lesion and 98.6% per patient.
- Open-Source Model Support: SEE-AI remains open-source, promoting transparency and reproducibility in AI applications for gastrointestinal imaging.
- Potential for Future Standard of Care: Researchers believe SEE-AI holds strong practical potential and could establish AI-assisted capsule endoscopy as a future standard for small bowel imaging.
- Need for Further Investigation: Prospective studies are needed to validate the AI model’s performance in routine clinical practice across diverse patient populations.
- Support for Reduced Clinician Workload: Integrating AI into capsule endoscopy workflows may streamline processes and reduce the workload for clinicians, while improving diagnostic accuracy.
