A Study on AI-Assisted Colonoscopy for Better Polyp Detection
Overview
Detecting and removing colon polyps during colonoscopy is vital for preventing colorectal cancer. This study focused on creating an AI program to improve polyp detection using the RetinaNet algorithm.
Study Details
The AI program was developed using a large dataset that was anonymized and split for testing. It included:
- 9,639 training images
- 1,070 validation images
- Video data from 56 patients
The final model works in real-time during endoscopy. A randomized controlled trial was conducted across six institutions with 805 patients to compare polyp detection rates (PDR).
Key Findings
The group using the AI model had:
- Higher PDR
- Higher adenoma detection rate (ADR)
Analysis showed an odds ratio (OR) of 1.50 for detecting polyps with AI assistance. This indicates that the AI program significantly improves polyp detection during colonoscopy, aiding in cancer prevention.
Practical Solutions and Value
Clinical trials are essential for developing safe treatments. Our AI platform, DocSym, integrates ICD-11 standards and clinical research into an accessible resource for healthcare providers.
In today’s healthcare landscape, efficiency is key. Our mobile apps facilitate:
- Scheduling
- Monitoring treatments
- Telemedicine services
This digital approach enhances patient care and reduces paperwork. Discover how we can assist you at aidevmd.com.