Evaluation of a real-time computer-aided polyp detection system during screening colonoscopy: AI-DETECT study

Area of Science:

• Gastroenterology
• Medical Imaging
• Cancer Screening

Background:

• Colorectal cancer (CRC) risk is reduced by polyp detection and resection during colonoscopy.
• Computer-aided detection (CADe) systems aim to improve polyp identification but require further clinical validation.
• This study evaluated CADe effectiveness within a national bowel cancer screening program (BCSP).

Purpose of the Study:

• To assess the impact of CADe on polyp detection rates during colonoscopies in a BCSP setting.
• To compare the performance of CADe-assisted colonoscopy versus standard colonoscopy.
• To determine if CADe offers additional benefits in a high-volume screening environment.

Main Methods:

• A prospective, randomized controlled trial was conducted at an English NHS BCSP center.
• Screening-accredited colonoscopists performed colonoscopies with either CADe or standard techniques.
• The primary outcome was the polyp detection rate (PDR); adenoma detection rate (ADR) was also analyzed.

Main Results:

• In intention-to-treat analysis, CADe showed a borderline significant increase in PDR (85.7% vs. 79.7%, P=0.05) but no significant difference in ADR (71.4% vs. 65.0%, P=0.09).
• Per-protocol analysis revealed no significant differences in PDR or ADR between CADe and standard colonoscopy groups.
• Procedure times did not differ significantly between the two groups.

Conclusions:

• CADe demonstrated a marginal improvement in PDR but not ADR among experienced colonoscopists in a BCSP, particularly when using adjuncts like Endocuff Vision.
• The clinical benefit of CADe appears limited in BCSP settings with high-performing endoscopists.
• Further research may be needed to identify specific patient populations or procedural contexts where CADe provides substantial advantages.