To CADe or not to CADe - evidence for the benefit of CAD for increased yield of colonoscopy

Computer -Aided Detection (CADe) has emerged as a promising tool for enhancing colonoscopy during colorectal cancer screening. This manuscript critically examines the evidence supporting CADe's role in improving the yield of colonoscopy, focusing on its impact on lesion detection, miss rates, and real-time decision-making, especially for less experienced endoscopists. We review relevant clinical trials, studies, and meta-analyses to provide a comprehensive assessment of CADe's impact on detection rates, diagnostic accuracy, cost-effectiveness, colonoscopy training, reimbursement and patient outcomes. Despite challenges such as the need for adequate bowel preparation, a degree of false-positive rate, and potential lower sensitivity in detecting sessile serrated lesions (SSL), CADe holds significant promise for improving colonoscopy outcomes by enabling constant detection of colorectal neoplasia and contributing to lower adenoma miss rate. This review also covers the factors influencing the decision to use or not use CADe in various clinical contexts. Further development of AI-powered tools can potentially extend the benefits of CADe by enabling polyp characterization and automated quality control during colonoscopy, addressing factors such as withdrawal time and bowel preparation adequacy.