The COMBO CAD Study: Characterization cOMparison Between twO CAD Systems

Implementation of clinical strategies based on optical diagnosis of <5 mm colorectal polyps may lead to a substantial saving of economic and financial resources. Despite this, 84.2% of European endoscopists reported not to use such strategies - also named as leave-in situ and resect- and-discard - in their practice due to the fear of an incorrect optical diagnosis.

Indeed, accuracy of optical diagnosis is operator-dependent, and values reported in the community setting are below the safety thresholds proposed for its incorporation in clinical practice.

Artificial intelligence (AI) is being increasingly explored in different domains of medicine, particularly those entailing image analysis. As optical diagnosis involves subitaneous processing of multiple images, searching for specific visual clues, and recognizing well-defined visual patterns, AI systems has the potential to help endoscopists in distinguish neoplastic from non-neoplastic polyps, making the characterization process more reliable and objective. Computer-Aided-Diagnosis systems aiming at characterization are called CADx.

Preliminary data on CADx showed a high feasibility and accuracy of AI for optical diagnosis of colorectal polyp, and initial experiences in clinical practice confirmed preliminary results.

To assess the potential benefit and risk of AI-assisted optical diagnosis with standard colonoscopy, we exploited two new Computer-Aided-Diagnosis systems (CAD-EYE® Fujifilm Co., and GI-Genius® Medtronic) that provide the endoscopist with a real-time polyp characterization without the need of optical magnification.