The prevalence of competence standards for the optical diagnosis of diminutive colorectal lesions (SODA competence standards) is unknown. We aimed to define the knowledge and adoption of these criteria and the role of artificial intelligence (AI) systems in this setting.  

We conducted a survey during a nationwide event on quality in colonoscopy held in Verona, Italy in May 2025 by a self-administered questionnaire. Descriptive analysis was conducted.

50 out of 58 endoscopists (28/50 F/M) attended the survey, with mean age of 52 years and equal distribution of age, regions and public or private practice. Most of participants (82%) perform colonoscopy within a specific program for screening of colorectal cancer. Over 90% stated real-time in-vivo diagnosis of diminutive colorectal polyps, usually by virtual chromoendoscopy, respectively with Narrow Band Imaging (NBI, 73%), i-scan (14%) and Blue Light Imaging (BLI, 13%). The competence is mainly based as self-training.

82% reports to monitor diagnostic accuracy and 77% knows SODA criteria; however, most of them is unable to define its own sensitivity and sensibility in clinical practice.  AI availability was confirmed in less than 30%; its role is expected to improve future monitoring of appropriateness and quality standards of colonoscopy.  

We found a general lack of adoption of SODA competence standards. Acquiring sensitivity and specificity in the optical diagnosis commonly performed in this setting should be the first step towards improving “resect and discard” or “leave in-situ” strategies in clinical practice.