A Novel Clinical Decision Support Tool to Predict Submucosal Invasive Cancer Within Large Non-Pedunculated Colorectal Polyps

Colorectal cancer (CRC) is a leading cause of death in the Western world. Large non-pedunculated colorectal polyps (LNPCPs, \>=20mm) represent 2-3% of colorectal polyps and require special attention. The presence of submucosal invasive cancer (SMI) determines whether endoscopic treatment is curative or whether referral for surgery is necessary. Current endoscopic classification systems (JNET, Kudo, NICE) are complex, require specialized equipment, and extensive training, making them underused in non-tertiary practice.

A novel, freely accessible web-based clinical decision support tool (available at gieqs.com/SMI) was developed based on established risk factors for SMI: the presence and characteristics of a demarcated area within the polyp, polyp size (\>=40mm), Paris classification, colonic location, and granularity. The tool uses only standard white-light and virtual chromoendoscopy available in the majority of Western endoscopy units.

This is a prospective, multi-centre study conducted across 8 centres in Belgium, the United Kingdom, Israel, and Italy. The study evaluates whether endoscopists of varying experience can accurately detect SMI within LNPCPs during live colonoscopy using this tool.

The study follows a pre-post design with randomization of the educational intervention:

* Phase 1 (Pre-training): Each endoscopist assesses 10 eligible LNPCPs using a standardized assessment and creates a 1-minute standardized video of each lesion. * Educational Intervention: Endoscopists are randomly allocated (1:1) to either (a) a 10-minute non-dynamic instructional video or (b) a 45-minute interactive educational session. * Phase 2 (Post-training): Each endoscopist assesses a further 10 LNPCPs using the clinical decision support tool. * Phase 3 (Durability): At 3 months, a further 10 assessments are performed to evaluate durability of the training effect.

Expert opinion applied to standardized videos serves as the reference standard. Histopathology results are recorded for all resected specimens. An interim analysis is planned when half the required participants are enrolled.

• Percentage of LNPCP Assessments in Which a Demarcated Area Is Correctly Detected Using the Clinical Decision Support Tool, Compared to Expert Opinion [Time frame: Through study completion, up to 48 months]
• Inter-observer Agreement Among Endoscopists in Detection of a Demarcated Area Within LNPCPs, Measured by Fleiss' Kappa [Time frame: Through study completion, up to 48 months]
• Percentage of Demarcated Areas in Which the JNET Classification Is Correctly Determined, Compared to Expert Opinion [Time frame: Through study completion, up to 48 months]
• Percentage of Lesions in Which Each Morphological Component (Size, Location, Granularity, Paris Classification) Is Correctly Determined, Compared to Expert Opinion [Time frame: Through study completion, up to 48 months]
• Correct Treatment Decision Based Upon the Tool Score (Rate of Unnecessary Surgery for Benign Disease or Unnecessary Endoscopic Treatment Requiring Subsequent Surgery) [Time frame: Through study completion, up to 48 months]
• Endoscopist-Reported Utility of the Clinical Decision Support Tool, Measured by a 5-point Likert-Scale Questionnaire [Time frame: Through study completion, up to 48 months]
• Difference in Scoring Accuracy Before and After the Educational Intervention, Comparing a 10-minute Instructional Video Versus a 45-minute Interactive Session [Time frame: Through study completion, up to 48 months]
• Percentage Accuracy of SMI Detection Using the Clinical Decision Support Tool at 3 Months Post-Intervention [Time frame: 3 months after educational intervention]