AI-Assisted Optical Diagnosis (CADx) for Diminutive Colorectal Polyps

Colonoscopy with polypectomy is the cornerstone of colorectal cancer (CRC) screening and prevention. The majority of polyps detected during colonoscopy are diminutive (≤5 mm) and carry a very low risk of harbouring advanced pathology. In Canada, over 1.9 million colonoscopies are performed annually, and sending all diminutive polyps for histopathological examination (at a cost of approximately $150-$300 CAD per specimen) generates a substantial economic burden with limited benefit for cancer prevention.

Two international guideline frameworks define the performance thresholds required for clinical adoption of optical diagnosis strategies for diminutive colorectal polyps. The ASGE PIVI initiative (2011) established (1) ≥90% negative predictive value (NPV) for adenomatous histology among diminutive rectosigmoid polyps diagnosed with high confidence ("diagnose-and-leave" threshold), and (2) ≥90% agreement between optical-diagnosis-based and pathology-based post-polypectomy surveillance intervals ("resect-and-discard" threshold). The more recent ESGE SODA position statement (2022) adopted sensitivity and specificity as the foundational, prevalence-independent performance measures for high-confidence real-time characterization of colorectal neoplasia, applicable across all colonic locations: sensitivity ≥90% and specificity ≥80% for the rectosigmoid subset, and sensitivity ≥80% and specificity ≥80% across all diminutive colorectal polyps regardless of location.

The investigators' research group has previously conducted several single-centre studies evaluating early-generation, Health Canada-approved computer-aided diagnosis (CADx) systems in routine clinical practice, generating one of the largest real-world CADx implementation datasets in the literature (868 patients, 1,660 diminutive polyps). These studies consistently demonstrated suboptimal overall diagnostic accuracy (62.7%-66.4%), attributable to restricted training datasets and limited polyp-classification capabilities in early-generation devices. GI Genius (Medtronic), the newer-generation, Health Canada-approved CADx module evaluated in this study, was trained on a larger and more diverse dataset of annotated endoscopic images and videos. Whether this translates into measurable gains in diagnostic performance sufficient to meet international competence thresholds has not been established in an independent, prospective, multicentre study, and is the question this study addresses.

This is a prospective, three-centre diagnostic accuracy study conducted in accordance with the Standards for Reporting Diagnostic Accuracy Studies (STARD 2015). It is a single-arm study validating a CADx-assisted optical-diagnosis strategy - rather than autonomous or standalone CADx performance - in which the endoscopist's final CADx-assisted optical diagnosis constitutes the primary index test. For each diminutive polyp (≤5 mm) identified during an already-scheduled elective colonoscopy, the following sequential steps are performed: (1) the endoscopist records a CADx-unassisted optical diagnosis (adenoma, hyperplastic polyp, or other) and confidence level (high or low), which is electronically locked before any device output is viewed; (2) the CADx module is activated and displays its own characterization of the polyp; (3) the endoscopist records a final CADx-assisted optical diagnosis and confidence level, remaining free to agree or disagree with the device; (4) a research assistant documents polyp size, morphology (Paris classification), and colonic location; and (5) the polyp is resected per standard technique and sent for histopathological examination, which serves as the reference standard. This sequential-locking design - recording and locking the unassisted diagnosis before CADx activation - is the study's key strategy for minimizing contamination bias between the two diagnostic strategies, and allows each polyp to generate two paired diagnostic data points (CADx-unassisted and CADx-assisted) compared against a single histopathological reference standard.

The study will enroll consecutive patients aged 45-80 years undergoing elective colonoscopy (screening, surveillance, or diagnostic indication) at three Canadian academic centres: Santé Québec - CHUM (Montreal), the McGill University Health Centre (MUHC/CUSM, Montreal), and St. Paul's Hospital (Vancouver), who provide written informed consent before the colonoscopy and before sedation.

Inclusion criteria: signed informed consent obtained before the colonoscopy and before sedation; age 45-80 years; indication for elective colonoscopy (screening, surveillance, or diagnostic); and at least one diminutive polyp (≤5 mm) detected during the procedure, which is required for inclusion in the analytic cohort (consented patients in whom no eligible polyp is detected are documented as screen failures and do not contribute to the primary analysis).

Exclusion criteria: known inflammatory bowel disease; active colitis; coagulopathy or thrombocytopenia (INR ≥1.5 or platelets \<50×10⁹/L); familial polyposis syndrome; American Society of Anesthesiologists (ASA) classification \>III; emergency colonoscopy; and inadequate bowel preparation (Boston Bowel Preparation Scale \<6).

All staff gastroenterologists and gastroenterology trainees performing elective colonoscopies at participating centres are eligible to participate as endoscopists, with no additional optical-diagnosis training provided, reflecting real-world implementation conditions.

The primary objective is to determine the sensitivity and specificity of CADx-assisted optical diagnosis for adenomatous histology among all diminutive colorectal polyps, benchmarked against the ESGE SODA competence thresholds (sensitivity ≥80%, specificity ≥80%), using histopathology as the reference standard. Secondary objectives include: the NPV of CADx-assisted diagnosis among high-confidence diminutive rectosigmoid polyps (ASGE PIVI threshold ≥90%); agreement between CADx-assisted and pathology-based surveillance interval recommendations (ASGE PIVI threshold ≥90%); sensitivity and specificity for the rectosigmoid subset (ESGE SODA threshold: sensitivity ≥90%, specificity ≥80%); a paired comparison of CADx-assisted versus CADx-unassisted diagnostic performance; diagnostic confidence calibration; and the proportion of CADx-assisted diagnoses made with high confidence. Prespecified exploratory subgroup analyses will evaluate diagnostic performance by polyp size (≤3 mm vs. 4-5 mm), morphology, colonic location, endoscopist experience level, patient biological sex, and study centre. A tertiary/exploratory objective is an integrated health economic evaluation - combining a decision tree and a state-transition Markov model, from both health-system and societal perspectives - assessing the cost-effectiveness of a possible future CADx-assisted "resect-and-discard"/"diagnose-and-leave" strategy compared with universal histopathological examination, with results reported per the CHEERS 2022 checklist.

A minimum of 826 patients is required to evaluate both co-primary outcomes (sensitivity and specificity) with an overall power of 80%, accounting for intra-patient clustering of polyps (design effect 1.3, based on an average of 1.91 diminutive polyps per patient and an intraclass correlation of 0.3). A recruitment target of 840 patients (approximately 1,605 diminutive polyps), enrolled evenly across the three centres (280 per centre) over an expected 24-month recruitment period, has been set to exceed this minimum. Sensitivity and specificity will be estimated using generalized linear mixed models with a logit link and a random effect for patient, to account for clustering of polyps within patients; each one-sided hypothesis test (H₀: sensitivity or specificity ≤80%) will be conducted at α=0.025, with study success requiring rejection of both null hypotheses. Prespecified sensitivity analyses include a per-protocol analysis, a high-confidence-only analysis aligned with formal ESGE SODA competence assessment, a reclassification of serrated lesions as neoplastic, models incorporating a random effect for endoscopist, and a temporal analysis for learning-curve or fatigue effects. No interim analyses are planned; all analyses will be conducted after complete enrolment.

There is no increased risk to participants beyond the standard risks of colonoscopy, since all polyps are resected according to usual clinical practice regardless of study participation, and no additional procedures (blood sampling, imaging, additional sedation, or follow-up visits) are performed for research purposes. This is an investigator-initiated study (Sponsor-Investigator: Dr. Daniel von Renteln, CHUM/CRCHUM); Medtronic Canada provides the GI Genius CADx equipment on loan for the study period, together with related installation and technical support only, and has no role in study design, recruitment, data collection, endpoint adjudication, statistical analysis, interpretation, or publication decisions.

• NPV of CADx-assisted optical diagnosis for adenomatous histology (rectosigmoid, high confidence) [Time frame: Within 14 to 45 days after polypectomy (colorectal polyp resection during the index colonoscopy), once histopathology results are available]
• Surveillance interval agreement [Time frame: Outcome assessed throughout the 24-month recruitment period, with total study duration of 36 months]
• Sensitivity and specificity, rectosigmoid subset [Time frame: Outcome assessed throughout the 24-month recruitment period, with total study duration of 36 months]
• CADx-assisted vs. CADx-unassisted diagnostic performance [Time frame: Outcome assessed throughout the 24-month recruitment period, with total study duration of 36 months]
• High-confidence prediction rate [Time frame: Outcome assessed throughout the 24-month recruitment period, with total study duration of 36 months]
• Confidence calibration [Time frame: Outcome assessed throughout the 24-month recruitment period, with total study duration of 36 months]