Post-polypectomy scar assessment traditionally relied on routine biopsies of the scar to exclude recurrent/residual polyp (RRP), but recent ESGE 2024 guidelines (1) now advocate selective biopsy based on high-confidence optical diagnosis. Existing evidence, including the ESCAPE trial (2) and a multicentre community study (3), focuses primarily on post-EMR scars for large non-pedunculated colorectal polyps (≥20 mm), with assessments performed by experts or endoscopists who underwent formal training modules. This creates a practical implementation gap, as most real-world scar assessments are performed by endoscopists with varying experience and without structured training. Our study extends optical diagnosis to all polypectomy techniques (snare, EMR, ESD, and hybrid) in a diverse national cohort without pre-assessment training, aiming to evaluate the reliability of optical diagnosis to exclude recurrence and support pragmatic adoption of guideline recommendations.

An online image-based questionnaire comprising 20 anonymised post-resection scar images (white light, narrow-band imaging and chromoendoscopy) was shared with UK endoscopists. The overall recurrence rate in the image set was approximately 20%, consistent with average recurrence rates reported in surveillance colonoscopies. Participants indicated recurrence (yes/no), with confidence (high/low), and histopathology serving as the gold standard. Experience was recorded based on whether endoscopists routinely performed large non-pedunculated colorectal polyp resections as experts or non-experts. Diagnostic performance was calculated using confusion-matrix analysis, including accuracy, negative predictive value (NPV), positive predictive value (PPV) and optical miss rate across all responses, high-confidence responses and experience subgroups.

Fifty-one endoscopists contributed 1,016 assessments. The overall negative predictive value (NPV) was high at 98.3%, with an optical miss rate of 3.4%. High-confidence evaluations demonstrated marginally better performance (accuracy 61.3%, NPV 98.7%, miss rate 1.6%), although this difference was not statistically significant on Chi-square testing. Among the 25 endoscopists who routinely performed large polyp resections (experts), diagnostic performance was superior, with an NPV of 99.7% and a miss rate of 2.1%, compared with 97.3% and 5.4%, respectively, in the 26 non-experts. Across all responses, the mean NPV was 98.5% (95% CI 96.9–100.0) and the mean optical miss rate was 3.1% (95% CI 0.4–5.8). Overall accuracy remained modest at 59.3% which was largely driven by false positives. The false negative rates remained consistently low across all subgroups, confirming the safety of biopsy avoidance when recurrence is optically excluded.

 Diagnostic performance of optical scar assessment by subgroup

Optical diagnosis demonstrated a consistently high NPV across all polypectomy scar types, even among endoscopists without structured training, supporting reliable exclusion of recurrence in real-world practice. These findings extend evidence beyond EMR-only data to include snare, ESD and hybrid resections, addressing a key implementation gap in current ESGE guidance. Our findings indicate that when an endoscopist judges a scar as recurrence-free, routine biopsy adds little clinical value, even for non-expert endoscopists, thereby supporting wider adoption of selective biopsy strategies. Incorporation of these strategies has potential to reduce unnecessary biopsies, streamline surveillance pathways and optimise resource utilisation.