EMR of large laterally spreading lesions (LSL) in the colon is a safe and effective alternative to surgery. Post-EMR scar assessment currently involves taking biopsy specimens of the scar to detect residual or recurrent adenoma (RRA). The accuracy of endoscopic imaging of the post-EMR scar is unknown. We aimed to determine the accuracy of a standardized imaging protocol in post-EMR scar assessment.

Prospective, single-center data from the Australian Colonic EMR study were analyzed. Consecutive patients undergoing first surveillance colonoscopy (SC1) after EMR of a large LSL were eligible. All scars were sequentially examined with high-definition white light (HD-WL) and narrow-band imaging (NBI) in a standardized fashion and then biopsies were performed. Endoscopic recurrence (recurrence at the post-EMR scar detected by systematic endoscopic assessment) was compared with the histologic findings.

One hundred eighty-three post-EMR scars were included. Thirty of 183 (16.4%) were confirmed to have RRA histologically at SC1. Thirty-seven of 183 (20.2%) post-EMR scars demonstrated RRA endoscopically. The sensitivity and specificity of endoscopic RRA detection were 93.3% (95% confidence interval [CI], 77.9%-99.2%) and 94.1% (95% CI, 89.1%-97.3%), respectively. The positive predictive value was 75.7% (95% CI, 58.8%-88.2%) and the negative predictive value was 98.6% (95% CI, 95.1%-99.8%). The diagnostic accuracy was 94.0%. Sensitivity was higher for the combination of HD-WL and NBI as opposed to HD-WL alone (93.3% vs 66.7%). The specificity was high for both HD-WL and HD-WL + NBI (96.1% and 94.1%, respectively). Flat morphology of RRA was better seen with NBI (P = .002).

Endoscopic detection of RRA in the post-EMR scar is highly accurate using a standardized imaging protocol with HD-WL and NBI. This allows real-time, accurate detection of recurrence and its concurrent treatment, and raises the possibility that routine biopsy of the post-EMR scar may not be necessary.