| Description: |
Background Endoscopic mucosal resection (EMR) is widely used for removing large colorectal polyps (≥20mm). The goal is complete resection with minimal adverse events and a low adenoma recurrence rate (ARR). Optimal initial evaluation of the polyps and subsequent resection by a therapeutic endoscopist is of crucial importance to achieve good outcomes. Aims To evaluate the accuracy of the initial evaluation of large polyps referred for EMR to a tertiary Centre. Subsequently, we analyzed the EMR technical success, complications, recurrence rate, final histology, appropriate follow-up, and need for surgery for patients referred for EMR of large colonic polyps. Methods This was a retrospective review of prospectively collected data of patients referred for EMR of colorectal polyps ≥20mm between 2017 and 2024. The procedures were performed by two therapeutic endoscopists. Results A total of 170 patients (median age 66; 52.3% female) with 187 polyps (median size 30mm) were included. Among referred polyps, 89 (48%) had accurate descriptions; 50 (27%) underestimated size, 41 (22%) overestimated size, 16 (8.6%) misclassified Paris type, and the location description was not accurate in 9 (5%). Initial manipulation of polyps at the first endoscopy was done in 58% of polyps; this included biopsies of the polyp (6.4%), tattooing nearby (8.4%), lifting (7.5%), and attempted resection in 15.7%. Median time to first EMR was 30 days. Median polyp size was 30mm (range 20-80mm). After EMR, high-grade dysplasia (HGD) was reported in 28 polyps (15%) and invasive carcinoma in 7 (3.7%). EMR could not be completed in 2 cases (1.1%) due to suspicion of submucosal invasion. Complications included mild intra-procedural bleeding in 37 patients (20%), post-polypectomy bleeding in 3 (1.7%), perforation in 1 (0.5%), which was managed conservatively and one case of post-polypectomy syndrome. Average time to follow-up was 7 months. Residual adenomatous tissue was present in 33 cases (18%), with HGD in 9 (5%). ARR increased according to ... |