Underwater endoscopic mucosal resection (UEMR) is an emerging technique for the management of large, benign colonic polyps, with higher en bloc and R0 resection rates and shorter procedure time compared to conventional EMR. Its role in the management of polyps suspicious for submucosal invasion is yet to be defined. Few studies have investigated the depth of resection with UEMR, with inconsistent results when compared to conventional EMR. No studies have compared UEMR with endoscopic submucosal dissection (ESD). We aimed to compare the resection depth of UEMR and ESD for large (20-30mm) non-pedunculated colonic polyps.

We performed a retrospective cohort study at a single UK tertiary referral centre from 2015 to 2024. All en bloc resections by UEMR or ESD of 20-30mm colonic polyps were included. Pedunculated or semi-pedunculated polyps, recurrent polyps and patients with inflammatory bowel disease were excluded. Histopathological samples were retrospectively reviewed, with measurement of maximum and minimum submucosal (SM) resection depths, measured from the muscularis mucosae. Primary outcome was the depth of SM in the resected specimen. Secondary outcomes included R0 resection rate, vertical margin (VM) positivity (defined as <0.1mm) in cases of adenocarcinoma, and rates of deep mural injury (DMI).

29 UEMR and 27 ESD resections were identified. Median lesion diameter was similar between the two groups (22 vs 25mm). Lesions resected by UEMR were more commonly protruding (69% vs 26%) and in the right colon (93% vs 48%), with ESD resections more commonly flat lesions in the left colon. Median maximum SM was significantly higher with UEMR (4.2 vs 1.75mm; p<0.001). There was no significant difference in median minimum SM depth (0.4 vs 0.5mm). Median maximum SM depth was significantly higher with UEMR for both protruding (5.5 vs 2.25mm; p=0.015) and flat lesions (2.3 vs 1.5mm; p=0.009). 20/29 UEMR and 27/27 ESD specimens were pinned. Comparing pinned vs unpinned UEMR specimens, there was no significant difference in median maximum SM depth (3.5 vs 5.5mm; p=0.22). There was no significant difference in R0 resection rates between UEMR and ESD (76% vs 93%; p=0.15). One T1b adenocarcinoma was resected by UEMR, with positive VM. Three T1a adenocarcinomas were resected by ESD, all with negative VM.  Three T1b adenocarcinomas were resected by ESD, 1/3 with positive VM. DMI (type III) was observed in one UEMR resection. There were no perforations in either group.

In our study, en bloc UEMR achieves significantly greater maximum SM depth than ESD in both flat and protruding colonic polyps 20-30mm in diameter. R0 resection and complication rates were similar between techniques. Prospective comparative studies are needed to define the role of UEMR in managing colonic lesions suspicious for superficial submucosal invasion.