Traditional endoscopic mucosal resection (T-EMR) is the primary endoscopic technique for resecting large size sessile colorectal lesions. However, the potential for residual adenomatous tissue to be found at the EMR site at surveillance colonoscopy remains problematic. Therefore, more recently, several modified EMR techniques have been developed including underwater EMR (U-EMR), cold EMR (C-EMR), tip-in EMR (TI-EMR), cap-assisted EMR (CA-EMR), and the addition of thermal ablation of resection margins following T-EMR using snare tip soft coagulation (STSC) or argon plasma coagulation (APC). However, few randomized controlled trials (RCTs) directly compare these modified techniques with T-EMR, thus limiting clinical decision making.Therefore, to compare the effectiveness and safety of different EMR techniques, we conducted a systematic review and network meta-analysis.

We searched PubMed, Embase, and the Cochrane Library from inception to April 2025 for RCTs comparing EMR techniques for the treatment of colorectal polyps ³15 mm. We estimated relative risks (RRs) with 95% credible interval (CrI) for dichotomous outcomes using Bayesian network meta-analysis. To assess the credibility of each comparison, we used Confidence in Network Meta-Analysis (CINeMA).

We included 17 RCTs with 2,832 polyps resected, using seven different EMR techniques: T-EMR (16 studies, 1,263 polyps); U-EMR (6 studies, 303 polyps); C-EMR (4 studies, 437 polyps); CA-EMR (1 study, 143 polyps); TI-EMR (2 studies, 92 polyps); and thermal ablation of resection margins (STSC: 3 studies, 440 polyps and APC: 3 studies, 154 polyps). As compared to T-EMR, U-EMR, CA-EMR, and T-EMR plus thermal ablation were all associated with significantly lower residual adenomatous tissue rates at surveillance colonoscopy, while C-EMR showed higher residual adenomatous tissue rates. SUCRA values for risk of residual adenomatous tissue with each technique were (higher SUCRA values = lower risk of residual adenomatous tissue): CA-EMR: 0.836; T-EMR plus STSC: 0.826; TI-EMR: 0.687; T-EMR with APC: 0.544; U-EMR: 0.413; T-EMR: 0.189; C-EMR: 0.007. There was no statistically significant difference between T-EMR and TI-EMR for residual adenoma rates, but TI-EMR was associated with higher rates of R0 resection than T-EMR.  As compared to T-EMR, en-bloc resection rates were superior using U-EMR, yet lower with C-EMR. Intraprocedural bleeding was lower with C-EMR, CA-EMR and T-EMR. C-EMR and T-EMR also had lower rates of delayed bleeding, and perforation rates were lower with C-EMR. 

As compared with T-EMR, all modified EMR techniques appear to be effective, except C-EMR, which consistently demonstrated higher residual adenomatous tissue rates at surveillance colonoscopy. However, these outcomes should be viewed with caution given the imprecision of TI-EMR outcomes and limited data for CA-EMR. Therefore, the choice of colorectal EMR technique should be individualized based on patient and lesion characteristics as well as endoscopist EMR technical experience.