Traction-assisted versus conventional endoscopic submucosal dissection for colorectal lesions: an updated systematic review and meta-analysis of randomized trials.

backgroundEndoscopic submucosal dissection (ESD) allows en bloc, organ-sparing removal of large superficial colorectal lesions; however, conventional ESD (C-ESD) is often time intensive, technically challenging, and associated with clinically relevant adverse events. Traction-assisted ESD (TA-ESD) was developed to enhance submucosal visualization and maintain a stable dissection plane. Still, its overall impact on efficiency, effectiveness, and safety, particularly across different levels of operator experience, remains unclear. We therefore compared TA-ESD with C-ESD for superficial colorectal lesions and examined whether endoscopist experience influences outcomes.

methodsIn line with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidance, PubMed, Scopus, Web of Science, the Cochrane Library, and ClinicalTrials.gov were searched from inception through 8 November 2025. Randomized controlled trials (RCTs) enrolling patients with superficial colorectal lesions undergoing TA-ESD versus C-ESD and reporting outcomes on an intention-to-treat (ITT) or modified intention-to-treat (mITT) basis were included. Pooled estimates were generated using random-effects models, with prespecified subgroup analyses according to endoscopist experience.

resultsFrom 930 screened records, 10 RCTs met the inclusion criteria, totaling 1008 participants (499 TA-ESD; 509 C-ESD). TA-ESD was associated with shorter procedure time than C-ESD (MD = -19.23 min; 95% CI -26.7 to -11.8; P < 0.0001), although heterogeneity was substantial and certainty was very low. Resection quality was high and similar between approaches, with no meaningful differences in en bloc or R0 resection rates. In the C-ESD arm, 13% of cases required rescue conversion to traction. TA-ESD increased overall resection speed (MD = 2.06 mm

conclusionsIn colorectal RCTs, TA-ESD was associated with shorter procedure time than C-ESD, consistent with prior meta-analyses. However, certainty for this endpoint was very low, and heterogeneity was substantial. The main demonstrated advantage was procedural efficiency, whereas resection quality and overall safety were broadly comparable. Implementation-related outcomes, including rescue conversion to traction and handover to another operator, were reported in only a limited number of trials and should therefore be considered exploratory and hypothesis-generating rather than definitive. Future multicenter RCTs should standardize workflow-related endpoint definitions and determine which lesions, procedural settings, and operator contexts derive the greatest incremental benefit from dedicated adjunctive traction.