To identify factors associated with therapeutic failure and complications after colorectal ESD in a single-center cohort covering the period after completion of the operator’s learning curve.

 Retrospective analysis of 212 consecutive patients (2019–2022). Endpoints included therapeutic success (R0/Rx), adverse events according to the Clavien–Dindo classification, and potential predictors (pre-procedural biopsy, hybrid resection technique [h-ESD], Paris classification, lesion size). The data derive from our own clinical series.

En bloc resection was achieved in 93.87%, R0 in 91.04%, and therapeutic success in 91.04%; complications occurred in 8.02% (bleeding 2.36%, perforation 3.30%). In univariable models, lack of success was more frequently associated with pre-procedural biopsy (OR 3.29; p=0.0168) and h-ESD (OR 8.31; p=0.0025). In the bivariable model, biopsy (OR 3.18; p=0.024) and h-ESD (p=0.0045) remained significant. Lesion size did not significantly affect efficacy (p=0.116), but correlated with the risk of complications; higher risk was observed, among others, for Paris classification 0-IIa-Is.

After completion of the learning curve, ESD provides high efficacy and acceptable safety also in a non-academic center. Avoiding routine biopsy before planned ESD and striving for full ESD (without hybrid resection) may increase the rate of radical resections and reduce complications.