Colorectal endoscopic submucosal dissection (ESD) allows en-bloc removal of early neoplastic lesions but remains technically demanding in Western non-academic settings, where exposure to high-volume training is limited. This study aimed to evaluate learning progression and performance determinants of untutored colorectal ESD in a European non-academic center. 

This retrospective cohort included all consecutive colorectal ESD procedures performed by a single operator from March 2016 to October 2025. Demographic, lesion, and procedural data were prospectively collected. Learning-curve assessment used case-by-case speed analysis and linear regression. The operator’s concurrent experience with additional third-space endoscopy procedures was also considered. Adverse events were classified by ASGE severity criteria and categorized as mild, moderate, severe, or fatal and immediate or delayed. 

A total of 114 colorectal ESDs were performed (mean age 66.5±12.0 years; 56.5% male). Key outcomes included: median lesion size 11.2cm (range 0.36–89.25), median max diameter 4.0cm (range 0.6— 10.5), median dissection speed 7.5cm /hour (range 1–48), high en-bloc and R0 resection rates (97.3% and 93.8%) and overall complication rate 13.2% (3.6% perforation rate), including two macro-perforations (the only treated surgically), two micro-perforations, four early and two delayed bleeding events and three strictures. The learning curve showed speed increasing progressively from cases 20–60 and peaking at 70–90 (cumulative experience from additional third-space endoscopy procedures, ~70 and ~50 cases during these respective periods). Dissection speed reached above 9 cm /hour after 30 colorectal ESD cases and became consistent after 60-70. Significant location-based differences were observed in lesion size (p=0.036), dissection speed (p=0.02), circumferential involvement (p=0.001), and perforation rates (p=0.04), with rectal and sigmoid lesions showing higher adjusted speeds and right-sided lesions being more challenging. En-bloc and R0 rates did not differ by location. Complications clustered between cases 15–30 before declining after case 35, reflecting an initial learning-curve vulnerability followed by sustained proficiency. 

In this single-operator European cohort, colorectal ESD exhibited a clear learning curve, with early variability progressively replaced by stable improvements in dissection speed and procedural efficiency, while en-bloc and R0 rates remained consistently high. When colorectal ESDs were combined with other ESD-related procedures, the broader technical exposure likely accelerated skill acquisition. Consequently, the true proficiency threshold for colorectal ESD is likely higher than the apparent cutoff observed here, with at least 100 procedures required to achieve an expert level.