Benchmarking sessile serrated lesion detection in colonoscopy: individual endoscopist performance has greater impact than specialty training - a retrospective analysis of 12,064 procedures.

BACKGROUND AND

aimsSessile serrated lesions (SSLs) are implicated in up to 30% of colorectal cancers through the serrated neoplasia pathway. Current guidelines lack established benchmarks for SSL detection rates (SSL-DR), with significant geographic variation reported internationally. The 2024 American Society for Gastrointestinal Endoscopy (ASGE) and American College of Gastroenterology (ACG) recommend a minimum SSL-DR of > 6% for patients aged ≥ 45, while the Gastroenterological Society of Australia (GESA) maintains a 4% minimum. This study aimed to determine optimal SSL-DR benchmarks for clinical practice, compare detection rates between gastroenterologists and general surgeons, and identify whether specialty training or individual endoscopist performance primarily drives detection variability.

methodsRetrospective analysis of 12,064 colonoscopies performed between 2020 and 2023 at a regional Australian centre. SSL-DR was defined as the proportion of colonoscopies in which at least one SSL was detected. Detection rates were compared between gastroenterologists and general surgeons, with additional analyses excluding high-performing endoscopists. Statistical comparisons used chi-square tests and multivariable logistic regression. Post-hoc power analysis and bootstrap confidence intervals validated statistical robustness of findings and proposed benchmarks. Sequential exclusion analysis was employed to isolate individual performance effects from specialty-based differences.

resultsOverall SSL-DR was 20.7% (2,502/12,064 procedures), substantially exceeding international averages. Initially, gastroenterologists achieved significantly higher SSL-DR than general surgeons (24.5% vs. 12.4%, difference 12.1% points, 95% CI: 10.5-13.7, p < 0.001). However, after excluding two high-performing gastroenterologists who performed 48.1% of procedures but detected 71% of all SSLs (individual SSL-DRs: 31.6% and 24.5%), the remaining gastroenterologists' SSL-DR (11.6%) was comparable to general surgeons (12.4%, p = 0.28). The study achieved > 99% statistical power for detecting the observed differences. Bootstrap analysis (10,000 iterations) validated proposed benchmarks: minimum standard 8.5% (95% CI: 7.7-9.2%) target standard 13.0% (95% CI: 12.1-13.8) high-performance 17.0% (95% CI: 16.1-17.9%), and aspirational range 25.0-30.0%.

conclusionsIndividual endoscopist factors have greater impact on SSL detection than specialty training. The dramatic reversal of specialty advantage after excluding just two practitioners demonstrates that apparent specialty-based differences may be driven by a small number of high-performing individuals. Quality improvement initiatives should focus on identifying and disseminating practices of high-performing individuals rather than emphasizing specialty distinctions. We propose benchmarks that align with the 2024 ASGE-ACG guidelines while accounting for Australian detection patterns and provide a quality improvement framework for clinical implementation.