Evaluating the learning curve of robot-assisted partial nephrectomy with the cumulative sum method

Abstract

Robot-assisted partial nephrectomy (RAPN) has become the standard treatment for small renal tumors, offering better perioperative outcomes than open surgery. However, objective evaluations of the RAPN learning curve are limited. While the Trifecta criteria-comprising negative surgical margins, no perioperative complications, and warm ischemia time (WIT) ≤ 25 min-are commonly used to assess surgical outcomes, they are inadequate for continuous proficiency assessment. This study aimed to evaluate the RAPN learning curve using the cumulative sum (CUSUM) method based on Trifecta achievement and its components. We retrospectively analyzed 119 RAPN cases performed by three surgeons at a single institution between 2017 and 2022. All surgeons (≥ 30 cases; ≥ 15 year experience) were included. CUSUM charts were created using Trifecta achievement rates with thresholds (p₀ = 0.4, p₁ = 0.8), and further analysis was performed on individual components. Distinct learning curve transitions were observed only in Surgeon B, with proficiency achieved at the 9th case for complication rates and the 4th case for overall Trifecta achievement. No clear transitions were seen in WIT or surgical margins, or in any component for Surgeons A and C. These findings suggest that Surgeons A and C may have already attained proficiency before the study period. The CUSUM method offers a practical tool for visualizing and quantifying individual learning curves in RAPN based on clinically relevant criteria. Despite some limitations, CUSUM enables continuous, surgeon-specific assessment. Future studies should integrate additional metrics to develop more comprehensive training programs and improve surgical safety and outcomes.

Keywords: CUSUM method; Education; Learning curve; RAPN; Trifecta.