Learning process analysis of robotic lateral pelvic lymph node dissection for local advanced rectal cancer: the CUSUM curve of 78 consecutive patients.

PURPOSE: Robotic lateral lymph node dissection (LLND) has been described as a safe and feasible procedure for local advanced rectal cancer. The aim of this study was to evaluate the learning curve for robotic-assisted LLND. METHODS: We collected data on 78 consecutive patients who underwent robotic-LLND at our hospital. The learning curve was analyzed using the cumulative sum (CUSUM) method to assess changes in the unilateral LLND operative times across the case sequence. RESULTS: Among the 78 patients, 52 underwent bilateral LLND and 26 underwent unilateral LLND. A total of 130 consecutive data were recorded. We arranged unilateral robotic-LLND operative times and calculated cumulative sum values, allowing the differentiation of three phases: phase I (learning period, cases 1-51); phase II (proficiency period, cases 52-83); and phase III (mastery period, cases 84-130). As the learning curve accumulated, the operation time and estimated blood loss of unilateral robotic-LLND decreased significantly with each phase (P < 0.05). By 12 months after surgery, the International Prostatic Symptom Score of patients at phase III was significantly lower than at phase I (P < 0.05). CONCLUSION: The CUSUM curve shows three phases in the learning of robotic-LLND. The estimated learning curve for robotic-assisted rectal-LLND is achieved after 51 cases.