LAUSR

Background Although some studies have assessed the learning curve of robotic-assisted total mesorectal excision for rectal cancer, most studies included limited sample sizes, no study used postoperative complications as an independent variable to analyze the learning curve of robotic rectal surgery, and no study evaluated the influence of the learning curve on long-term oncologic outcomes. Methods Clinical data on consecutive patients who underwent robotic-assisted total mesorectal excision for rectal cancer by a single surgeon between January 2015 and December 2018 at the First Affiliated Hospital of Nanchang University were retrospectively collected. The cumulative sum (CUSUM) and risk-adjusted cumulative sum (RA-CUSUM) were used to visualize the learning curve of operation time and postoperative complications (CD ≥ grade II). Comparisons of clinical outcomes at different learning phases analyzed by RA-CUSUM were performed after propensity score matching. Results A total of 389 consecutive patients were included in the analysis. The numbers of patients needed to overcome the learning curves of operation time and postoperative complications of robotic-assisted laparoscopic surgery for rectal cancer were 34 and 36, respectively. The learning process was divided into two phases based on RA-CUSUM: the learning phase (1st-36th cases) and the mastery phase (37th-389th cases). Before matching, the mastery phase had more patients with older age, lower tumor location, and neoadjuvant therapy. After matching, the two phases exhibited similar characteristics. The operation time, intraoperative blood loss, postoperative hospital stay, and postoperative complications in the mastery phase were reduced compared with the learning phase, with a median follow-up of 35 months, and the long-term oncologic outcomes were not significantly different between the two phases. Conclusions An experienced laparoscopic surgeon initially implements robotic-assisted total mesorectal excision for rectal cancer, surgical outcomes improved after 36 cases, and the learning curve seemingly did not have an obvious impact on long-term oncologic outcomes.