Learning curve for robotic assisted laparoscopic partial nephrectomy: a single operator experience / 中华泌尿外科杂志

ABSTRACT

Objective:To present the learning curve of robotic assisted laparoscopic partial nephrectomy by a single surgeon.Methods:The clinical data of 100 patients with renal tumor who underwent robot assisted laparoscopic partial nephrectomy in Peking Union Medical College Hospital from February 2016 to April 2018 were retrospectively analyzed. There were 64 males and 36 females. The average age was 51.5(18-79) years. The average body mass index (BMI) was 24.9(19-31)kg/m 2. TNM stage was T 1N 0M 0, which indicated partial nephrectomy. The tumors were located on the left in 38 cases and on the right in 62 cases. The average maximum diameter of tumor was 3.9(1.0-6.7) cm, and the average R. E.N.A.L. score was 6.7(4-11). All operations were performed by a single surgeon. This group of operations were the first 100 robot cases for this surgeon. The mean machines docking time was 14(10-30) min, the mean operation time was 119.2(60-240) min, the mean warm ischemia time was 16.7(0-45) min, and the estimated mean blood loss was 105.2(30-500) ml. There was no conversion to open surgery or laparoscopic radical nephrectomy. No serious complications occurred (Dindo calvein grade ≥ grade Ⅲ). There were 1 case of lower extremity intramuscular venous thrombosis, 1 case of urinary fistula, 1 case of pulmonary infection and 1 case of peri-kidney hematoma after operation. All patients were cured after conservative treatment without further operation or intervention. Cumulative Sum (CUSUM) test was used to fit the learning curve of docking time and operation time, and to judge the three different learning stages of robot surgery according to the inflection of the curve (CUSUM stop rising, and CUSUM begin to decline). The characteristics of patients and perioperative data of different learning stages were compared. Result:According to the CUSUM, the learning curve of robot assisted laparoscopic partial nephrectomy was 19 cases. According to the inflection point of operation time learning curve, 100 cases could be divided into three stages, 1st-19th cases were learning stage, 20th-43rd cases were mastering stage and 44th-100th cases are proficient stage. There were no significant differences in age, gender, tumor side, intraoperative bleeding volume, postoperative pathological type, total cost of hospitalization and incidence of complications among the three stages ( P>0.05). The median BMI in the learning stage was significantly lower than that in the mastery stage and the proficiency stage (23 kg/m 2, 26 kg/m 2, 25 kg/m 2, P=0.02). The median docking time(20 min, 12 min, 12 min), median operation time (150 min, 120 min, 100 min) and median warm ischemia time (21 min, 18 min, 15 min)were gradually shortened in the learning stage, mastering stage and proficient stage ( P<0.001). The median length of postoperative hospital stay in mastering stage and proficient stage was shorter than that in learning stage (7 d, 6 d, 6 d, P=0.011). The median maximum diameter of tumor (3.7 cm, 3.9 cm, 4.0 cm)and median R. E.N.A.L. score (6, 7, 7)increased gradually in learning stage, mastering stage and proficient stage, but the difference was not statistically significant( P=0.75, P=0.16). Conclusions:The learning curve of robotic assisted laparoscopic partial nephrectomy for an experienced surgeon is about 19 cases. After the completion of the learning curve, the docking time, operation time, warm ischemia time and postoperative hospital stay of patients can be significantly shortened than the initial period, and more difficult renal tumor operations can be accomplished.