达芬奇机器人手术系统辅助胃癌根治术的学习曲线

Learning curve of Da Vinci robot-assisted radical gastrectomy for gastric cancer

目的探讨达芬奇机器人手术系统辅助胃癌根治术的学习曲线.方法采用回顾性队列研究方法.收集2017年10月至2018年8月西安交通大学第一附属医院收治的42例行达芬奇机器人手术系统辅助胃癌根治术患者的临床病理资料;男30例,女12例;平均年龄为59岁,年龄范围为36~84岁.采用累积和分析法(CUSUM)及最佳拟合曲线分析手术的学习曲线;以度过学习期所需的最少手术例数为界限,将患者分为学习提高阶段组和熟练掌握阶段组,比较两组患者一般资料及手术疗效.观察指标:(1)手术情况.(2)CUSUM分析结果.(3)两组患者一般资料比较.(4)两组患者手术疗效比较.(5)随访情况.采用门诊或电话方式进行随访,了解患者术后并发症、肿瘤复发转移及生存情况.随访时间截至2019年2月.正态分布的计量资料以Mean±SD表示,组间比较采用独立样本t检验.计数资料以绝对数表示,组间比较采用χ^2检验或Fisher确切概率法.等级资料组间比较采用Mann-WhitneyU检验.结果(1)手术情况:42例患者均顺利完成达芬奇机器人手术系统辅助胃癌根治术,无中转开腹手术,无围术期死亡.42例患者中,行达芬奇机器人手术系统辅助全胃癌根治术14例,行达芬奇机器人手术系统辅助远端胃癌根治术28例.42例患者手术时间为(213±31)min,安装时间为(26±11)min.手术时间和安装时间均随着手术例数的增加呈现整体下降趋势.(2)CUSUM分析结果:CUSUM学习曲线最佳拟合方程分别为CUSUM(手术时间)=0.0169X^3-1.9133X^2+50.985X-16.595,CUSUM(安装时间)=0.0128X^3-1.0707X^2+22.189X-23.097(X为手术例数).拟合模型检验P值均<0.05,拟合优度决定系数R2分别为0.960、0.985.CUSUM(手术时间)在手术例数累积至第19例时达到峰值,19例为跨越学习曲线熟练掌握手术所需要累积的最低手术例数.CUSUM(安装时间)在手术例数累积至14例时达到峰值,14例为跨越学习曲线熟练掌握达芬奇机器人手术系统安装所需要累积的最低手术例数.(3)两组患者一般资料比较:以19例为分界点,分为学习提高阶段组和熟练掌握阶段组.学习提高阶段组患者性别(男、女),年龄,体质量指数,美国麻醉医师协会评分(1级、2级、3级),既往腹部手术史,合并基础疾病,术前超声胃镜肿瘤T分期(T1期、T2期、T3期、T4a期),肿瘤最大直径,术后临床分期(Ⅰ期、Ⅱ期、Ⅲ期),手术方式(全胃切除、远端胃切除)分别为14、5例,(60±13)岁,(23.7±2.9)kg/m2,1、16、2例,3例,8例,5、3、3、8例,(4.1±3.5)cm,6、7、6例,10、9例.熟练掌握阶段组患者上述指标分别为16、7例,(58±10)岁,(23.7±1.3)kg/m2,1、17、5例,2例,14例,3、6、9、5例,(4.7±2.7)cm,8、9、6例,18、5例.两组患者性别、年龄、BMI、美国麻醉医师协会评分、合并基础疾病、术前超声胃镜肿瘤T分期、肿瘤最大直径、术后临床分期、手术方式比较,差异均无统计学意义(χ^2=0.086,t=0.475,-0.007,Z=-0.884,χ^2=1.469,Z=-0.301,t=-0.651,Z=-0.079,-0.236,χ^2=3.076,P>0.05),两组患者既往腹部手术史比较,差异无统计学意义(P>0.05).(4)两组患者手术疗效比较:学习提高阶段组患者手术时间、术中出血量、淋巴结检出数目、术后首次进食流质食物时间、术后并发症、术后住院时间分别为(230±25)min、(176±103)mL、(21±7)枚、(5.1±2.0)d、2例、(9.3±2.5)d.熟练掌握阶段组患者上述指标分别为(191±18)min、(95±41)mL、(21±6)枚、(4.7±1.7)d、3例、(8.4±2.1)d.两组患者手术时间、术中出血量比较,差异均有统计学意义(t=5.951,-3.359,P<0.05);两组患者淋巴结检出数目、术后首次进食流质食物时间、术后住院时间比较,差异均无统计学意义(t=-0.120,0.538,1.303,P>0.05),两组患者术后并发症比较,差异无统计学意义(P>0.05).(5)随访情况:42例患者均获得术后随访,随访时间为6~16个月,中位随访时间为11个月.随访期间所有患者无严重远期并发症、无肿瘤复发转移及死亡.结论达芬奇机器人手术系统辅助胃癌根治术CUSUM学习曲线可分为学习提高和熟练掌握两个阶段;19例为跨越学习曲线熟练掌握手术所需要累积的最低手术例数.

Objective To investigate the learning curve of Da Vinci robot-assisted laparoscopic radical gastrectomy for gastric cancer. Methods The retrospective cohort study was conducted. The clinicopathological data of 42 patients who underwent Da Vinci robot-assisted radical gastrectomy for gastric cancer in the First Affiliated Hospital of Xi′an Jiaotong University from October 2017 to August 2018 were collected. There were 30 males and 12 females, aged from 36 to 84 years, with an average age of 59 years. The learning curve was evaluated using the cumulative sum (CUSUM) analysis and the best fitting curve method. According to the minimum number of surgeries required to cross the learning curve, the patients were divided into learning stage group and mastery stage group. Then general data and surgical efficacy of the two groups were compared. Observation indicators:(1) surgical situations;(2) results of CUSUM analysis;(3) comparison of general data between the two groups;(4) comparison of surgical efficacy between the two groups;( 5) follow-up. Patients were followed up by outpatient examination or telephone interview to detect the postoperative complications, tumor recurrence and metastasis up to February 2019. Measurement data with normal distribution were presented as Mean±SD, and comparison between groups was done using the independent sample t test. Count data were represented as absolute number, and comparison between groups was analyzed using the chi-square test or Fisher exact propability. Comparison of ordinal data between groups was analyzed using the Mann-Whitney U test. Results (1) Surgical situations: all the 42 patients underwent Da Vinci robot-assisted radical gastrectomy for gastric cancer successfully, without conversion to open surgery or perioperative death. Fourteen out of 42 patients underwent Da Vinci robot-assisted total radical gastrectomy and 28 underwent Da Vinci robot-assisted distal radical gastrectomy. The operation time and docking time were ( 213 ± 31) minutes and ( 26 ± 11) minutes. The operation time and docking time had a tendency to decreasing as the surgical cases increasing.( 2) Results of CUSUM analysis. The CUSUM learning curve were best modeled as a polynomial with equation: CUSUM ( operation time)= 0. 016 9X^3 - 1. 913 3X^2 + 50. 985X-16. 595, CUSUM( docking time)= 0. 012 8X^3 -1. 070 7X^2 +22. 189X-23. 097 respectively (X means the surgical case). The P value of fitting test of models was < 0. 05, with goodness-of-fit ( R 2 ) as 0. 960 and 0. 985. The CUSUM learning curve of operation time reached its peak when the number of surgical cases accumulated to the 19th case. Nineteen cases were the minimum number of surgeries required to cross the learning curve. Similarly, The CUSUM learning curve of docking time reached its peak when the number of surgical cases accumulated to the 14th case, and 14 cases were the minimum number of surgeries required to skillfully master robot installation across the learning curve.(3) Comparison of general data between the two groups: patients were divided into learning stage group and mastery stage group with 19 cases as the cut-off point. Males, females, age, body mass index ( BMI), cases in grade 1, 2, 3 of American society of anesthesiologists (ASA), cases with previous abdominal surgery history, cases with basic diseases, cases in T1, T2, T3, T4 stages of preoperative ultrasonic gastroscopic tumor T staging, maximum tumor diameter, cases in Ⅰ,Ⅱ,Ⅲ stages of postoperative clinical staging, cases with total gastrectomy and distal gastrectomy ( surgical method) were 14, 5,( 60 ± 13)years,(23. 7±2. 9)kg / m 2 , 1, 16, 2, 3, 8, 5, 3, 3, 8,(4. 1±3. 5) cm, 6, 7, 6, 10, 9 in the learning stage group, and 16, 7,(58±10)years,(23. 7±1. 3)kg / m 2 , 1, 17, 5, 2, 14, 3, 6, 9, 5,(4. 7±2. 7)cm, 8, 9, 6, 18, 5 in the mastery stage group, respectively. There was no significant difference in the sex, age, BMI, ASA score, basic diseases, preoperative ultrasonic gastroscopic tumor T staging, maximum tumor diameter, postoperative clinical staging, and surgical method between the two groups (χ 2 = 0. 086, t = 0. 475,-0. 007, Z =-0. 884,χ 2 = 1. 469, Z =- 0. 301, t =- 0. 651, Z =- 0. 079,- 0. 236,χ 2 = 3. 076, P > 0. 05). There was no significant difference in the previous abdominal surgery history between the two groups ( P > 0. 05 ).( 4 ) Comparison of surgical efficacy between the two groups: operation time, volume of intraoperative blood loss, number of lymph nodes harvested, time to first liquid food intake, cases with postoperative complications and duration of postoperative hospital stay were (230±25)minutes,(176±103)mL, 21±7,(5. 1±2. 0)days, 2,(9. 3±2. 5)days in the learning stage group, and (191±18)minutes,(95±41)mL, 21±6,(4. 7±1. 7)days, 3,(8. 4± 2. 1)days in the mastery stage group, respectively. There were statistically significant differences in the operation time and volume of intraoperative blood loss between the two groups (t = 5. 951,-3. 359, P<0. 05). There was no statistically significant difference in number of lymph nodes harvested, time to first liquid food intake, and duration of postoperative hospital stay between the two groups ( t =- 0. 120, 0. 538, 1. 303, P > 0. 05). There was no significant difference in the cases with postoperative complications between the two groups (P>0. 05).(5) Followup: all the 42 patients were followed up for 6-16 months, with a median time of 11 months. No serious long-term complications, tumor recurrence and metastasis or death occurred during the follow-up. Conclusions The CUSUM learning curve of Da Vinci robot-assisted radical gastrectomy for gastric cancer can be divided into the learning stage and the mastery stage. It is suggested that the surgeons need to finish 19 cases or more to master Da Vinci robot-assisted radical gastrectomy for gastric cancer.