The Influence of Surgical Complexity and Center Experience on Postoperative Morbidity After Minimally Invasive Surgery in Gynecologic Oncology: Lessons Learned from the ROBOGYN-1004 Trial

Abstract

Background: This study was a secondary analysis of the ROBOGYN-1004 trial conducted between 2010 and 2015. The study aimed to identify factors that affect postoperative morbidity after either robot-assisted laparoscopy (RL) or conventional laparoscopy (CL) in gynecologic oncology.

Methods: The study used two-level logistic regression analyses to evaluate the prognostic and predictive value of patient, surgery, and center characteristics in predicting severe postoperative morbidity 6 months after surgery.

Results: This analysis included 368 patients. Severe morbidity occurred in 49 (28 %) of 176 patients who underwent RL versus 41 (21 %) of 192 patients who underwent CL (p = 0.15). In the multivariate analysis, after adjustment for the treatment group (RL vs CL), the risk of severe morbidity increased significantly for patients who had poorer performance status, with an odds ratio (OR) of 1.62 for the 1-point difference in the WHO performance score (95 % CI 1.06-2.47; p = 0.027) and according to the type of surgery (p < 0.001). A focus on complex surgical acts showed significant more morbidity in the RL group than in the CL group at the less experienced centers (OR, 3.31; 95 % CI 1.0-11; p = 0.05) compared with no impact at the experienced centers (OR, 0.87; 95 % CI 0.38-1.99; p = 0.75).

Conclusion: The findings suggest that the center's experience may have an impact on the risk of morbidity for patients undergoing complex robot-assisted surgical procedures.

Keywords: Conventional laparoscopy; Gynecologic oncology; Morbidity; Prognostics factors; Randomized phase III trial; Robotic-assisted laparoscopy.

Fig. 1

Study participant flow

Fig. 2

Forest plot of the treatment effect on severe postoperative morbidity by subgroup (main analysis). The lower boundary of the 95% confidence interval of WHO 2– is not represented because of the scale. The 95% confidence interval (CI) is 0.01–2.42. For each factor successively, the treatment effect of RL versus CL in the different subgroups was estimated in a multivariable model, including WHO performance status, type of surgery, treatment arm, the considered covariate, and an interaction term between treatment and the covariate. All models were hierarchical, considering the center as a random effect, except when the treatment effect was studied according to prior center RL experience (no random center effect). For each factor successively, the p value corresponds to the interaction test of the treatment effect (RL vs CL) by the considered factor. Marker size is scaled according to the number of patients in each subgroup. Detailed results of this analysis are available in Appendix Table S4. BMI, body mass index; WHO, World Health Organization performance status; TH, total hysterectomy; LND, lymphadenectomy; PeLND, pelvic lymph node dissection; AoLND, aortic lymph node dissection; RH, radical hysterectomy; RL, robot-assisted laparoscopy; CL, conventional laparoscopy

Fig. 3

Forest plot of the treatment effect on severe postoperative morbidity according to the center experience, with simple and complex surgical acts considered separately. Treatment effect of RL versus CL was estimated by subgroups (centers with < 50 vs ≥ 50 robot-assisted laparoscopies before the first inclusion in the trial) in multivariable models, including the treatment arm, WHO performance status, prior center experience in robot-assisted laparoscopy, and an interaction term between treatment group and prior center RL experience, separately in the strata of simple surgical acts (TH alone or PeLND ± TH) and the strata of difficult surgical acts (AoLND ± TH or RH ± LND). The p value corresponds to the interaction test of the treatment effect (RL vs CL) and the center experience separately for (1) surgical acts classified as “simple surgery” (i.e., TH alone or PeLND ± TH) and (2) surgical acts classified as “difficult surgery” (i.e., AoLND ± TH or RH ± LND). Marker size is scaled according to the number of patients in each subgroup. Detailed results of this analysis are available in Appendix Table S5. RL, robot-assisted laparoscopy; CL, conventional laparoscopy; WHO, World Health Organization; TH, total hysterectomy; PeLND, pelvic lymph node dissection; AoLND, aortic lymph node dissection; RH, radical hysterectomy