Bayesian network analysis of open, laparoscopic, and robot-assisted radical cystectomy for bladder cancer

Abstract

We have performed the direct and network meta-analysis to evaluate the safety and efficacy of robot-assisted (RARC) versus laparoscopic (LRC) versus open radical cystectomy (ORC) for bladder cancer (BCa). A systematic search of PubMed, Cochrane Library, and Embase was performed up until Dec 20, 2019. Outcome indexes include oncologic outcomes (the recurrence rate, mortality), pathologic outcomes (lymph node yield (LNY), positive lymph node (PLN), positive surgical margins (PSM)), perioperative outcomes (operating time (OP), estimated blood loss (EBL), blood transfusion rate, the length of hospital stay (LOS) and the time to regular diet) and postoperative 90-day complications. We have analyzed 6 RCTs, 23 prospective studies, and 25 retrospective studies (54 articles: 6382 patients). On one hand, the direct meta-analysis shows RARC is better than LRC or ORC. On the other hand, the clinical effects of the recurrence rate, Morbidity, PSM, LNY, PLN, and postoperative 90-day complications of RARC, LRC and ORC are all no statistical significance by network meta-analysis. Moreover, the probability rank shows that the comprehensive rank of RARC is better than LRC or ORC. The clinical effects of OP, EBL, LOS, blood transfusion rate and the time to regular diet are all statistical significance by network meta-analysis. There are ORC > LRC > RARC in the EBL ranking. Patients with RARC exhibited a decrease of LOS compared to those with LRC or ORC. Patients with RARC exhibited a decrease in blood transfusion rate and the time to regular diet compared to those with ORC. Patients with ORC exhibited an increase of OP compared to those with RARC or LRC. The heterogeneity tests of most studies are < 50%. Most studies have no publication bias and the quality of the selected studies is good. The direct meta-analysis and network meta-analysis suggest that RARC is better than LRC or ORC according to comprehensive analysis. However, we need a large sample size and more high-quality studies to verify and improve in the further.