Laparoscopic and robotic pyeloplasty as minimally invasive alternatives to the open approach for the treatment of uretero-pelvic junction obstruction in infants: a multi-institutional comparison of outcomes and learning curves

Background Since the development of minimally invasive surgery (MIS), laparoscopic and robotic approaches have been widely adopted. However, little has been published detailing the learning curve of MIS, especially in infants. Objective To quantify the learning curve of laparoscopic (LP) and robot-assisted laparoscopic pyeloplasty (RAL-P) for treatment of uretero-pelvic junction obstruction (UPJO) in infants evidenced by number of cases, operative time, success and complications. Patients and methods Between 2009 and 2017, we retrospectively reviewed pyeloplasty cases for treatment of UPJO in infants at three academic institutions. The primary outcome was success. Secondary outcomes were UPJO recurrence, complications, and operative time as a surrogate of skill acquisition. Continuous variables were analyzed by t test, Welch-test, and one-way ANOVA. Non-continuous variables were analyzed by Chi-squared test or Fisher's exact test. Learning curves (LC) were studied by r-to-z transformation and CUSUM. Results Thirty-nine OP, 26 LP, and 39 RAL-P had mean operative times (OT) of 106, 121, and 151 min, respectively. LCs showed plateau in OT after 18 and 13 cases for LP and RAL-P, respectively. RAL-P showed a second phase of further improvements after 37 cases. At 16 months follow-up, there were similar rates of success and complications between the three groups. Conclusions Despite different duration of learning phases, proficiency was achieved in both LP and RAL-P as evidenced by stabilization of operative time and similar success rates and complications to OP. Before and after achievement of proficiency, LP and RAL-P can be safely learned and implemented for treatment of UPJO in infants.

Automated summary

The study compared the learning curves of three methods for UPJO treatment in infants at three academic institutions from 2009 to 2017. The results showed that despite different learning curve durations for LP and RAL-P, similar proficiency was achieved in terms of stabilization of operative time, success rates, and complications for both methods.