Surgical outcomes
Operation time significantly decreased with increasing case volume: cases in group 1 had a significantly longer operation time compared to group 2 and group 3 (both p <0.001). Operation time in the newly added group 3 was similar to group 2 (p =1.00) (Figure 2 ).
In 50 patients a grade 3-4 (severe) adverse event occurred at any time during follow-up of whom 31 patients (19.4%) had a grade 3-4 adverse event that occurred intraoperatively or within the initial 42 postoperative days. No grade 5 event (death related to surgery) occurred. The number of severe adverse events was highest in group 1 but did not significantly differ between the three groups: 13 patients (21.3%) in group 1, 10 (9.6%) patients in group 2, and 8 (13.1%) patients in group 3 (p =0.11)(see Table 1 ). When comparing the three groups separately, the proportion of severe adverse events in group 3 was not significantly different from group 1 and group 2, but severe adverse events were significantly lower in group 2 compared to group 1 (p =0.04).
One case was converted to conventional laparoscopy because technical issues hindered robot-assisted surgery (i.e. clashing robotic arms). No intraoperative surgical adverse events necessitated a direct conversion to laparotomy or laparoscopy. Table 2 provides an detailed overview of all intraoperative and postoperative grade 3-5 adverse events within 42 days after surgery, of which multiple adverse events could occur in one patient. In total 36 severe adverse events occurred within 42 days in 31 patients: 28 grade 3 events in 23 patients and eight grade 4 events in eight patients. Six patients underwent laparotomy as a secondary procedure to manage their grade 4 adverse event: three with ureter injury, two with postoperative hypovolemic shock due to intraabdominal haemorrhage, and one with perforation of the sigmoid. One patient with a grade 4 event needed surgical intervention for a severe cervical bleeding but no laparotomy. The eighth patient with a grade 4 event had an obstructed airway caused by a pneumomediastinum. This was detected directly after detubation and managed by reintubation and ICU admission where laryngoscopy excluded an airway injury.
Regarding the distribution of severe adverse events across the three groups, a shift from intraoperative to postoperative adverse events was observed: in group 1 four intraoperative events occurred compared to zero in group 3. This shift is inherent to the shift from severe urological injuries to postoperative symptomatic lymphoceles (seeSupplementary Table S1 ). Subgroups of adverse events were too small to perform reliable statistical analysis on.