The European Association of Urology Pediatric guidelines acknowledge for pyeloplasty procedure that “in good and experienced hands, the open, laparoscopic, or robotic approaches have the same good outcome” (1). This statement is not based on level 1 evidence and pertains only to pyeloplasty. Thanks to its precise suturing, the robot-assisted approach is used in the majority of minimally invasive pyeloplasty (more than 80%) of teenagers in the United States (2–4).

It remains controversial whether to use a transperitoneal or retroperitoneal approach for pyeloplasty in children.

A large single-center series with long-term follow-up addressed the impact of 10-year retroperitoneal laparoscopic pyeloplasty experience in a pediatric teaching center, demonstrating that it is a safe, reliable, and efficient procedure with an excellent outcome (5). Surgeons from our department have been trained in the retroperitoneal laparoscopic approach. Retroperitoneoscopic pyeloplasty, as described by Yeung et al. has been the standard treatment for ureteropelvic junction obstruction (UPJO) at our institution since 2010 and remained the approach of choice when we started the multidisciplinary pediatric robotic program in our center (6). Publications evaluating the retroperitoneal approach for robot-assisted laparoscopic pyeloplasty are limited (7, 8).

We present our findings in terms of safety and efficacy during the first 2 years of a multidisciplinary pediatric robotic program.

Between December 2016 and November 2018, with the Da Vinci Xi Surgical System, we performed 50 robot-assisted laparoscopic pyeloplasties (RALP): 37 by a retroperitoneal approach and 13 by a transperitoneal approach for redo procedure, horseshoe kidney, ectopic kidney, based on a previously published algorithm that we have since modified (Figure 1) (5).

All 50 cases were either done by or with the assistance of one surgeon (TB) who had no prior experience in robotic surgery but who had extensive experience in the retroperitoneal approach (nephrectomy, pyeloplasty). Before starting robotics, the main author had performed almost one hundred laparoscopic retroperitoneal or transperitoneal pyeloplasties (Unpublished data). He also observed an experienced operator in robotics (one of the co-authors Christophe VAESSEN), who later on assisted him with his first procedure.

This retrospective analysis of prospectively collected data received approval from an independent ethics committee (Comité de Protection des Personnes, CPP Ile de France VII). The sponsor was Assistance Publique—Hôpitaux de Paris (APHP, Clinical Research and Innovation Delegation) and this project was funded by a grant from Necker Hospital. It is registered with the ClinicalTrials.gov identifier NCT03274050.

The diagnosis of UPJO was confirmed by renal ultrasound, Uro-magnetic resonance imaging (Uro-MRI) and technetium Tc 99m mercaptoacetyltriglycine-3 (MAG-3) renal scan (RS). Those with equal differential renal function (DRF) on RS were symptomatic (ipsilateral flank pain and/or recurrent febrile urinary tract infections and/or high blood pressure) with pyelocaliceal dilatation on ultrasound.

Table 1 shows the demographics, indication for surgery, and surgical variables. All cases were completed with the robotic system, with no conversion. We did not experience any robot malfunction, system failure, or complication. No emergency undocking was needed.

The median age was 7.9 years (5.1–13.8); the youngest was 2 years old. The median weight was 23 kg (17–41) with the smallest weighing 12.4 kg. Aberrant crossing vessels were present in 18 children.

There was no significant blood loss.

The peritoneum was opened once during the 3rd port placement, and once during the procedure. To overcome this difficulty, the assistant lifted up the kidney through the assistant port, with no need for conversion to open surgery or laparoscopy

Neither parenchymal injury nor vascular injury in case of polar vessel occurred.

In one patient, preoperative evaluation (ultrasound, MRI, and MAG-3 RS) missed a HSK which was discovered during RALP via the retroperitoneal approach and the procedure was continued using an anterior extraperitoneal approach to the left renal pelvis.

The Black-Star^® magnetic stent was inserted in 17 children.

Median set-up time, from skin incision until the end of the 4-port insertion, was 33 min (29–48). Initially taking more than an hour, over time the operator became more than twice as proficient (Figure 4A).

The CUSUM methodology demonstrated a biphasic learning curve (Figure 4B). In the first phase, a steep positively sloping curve appeared until the 15th case, indicating the process of overcoming the learning period. It was followed by a flatter negatively sloping curve indicating a period of gaining competence and consolidation without gaining mastery.

Median surgeon's console time was 151 min (136–182). For 5 patients, console time was < 120 min; while for 4 patients, console time was >210 min. Median complete anastomosis time was 79 min (68–90). Despite a trend of a flat negatively sloping curve (Figures 5, 6), the CUSUM methodology could not demonstrate 2 phases. Practical training of other colleagues was possible after 10 cases performed by the same surgeon. One senior surgeon and 2 fellows, with extensive experience in laparoscopy, took part in the retroperitoneal RALP operation and were assisted in different steps of the procedure. However, they have not yet performed the procedure independently.

The median hospital stay was 1 day (1–1); all but 4 patients (90%) were discharged the day after the procedure.

The one blind-ended JJ stent was removed after a median of 37 days (31–44) under general anesthesia as a day procedure and the Black-Star^® magnetic stent was removed after 34 days (31–35) in the outpatient clinic.

All patients have been followed up regularly (median follow-up 9 months (5–13) and remain asymptomatic. All had a decrease of the AP diameter on ultrasound in the postoperative course (Figure 7). Based on our protocol, because of this decrease in AP diameter, we have done post-operative MAG-3 only in cases of asymmetric preoperative function to evaluate the effect of the pyeloplasty on renal function. A total of 14 children have had postoperative MAG-3. No patient has had a decrease in renal function. In 9 children, the operated DRF has improved (>3%). No redo procedure has been necessary.

Six children (4 girls and 2 boys) were treated for 8 febrile UTI in total (Clavien Grade 2) with oral antibiotics. These were not immediate postoperative UTIs, rather they occurred during the period in which the Double-J catheter was indwelling.

Weight was the only significant factor predicting console time [19 kg (16–27) for time < 150 min vs. 31 kg (22–48) for time > = 150 min, p = 0.03) (Table 2).

The preliminary results of retroperitoneal RALP in children in the debutant 2 years of a multidisciplinary pediatric robotic program in our center demonstrates that this approach can achieve results comparable with those reported with open pyeloplasty, laparoscopy or RALP using the transperitoneal approach (11–13).

There is still some controversy concerning which approach to choose: transperitoneal or retroperitoneal (14).

While the transperitoneal approach offers the advantage of a larger working space, retroperitoneal approach accesses the urinary tract directly and detection of crossing vessels is easier. The potential risk of intra-abdominal organ injury during transperitoneal access, though a rare event, is also avoided. It is a matter of personal preference, based on the experience of the surgeon who should ideally be familiar with both approaches. Since the beginning of the multidisciplinary pediatric robotic program in our center, we have modified the previously published algorithm (Figure 1).

Olsen and Jorgensen published the only series of retroperitoneal RALP cases in 13 pediatric patients (7). Median operative time was 173 min with no obstruction observed at follow-up. Median hospitalization was 2 days and only 1 complication of ureteral stent occlusion occurred. The authors published an expanded series of 65 children with a follow-up of 5 years in 2007, affirming their earlier results (8).

As already highlighted by Olsen et al. previous experience with retroperitoneal pyeloplasties using standard laparoscopic instruments facilitated our transition to this new technology (8). Both approaches share similar basic procedural and technical elements, with the same three instruments being used: monopolar scissors, bipolar forceps, and needle holder. However, there are significant differences related to port placement and size.

For laparoscopic pyeloplasty, we use 3 ports (5-3-3 or 10-5-5 mm) (5). Retroperitoneal access is achieved through the first incision, one finger width from the lower border of the tip of the 12th rib. The Gerota's fascia is opened under direct vision and the first blunt trocar is introduced directly inside the opened Gerota's fascia. A working space is created by gas insufflation dissection. The second trocar is inserted posteriorly in the costovertebral angle, in front of the lumbosacral muscle. Third trocar insertion is in the anterior axillary line, a finger width from the top of the iliac crest.

For robotic pyeloplasty, we have chosen our port placement based on surgical techniques in adult centers and not as described by Olsen et al in 2004 (7). We use 3 8-mm robotic ports and one 8-mm AirSeal^® iFS System port, placed linearly from the iliovertebral angle to the iliac fossa, as already described. While this accessory (4th) port is optional given that suction/irrigation devices can be passed through the instrument ports of the system in the case of bleeding or need for double-J stent or suture material, it allows us also to operate in greater security and maintain time efficiency. It is very important to keep meticulous hemostasis maintaining a clear vision and thereby orientation in a space with the psoas muscle being a major landmark which is always located on the lower part of the screen.

The robotic trocar diameter is larger (8 vs. 3 or 5 mm) which limits its application to very small children. In our study, the youngest child was 2 years old and 12 kg, older than the infants operated in retroperitoneal laparoscopic pyeloplasty (15). In the future smaller instruments will hopefully be available, making the procedure more suitable for smaller children and infants, as with standard laparoscopic instruments.

However, we feel that instrument size is not the limiting factor as the procedure is performed in a restricted workspace with limited instrument movement, meaning a low risk of collision of the various parts of the robot. The Da Vinci Xi arms are less bulky and therefore much more adapted to pediatric surgery. Thakre et al. have evaluated the performance of robot assisted laparoscopic skills in different workspace sizes with the first generation of the da Vinci Surgical System. While small cubes measuring 40 and 45 mm in diameter have posed difficulties to the surgeon due to arm collisions rendering it impossible to perform drills, these same skills have been replicated with difficulty but no arm contact in cubes measuring 50 and 60 mm, and finally executed with ease in cubes larger than 70 mm, the latter being comparable to the retroperitoneal space in a small child (16).

We suggest that the major limitation lies more in the length/depth needed to operate the robotic instrument in the restricted area that is the retroperitoneal space in small children.

The robotic system mainly speeds up performance of the anastomosis. The LC in robotic surgery is increasingly being analyzed though, for the moment, operative time has been the single parameter used to assess proficiency in studies. Given the disparities within this measure, a case load ranging from 15 to 58 has been thus far proposed as the LC target (17, 18).

We have demonstrated a significant learning curve for the set-up time and a moderate but insignificant learning curve for anastomosis and console times.

In essence, OT is not a reliable measure of learning, and OT alone may not be a useful indicator of good practice (19). Kassite et al. have evaluated the learning curve (LC) over a 10 year period for RALP in children by adopting a multidimensional approach and accounting for patient complexity factors (20). Using the cumulative sum (CUSUM) methodology analyzing multi-outcome parameter, they showed that there were three separate phases for the LC in RALP. Initially, the surgeon was in a learning phase which was followed by a consolidation period and finally progression to increased competence, whereby a significantly shorter operation with reduced hospital stay and less postoperative pain was demonstrated with increased surgical experience.

The role of comfort should not be underestimated in lengthy complex laparoscopic procedures. Pediatric laparoscopic surgeons, who work in very bad ergonomic environments due to the small size of their patients, may have a further increase in static postural stress (21). In 2013, a multicenter survey confirmed a strong association between work-related musculoskeletal symptoms (mainly shoulder symptoms) and the number of laparoscopic procedures performed. Skilled laparoscopic surgeons had more pain than less skilled laparoscopic surgeons. These symptoms were more frequent after laparoscopy than after open procedures. Approximately one in four (27%) of these pediatric laparoscopic surgeons didn't sleep well-due to pain which could potentially hinder surgical performance (22). With the robot, the surgeon is in a more comfortable working position.

Retroperitoneal RALP in children is feasible, safe and effective. Developing competence and ease with bulky instruments poses a significant learning curve on the surgeon. Longer term follow up as well as continued practice will allow surgical mastery and address any challenges associated with this procedure which is still in its early days.

The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

This retrospective analysis of prospectively collected data received approval from an independent ethics committee (Comité de Protection des Personnes, CPP Ile de France VII). ClinicalTrials.gov identifier NCT03274050.

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.