Perioperative and Functional Outcomes of Robot-assisted Ureteroenteric Reimplantation: A Multicenter Study of Seven Referral Institutions

Take Home Message This is the first multicenter study in the literature that specifically investigates the feasibility of robot-assisted ureteroenteric reimplantation for the management of ureteroenteric strictures. This minimally invasive procedure is feasible and could be an alternative to open revisions.

mo. Of the UESs, 28 (44%) failed an endourological attempt (balloon dilatation/endoureterotomy). The median RUER operative time was 195 (175-269) min. No intraoperative complications or conversions to open approach were reported. Twenty-three (37%) patients had postoperative complications (20 [32%] were minor and three [5%] major). The median length of hospital stay was 3 (1-6) d and readmissions were 5%. After a median follow-up of 19  mo, 84% of cases were stricture free. Lack of prior RT was the only variable associated with better stricture-free survival after RUER (hazard ratio 6.8, 95% confidence interval 1.10-42.00, p = 0.037). The study limitations include its retrospective nature and the small number of patients. Conclusions: RUER is a feasible procedure for the management of UESs. Prospective and larger studies are warranted to prove the safety and efficacy of this technique. Patient summary: In this study, we investigate the feasibility of a novel minimally invasive technique for the management of ureteroenteric strictures. We conclude that robotic reimplantation is a feasible and effective procedure.

Introduction
Radical cystectomy (RC) with pelvic lymph node dissection and ileal urinary diversion is the standard of care for the management of muscle-invasive bladder cancer and refractory non-muscle-invasive bladder cancer [1]. Benign ureteroenteric strictures (UESs) occur in 3-19% of patients according to large cystectomy series and represent the most common cause for reoperations after RC [2][3][4]. Owing to the low success rate of the endourological approach, open ureteroenteric reimplantation is the gold standard treatment for UESs [5]. However, open revisions are challenging surgeries with a non-negligible risk of complications [6]. Recently, utilization of minimally invasive approaches for ureteral reconstructive procedures has increased. Although only few single-center series of robot-assisted ureteroenteric reimplantation (RUER) were reported previously, outcomes are promising [7]. The potential benefits of this procedure could be those related to the minimallyinvasive approach as well as the three-dimensional vision, tremor filtering, ability to suture with precision, and use of indocyanine green (ICG).
Herein, we aimed to report the perioperative and functional outcomes of RUER for the management of UESs after RC in a large series of patients from seven institutions.

Surgical technique
Supine or lithotomy in Trendelenburg position was utilized. Four 8-mm robotic ports were placed transperitoneally in triangulation around the internal end of the ileal conduit or in the same way as in robotassisted radical cystectomy (RARC) in case of a neobladder ( Fig. 1) [8].
One assistant port of 12 mm and another of 5 mm could also be used.
Extensive adhesiolysis was often necessary to identify the key anatomical components. Intraureteral or intradiversion administration of ICG as described by Lee et al. [9] could help identify anatomical landmarks and the site of stricture. Ureterolysis was performed circumferentially toward the urinary diversion. The dilated ureter proximal to the stricture was dissected and transected. The distal extent of the UES was excised for pathology. The intravenous ICG injection could help in identifying healthy distal ureteral tissue [10]. The healthy end of the ureter was spatulated and reimplanted at a new site on the diversion using monofilament absorbable running or interrupted suture (4-0 or 5-0). For unilateral UESs, a Bricker anastomosis was recommended, and for bilateral UESs either a Wallace or a Bricker anastomosis could be used. After completing the posterior half of the anastomosis, a ureteral stent was placed transabdominally across the anastomosis, with a guidewire inserted through the assistant port. The proximal end of the stent was advanced till the renal pelvis, and the distal end was inserted into the urinary diversion under direct visualization. In ileal conduit cases, the stent could be placed retrogradely through a sucker placed in the conduit. A nephroureteral stent was used for ileal conduits, while both nephroureteral and double-J stents could be used for neobladders.

Study outcomes
Data were reviewed for demographics and perioperative outcomes. Previous abdominal/pelvic surgery (apart from RC) or radiotherapy (RT) was reported. We also reviewed patients who failed endourological manage-

Statistical analysis
Categorical variables were reported as frequencies and proportions, and continuous data as medians and interquartile range (IQR

Perioperative characteristics and outcomes
The median time from cystectomy to stricture diagnosis was 5 mo (IQR 3-11 mo). Twenty-eight cases (44%) failed initial endoscopic and/or percutaneous management. The median time between the diagnosis and reimplantation  was 4 mo (IQR 2-12 mo) and the median stricture length was 2 cm (IQR 1-3 cm). The pathological report of the resected ureter showed benign tissue in all cases. The median OT was 195 min (IQR 175-269 min). No intraoperative complications, conversions to open approach, or intraoperative blood transfusions were reported. Two patients received a blood transfusion after the surgery as a consequence of the intraoperative blood loss, but none of them had signs of postoperative active bleeding. Twentythree patients (37%) had postoperative complications, of which 20 (32%) were minor (Clavien I-II) and three (5%) were major (Clavien III). The minor complications comprised urinary tract infections, ileus (managed conservatively), and urine leak (managed with prolonged stenting).

3.3.
Predictors of success Table 3 shows the univariate analysis of the potential predictors of success. Patients with prior abdominal/pelvic RT had a success rate of 60% compared with 86.2% in those without RT (p = 0.12). A multivariate Cox regression analysis for assessing potential variables associated with stricturefree survival after RUER showed that the lack of prior abdominal/pelvic RT was the only significant variable (hazard ratio 6.80, 95% confidence interval 1.10-42.00, p = 0.037; Table 4).

Discussion
To our knowledge, this is the largest series of RUER for UESs after RC. A benign UES is a common complication that can lead to a variety of health conditions and impairment of quality of life. Management of this complication is challenging, with several options described in the literature but without specific recommendations in clinical guidelines. UESs can occur in up to 19% of patients after RC. Previous studies reporting UESs varied in the surgical technique or experience, and the frequency and duration of follow-up [11,12]. Prior abdominal RT is a well-known risk factor for UESs [13]. Our series included both open RC (ORC) and RARC patients. Prospective randomized controlled trials have shown similar perioperative outcomes between ORC and RARC, while a recent prospective study demonstrated no difference in stricture rate [14,15]. However, the association between RARC (ICUD or ECUD) and UESs is still unclear. It was suggested that ECUD could increase the risk of UESs as a consequence of the higher dissection of the ureters compared with ICUD [16]. In contrast, Ahmed et al. [12] found a higher rate of UESs after ICUD, especially during their early learning curve. It has been demonstrated that the endoscopic procedures have poor long-term efficacy for the management of UES [17]. Nevertheless, it is usually used as the initial management in order to avoid the morbidity of an open surgery and the difficulty associated with operating in a previously operated abdomen. Gin et al. [18] suggested that prior endoscopic attempts could worsen the outcomes during subsequent ureteroenteric reimplantation and recommended performing a primary open/robotic revision. Although prior failed endoscopic attempts were not a predictor of recurrence in our study, >50% of our patients underwent a primary RUER. Our findings suggest that, currently, the primary robotic revision of a UES could be increasing as a result of its low morbidity in the hands of experienced surgeons.  As a consequence of the low efficacy of conservative procedures, open revision is considered the gold standard treatment for UESs [17]. The success rate of open surgical revision is up to 85% [5,18,19], although it could decrease to 78% at a longer-term follow-up [20]. The intraabdominal adhesions with the risk of bowel injury and the surrounding major vessels around the stricture area could deter surgeons from performing this procedure. With the development of minimally invasive techniques, multiple small single-center series of RUER have shown a success rate of up to 80% [7,9,12,21]. A unique case of successful RUER performed by using a pure single-site approach (daVinci SP) and the initial experience of 11 laparoscopic ureteroenteric reimplantations have been also reported [22,23]. However, the role of robotic surgery in the management of UESs is still undefined, and there is a lack of studies comparing the outcomes between open and robotic revisions. We found that 84% were stricture free 19 mo after RUER, demonstrating that this approach could be a reasonable alternative to open surgery. Our work is the first multicenter study in the literature that specifically investigates the feasibility of this procedure in a relatively large series of patients.
Several procedures [18]. The other showed comparable perioperative outcomes between six open and 16 robotic approaches [12]. A small number of robotic procedures were included in both studies. In contrast, a recent series of eight patients treated with RUER reported five cases with postoperative complications. Despite the small sample, the authors concluded that ureteroenteric reimplantation might be a morbid procedure regardless of the operative approach [9]. In our series, postoperative complications occurred in 37% and only three (5%) were major. These results support the belief that although RUER is a morbid surgery, the risk of high-grade complications is lower than in the open approach. RUER has also been associated with a shorter LOS than open reimplantation (3 vs 6-10 d) [5,16,18]. In the current study, the median LOS was 3 d with a readmission rate of 5%. However, the low number of patients, and the variety in population and surgeons might decrease the generalizability and strength of our findings.
Identification of the ureters and urinary diversion in UES patients could be challenging because of the intraperitoneal adhesions and periureteral fibrosis. The use of ICG (nontoxic tracer) could facilitate the identification of the ureter and the site of the stricture. Two studies reported the outcomes of eight and ten patients, who underwent RUER with the use of ICG injected in the ureter. They suggested that this approach provided a quick and precise identification of the ureter and could reduce the risk of vascular and bowel injuries [7,9]. Although ICG was injected into the ureter only in 17% of our procedures, we did not report any case of conversion to open approach or intraoperative complication in the overall cohort. On the contrary, a UES is most commonly attributed to compromised vascularity of the ureter. A recent study reported a significant reduction in the UES rate following the implementation of intravenous (IV) ICG dur- ing RARC [10]. Even though we agree that this maneuver is useful for the assessment of distal ureter vascularization, we did not find a higher success rate in patients who received intravenous ICG. However, our results have several limitations, including the low number of patients who received IV ICG (10%) and that they were all from the same institution. ICG is helpful in assessing the vascularity only when injected IV rather than intraureteral. Prior intraureteral ICG may preclude its benefit to assess vascularity.
Limitations of robotic surgery remain the difficulty of access in a previously operated abdomen and the lack of tactile feedback. Dangle and Abaza [21] suggested that patients who had undergone previous RARC rather than ORC may had reduced intraperitoneal adhesions, and this in fact could facilitate ureteroenteric reimplantation. Our series included mainly patients having undergone minimally invasive RC, and the reported results are hardly comparable with other results available in the literature. Nevertheless, although nine and ten patients of our series were previously treated with ORC and RARC with ECUD, respectively, none of them had to be converted to an open approach or had intraoperative complications. We believe that RUER is feasible with either approach to RC.
In accordance with previous studies of open revisions, we found that the success rate of RUER did not differ by reimplantation side, failed endoscopic attempt, type of urinary diversion, and previous abdominal surgery [5]. One study reported worse outcomes in open revisions among patients who had a preoperative nephroureteral catheter than in those with percutaneous nephrostomy, in contrast to our findings [6]. We used both univariate analysis and multivariate Cox regression analysis due to the low event rate. On multivariate analysis, lack of prior abdominal RT was the only variable significantly associated with being stricture free after RUER. Given the difficulty of operating in a previously irradiated abdomen and the lower success rate, patients should be counseled about the risk-benefits of undergoing RUER. It has been suggested that successful endourological treatment for stricture depends primarily on stricture length [5]. This finding was not confirmed in prior open or robotic series, and was not evaluated in the current study as it was missing for many patients.
To our knowledge, this is the largest report of RUER. However, several limitations exist. The retrospective nature has its known limitations. Another limitation is the variation among institutions in surgical volume, experience, surgical technique, and follow-up protocols. However, all the procedures were performed at high-volume centers using high-volume robotic surgeons who followed similar standardized surgical techniques. Furthermore, our study is strengthened by the fact that it is difficult to generate a large single-institutional series of RUER with optimal follow-up. The relatively small number of patients and the variety in population can limit generalizability of the results.

Conclusions
RUER is a feasible procedure for the management of UESs. Prospective and larger studies are warranted to prove the safety and efficacy of this technique.
Author contributions: Enrique Trilla had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Analysis and interpretation of data: Carrion, Lozano.