Robotic Ureteral Reconstruction Distal to the Ureteropelvic Junction: A Large Single Institution Clinical Series with Short-Term Follow Up

Introduction

Ureteral reconstruction for a variety of benign and malignant conditions has traditionally been performed using open surgical techniques. Initial experience with laparoscopic ureteral reconstruction was first described in the early 1990s. ^1,2 Much of the initial experience was highlighted by laparoscopic repair of ureteropelvic junction obstruction (UPJO) and distal ureteral reconstruction, with larger laparoscopic series showing results comparable to open surgery. ^{3 –6} Despite these initial successful series, laparoscopic ureteral reconstruction was technically challenging, which limited its universal applicability. ^7,8

With the introduction of robotic-assisted technologies, some of the limitations of conventional laparoscopic surgery could be overcome, with the most significant factor being easier intracorporeal suturing facilitated by the robotic Endowrist^®. Initial urologic experience with robotics was largely limited to prostatectomy, and now, around 80% of prostatectomies in the United States are performed robotically. ⁹ As surgeon comfort and experience with robotic prostatectomy grew, robotic technology was adapted to pyeloplasty, ^10,11 followed thereafter by reports of robotic ureteroureterostomy, ureterolysis, ureteroneocystostomy, ureterocalicostomy, and others. ^{12 –14} Most series of robotic ureteral reconstruction in adults have been limited to studies on pyeloplasty or small cohorts of patients undergoing distal ureteral reconstruction. ^{15 –20}

Here, we present our single-institution experience with robotic ureteral reconstruction distal to the UPJ for benign and malignant conditions.

Patients/Methods/Technique

With approval from the Institutional Review Board, we created a retrospective database of all patients at our institution who underwent robotic surgery on the ureter distal to the UPJ. Current procedural terminology (CPT) codes used to generate the list of patients are supplied in Appendix Table A1. Pediatric patients and planned open procedures were excluded. Demographic information included age, gender, body mass index, and affected side. Causative factors consisting of benign (iatrogenic, traumatic, and idiopathic) and malignant etiology were noted as well as interval from date of injury, if known, and presence of a nephrostomy tube at the time of surgery. Perioperative data extracted included total operative time, concurrent endoscopy, estimated blood loss (EBL), intraoperative and postoperative complications, hospital length of stay, and pathological margin status. The follow-up period was defined as the furthest date of imaging or clinical visit postoperatively. Statistical comparisons of operative times were performed using Mann–Whitney U-tests.

Patient, robot, and port positioning varied depending on the suspected level of obstruction. ²¹ In general, for distal lesions, a standard robotic prostatectomy template with a slightly more cephalad port placement was used. The third operative arm of the robot was placed on the ipsilateral side of intervention, if needed. Alternatively, the robotic cart could be positioned obliquely over the ipsilateral hip. For proximal ureteral procedures, patient positioning was similar to that for standard pyeloplasty or partial nephrectomy. Any indwelling nephrostomy tubes were accessible intraoperatively to aid in ureteral identification with an injection of saline or indocyanine green fluorescent dye, if needed.

Endoscopy may be performed in the same operative setting as the formal repair. Cystoscopy with fluoroscopic guidance before positioning the patient for the robotic portion of the procedure enables a wire or ureteral catheter to be placed, or a retrograde pyelogram to be performed. Typically, an open-ended ureteral catheter is placed that can then be accessed by the bedside assistant during the case to pass wires, retrograde stents, or a flexible ureteroscope. Just before completion of an anastomosis, the ureteral catheter can be exchanged for a guide wire to enable retrograde placement of a Double-J ureteral stent. The Tile Pro^® feature of the da Vinci robot is very useful for facilitating on-table endoscopy, which enables both the robotic and bedside surgeons to see both the laparoscopic and endoscopic video concurrently. Flexible ureterorenoscopy can also be performed by passing a flexible scope through an existing trocar, and proximal ureteral or renal stones can be extracted in this manner as well. ²²

Results

Patient demographics are presented in Table 1, and Table 2 lists the breakdown of the various procedures performed. All 55 cases were completed robotically with no conversions to open and no aborted procedures. For the 22 patients with a known date of injury, the median time to repair was 91 days. The earliest repair was 6 days after a robotic hysterectomy complication, and the longest time was more than 6 years for gunshot wound injury managed with chronic nephrostomy tubes. Perioperative metrics are presented in Table 3. The median total operative time was 233 minutes. Any endoscopy utilization increased the median operating room (OR) time by 32 minutes; 234 with versus 202 without (p=0.0775). The median OR time for cases involving malignant etiologies was 224 minutes, which was equivalent to the median time for benign cases of 221 minutes (p=0.8958). The median EBL was 50 mL, and there were no intraoperative transfusions. Two patients required a transfusion postoperatively (Clavien II). ²³ For upper tract malignancy, all surgical margins were negative.

There were no intraoperative complications, but two serious (Clavien≥III) postoperative complications (3.6%) occurred. One patient with multiple medical co-morbidities who underwent a Boari flap and lysis of adhesions for greater than one hour due to previous surgical treatment for diverticulitis experienced an acute oxygen desaturation on postoperative day (POD) 2 and was transferred to the surgical ICU (Clavien IV). He responded well with only oxygen supplementation, was transferred back to floor after less than 24 hours in the SICU, and was discharged home on POD 6. Another patient underwent a ureteroneocystostomy for a stricture resulting from stones with no intraoperative complication and minimal blood loss (100 mL). Several hours after his surgery, he was noted to be tachycardic with increased sanguinous drain output. He was transfused and taken back to the OR for exploratory laparoscopy using the same port sites (Clavien IIIb). At that time, an estimated 300 ml of organized clot was evacuated, but no active bleeding was found. Hemostatic agents were placed in the operative field, and he had no further complications until his discharge on POD 2.

Nephrostomy tubes were present preoperatively in 25 patients. In all but one, the nephrostomy was removed before completion of the procedure. In that patient with a solitary kidney, the nephrostomy tube was left postoperatively after a successful proximal ureteroureterostomy as a safety precaution and removed 1 week later after a nephrostogram.

The average length of stay was 1.6 days, with 71% of patients able to be discharged on POD 1. Two patients returned to the referring provider for continued care, and for the remaining 53 patients the median length of follow up from surgery to last imaging was 181 days. Postoperative imaging varied among the cohort, consisting of a computed tomography urogram, renal ultrasound, diuretic renogram, magnetic resonance imaging, intravenous pyelogram, or retrograde pyelogram. Patients with upper tract urothelial carcinoma were followed with imaging and surveillance cystoscopy. A total of four patients developed bladder recurrence during follow up without evidence of ureteral recurrence.

Our overall success rate was 94.7%. We have had three failures to date defined by continued patient symptoms or radiographic evidence of stricture or obstruction requiring reintervention. All three failures occurred in patients treated with a Boari flap. One failure occurred at 24 months after surgery with acute onset flank pain and was successfully treated with ureteroscopy and balloon dilation of an anastomotic stricture. One patient had lingering flank pain and persistent hydronephrosis after stent removal and required replacement. No other interventions were needed, and the stent was permanently removed without further complication. The third patient had recurrence of pain 3 months after her stent was removed after a long Boari flap plus nephropexy. She underwent initial endoscopic dilation but had persistent pain and hydronephrosis, so then elected to undergo a laparoscopic simple nephrectomy.

Discussion

Our institutional experience with adult robotic ureteral reconstruction represents perhaps the largest reported series of such patients at a single institution. It is a heterogeneous group comprising both malignant and benign etiologies. We have been able to reproduce a large variety of open procedures in a minimally invasive fashion utilizing the da Vinci robot, with a high success rate and a low complication rate in short-term follow up. This demonstrates that robotic reconstruction is a viable option for repairing a range of different pathologies and offers a diverse choice of technical procedures to the surgeon so that the repair can be individually tailored for a particular patient's needs. This procedural flexibility is highly valuable when dealing with ureteral reconstructions, particularly in cases with previous attempts at repair.

Our perioperative metrics, complication rate, and early reintervention rates are similar to reported series for both robotic pyeloplasty ²⁴ and robotic distal ureteral repairs for both benign and malignant pathologies. ^{16,17,25 –30} A comparison of our outcomes to those of several other robotic ureteral reconstruction series is presented in Table 4 and demonstrates the ability to maintain similar excellent results across a variety of etiologies and repair techniques throughout our larger series.

Certainly, nephroureterectomy remains the standard treatment for upper tract urothelial carcinoma, but there remain select patients (such as low-grade isolated tumors, competing health issues, etc.) for whom a segmental ureterectomy and reconstruction provides an attractive alternative. Other groups have reported this approach as well with similar results. ²⁸ These patients warrant close follow up and future outcomes reporting to determine the long-term functional and oncologic success. Maintaining negative surgical margins remains crucial when electing to perform a segmental ureterectomy.

Although there was no statistically significant difference in the operative times, benign etiologies were perceived to be more difficult surgical cases. This may be due to bowel adhesions, scar tissue, and peri-ureteral inflammation that often accompany iatrogenic injuries from previous surgeries or urine leaks. These features are less likely to be present in an untreated malignant obstruction and may account for the added complexity.

Iatrogenic injury is a common cause of ureteral obstruction, ³¹ and surgical dogma typically dictates a delay before repair. ^32,33 We have had good success with repairs at even less than 1 week after the injury, and our preference has shifted toward early repair after identification of such injury rather than delaying for several weeks or months.

All three of our failures involved a Boari flap repair. Although there are not enough patients to power any true statistical meaning to this, it bears consideration. Creation of a bladder flap is certainly a more technically challenging procedure than a simple reimplant, and it is possible that surgeon factors at this time or during the anastomosis contribute to the development of a restenosis. Alternatively, since Boari flaps are selected for longer segment or more proximal strictures, the desire to preserve as much ureteral length as possible may lead to performing the flap anastomosis to the ureter that is not fully healthy even if it appears normal. This is an area for us to monitor and report on in the future.

Although the short-term results have been promising, it is early in the follow-up period and the long-term success has yet to be established. The retrospective nature of this series raises the possibility of selection bias factoring into the results, and the creation of the database was dependent on proper coding of CPT codes at the time of the surgery. The relatively large number of patients, single institutional setting, and inclusion of a wide variety of pathologies and repairs are a benefit for this case series report.

Conclusions

Robotic reconstruction of the ureter distal to the UPJ is feasible, safe, and able to achieve excellent short to intermediate outcomes with a low rate of complications when performed by experienced robotic urological surgery teams. Additional studies with larger populations and extended follow up will help further determine the long-term success of these operations.

The wide variety of procedures able to be performed provides great flexibility to the robotic surgeon to choose that procedure which will best suit a particular patient and the underlying pathology.