Prospective clinical trial of the feasibility and safety of modified retroperitoneal lymph node dissection at time of nephroureterectomy for upper tract urothelial carcinoma


Philippe E. Spiess, Department of Genitourinary Oncology, Moffitt Cancer Center, 12902 Magnolia Drive, Office 12538, Tampa, FL 33612, USA. e-mail:


Study Type – Therapy (cohort)

Level of Evidence 2b

What's known on the subject? and What does the study add?

Very little is known about the safety and potential oncological benefit of performing a retroperitoneal lymph node dissection at time of nephroureterectomy for upper tract tumours.

This study is the first clinical trial to prospectively validate the safety and feasibility of a retroperitoneal lymph node dissection at time of nephroureterectomy for upper tract tumours. The onus is now on the scientific community at large to conduct adequately powered multicentre clinical trials to evaluate the potential oncological benefit it may impart to patients with upper tract tumours.


  • • To determine the safety and feasibility of modified retroperitoneal lymph node dissection (RPLND) at the time of radical nephroureterectomy (RNU).


  • • Between 2009 and 2011, 20 patients with suspected upper urinary tract urothelial carcinoma (UUT-UC) underwent open (n= 10), laparoscopic (n= 4), or robot-assisted (n= 6) RNU with modified RPLND.
  • • Demographic, clinical and pathological data, histological nodal status, peri-operative complications and recurrence data were collected.


  • • On histopathological review, one patient had a benign angioma and was excluded from the final data analysis. Of the remaining 19 patients, 10 had pTa, five had pT1, one had pT2, and three pT3 disease.
  • • The mean (range) lymph node count was 7 (2–17), with one patient having pathologically proven lymph node metastasis. The mean (range) operating time was 279 (146–500) min. The mean EBL was 396 (100–1100) mL, with the mean (range) hospital stay 7.1 (4–18) days. The mean (range) duration of follow-up after surgery was 12 (2–24) months.
  • • Overall, nine patients developed postoperative complications, which included eight minor (Clavien Grade I-II) and one major complication (Clavien grade IIIb). The major complication was a postoperative chylous lymphatic leak requiring surgical exploration.


  • • The present results indicate that modified RPLND during RNU for UUT-UC is a feasible procedure with acceptable morbidity. A larger prospective clinical trial is needed to adequately assess its potential therapeutic benefit.

retroperitoneal lymph node dissection


radical nephroureterectomy


upper urinary tract urothelial carcinoma


Primary upper urinary tract urothelial carcinoma (UUT-UC) accounts for ≈5% of all urothelial neoplasms with a tendency for multifocality and frequent recurrences [1]. About 30% of patients have tumours invading the musculature of the renal pelvis or ureter at presentation and 30–40% have involvement of regional lymph nodes at time of surgery [2,3]. Nodal metastasis is an adverse prognostic indicator and results in poor outcome, irrespective of the use of systemic chemotherapy or radiation [4]. The ‘gold standard’ treatment for invasive UUT-UC is radical nephroureterectomy (RNU) with bladder cuff removal.

Extended pelvic lymph node dissection in urothelial carcinoma of the bladder provides staging and, in selected patients, a survival benefit [5,6]. By extrapolation, it would seem intuitive that a regional lymphadenectomy is as important in UUT-UC because both are histologically of a similar phenotype. However, the role of retroperitoneal lymph node dissection (RPLND) in UUT-UC remains highly debated. Some authors have suggested that it could provide important clinical staging and therapeutic benefit in patients with UUT-UC [7–11]. By contrast, there are studies showing that even an extensive lymph node dissection does not affect patient survival [12]. Similarly, no prospective studies have standardized the ideal extent of RPLND or the optimum number of total lymph nodes that should be removed in patients with UUT-UC. Some investigators believe that the template has a more important impact on survival than the total number of lymph nodes removed [9]. Hence, controversies exist and it is difficult to draw definitive conclusions about the therapeutic benefit achieved by performing a modified RPLND in the clinical context of UUT-UC based on previous retrospective studies.

The aim of the present study was to determine the safety and feasibility of performing a modified template RPLND at the time of RNU for UUT-UC in a prospectively collected cohort of patients. Based on previous reports from bladder UC, we also hypothesized that, in patients with UUT-UC, RNU with modified RPLND has a potential benefit over RNU alone in terms of cancer-specific outcomes.


The aim of the present prospective single-arm clinical trial was to evaluate the safety of performing a modified RPLND in patients with primary UUT-UCC at the time of open or minimally invasive RNU with bladder cuff excision. After institutional review board approval, 20 consecutive patients with primary UUT-UCC underwent this procedure at our centre between March 2009 and April 2011. The sample size was established before conducting the present study.

Preoperative staging evaluation included a history, physical examination, serological studies including liver function tests, chest x-ray or CT scan of the thorax, and a CT scan or MRI of the abdomen and pelvis (with i.v. contrast). Patients with suspicious lymph node(s) >1 cm in largest diameter or distant metastasis on preoperative radiographic imaging were excluded from this study. Patients with suspected locally advanced disease (clinical stage ≥ T3) are routinely recommended to have neoadjuvant systemic chemotherapy at our institution. Demographic data, clinical stage, pathological stage including lymph node status and length of hospital stay were recorded. The Clavien system was used to grade postoperative complications occurring within 30 days of surgery [13]. Postoperative data including ensuing recurrence, use of adjuvant/salvage chemotherapy and disease status at last follow-up were collected.


All patients underwent RNU with bladder cuff excision with concomitant modified RPLND performed by one of three genitourinary surgeons at our tertiary care referral centre. The approach was open in nine cases, laparoscopic in four, and robot-assisted in seven. Patients who had no previous abdominal surgeries or contra-indications to laparoscopic surgery (e.g. severe chronic obstructive pulmonary disease) were offered a minimally invasive approach. In 17 cases, the distal ureterectomy was by endoscopic resection with antegrade ‘pluck’ technique. In three cases (two open and one laparoscopic), the distal ureterectomy was by the open bladder cuff approach. Lymph nodes were usually submitted for frozen section, with areas of resection sent separately according to the surgical templates, but this was left at the surgeon's discretion.


The anatomical boundaries of the lymph node dissection were defined by the ipsilateral side of UUT-UC. In patients with right-sided UUT-UC, the template of dissection consisted of (i) right perihilar lymph nodes, (ii) paracaval lymph nodes, and (iii) right pelvic lymph nodes (common, external and obturator lymph nodes [Fig. 1]). In patients with left-sided UUT-UC, the template of dissection included (i) left perihilar lymph nodes, (ii) para-aortic lymph nodes, and (iii) left pelvic lymph nodes (common, external and obturator lymph nodes [Fig. 2]). Removal of interaortocaval lymph nodes was left to the discretion of the individual surgeon, depending on the presence of positive paracaval lymph nodes (right) or para-aortic lymph nodes (left) as determined on intra-operative frozen section or based on the suspicion of lymphadenopathy within the interaortocaval region at time of surgery.

Figure 1.

Schematic diagram of a surgical template for modified RPLND in patients with right-sided UUT-UC.

Figure 2.

Schematic diagram of a surgical template for modified RPLND in patients with left-sided UUT-UC.


Lymph nodes were frequently submitted for frozen section with areas of resection sent separately according to the surgical templates but this again was left to the discretion of the treating genitourinary surgeon. All surgical specimens were processed according to standard procedures and reviewed by a single experienced genitourinary pathologist. Pathological staging was performed according to the 2002 American Joint Committee on Cancer staging system and grading using the 1998 WHO classification [14,15].


Patients were assessed postoperatively by history and physical examination, urinary cytology, and cystoscopy every 3 months for the first 2 years. Radiographic studies including chest x-ray and abdomino-pelvic CT scan or MRI were performed every 6 months for the first 2 years and yearly thereafter.


Descriptive statistics were applied for summarizing the study results. For continuous variables, mean, sd and range data are shown and, for categorical variables, we show the frequency.


Baseline characteristics of our patient cohort are shown in Table 1. The mean (range) patient age at diagnosis was 69 (42–85) years. The tumour was located in the renal pelvis and proximal ureter in 15 and in the mid to distal ureter in four patients. On final pathology review, 10 patients had pTa disease, five had pT1, one had pT2, and three had pT3 disease. Ten patients had low grade and nine had high grade tumours. Two patients had an interaortocaval node dissection. The median number of lymph nodes was 5 in the paracaval/para-aortic region, 2 in the pelvic nodes and 6 in the interaortocaval region. The mean (range) total lymph node count was 7 (2–17), with one patient having pathologically proven lymph node metastasis at the time of RNU and modified RPLND.

Table 1. Clinical and pathological characteristics of the study population
Variable, N= 19 
Mean (range) age at diagnosis in years69 (42–85)
Clinical stage, n 
Pathological stage, n 
Pathological grade, n 
 Low grade10
 High grade9
Nodal Status, n 
Tumour location, n 
 Renal pelvis/upper ureter15
 Mid to distal ureter4


The mean (range) operating time was 279 (146–500) min and the mean (range) intra-operative blood loss was 396 (100–1100) mL, as shown in Table 2. The mean hospital stay was 7 (4–18) days. The mean (range) follow-up from time of surgery was 12 (2–24) months.

Table 2. Surgical outcomes of the patient population
Variables, N= 19Mean (range)
Total lymph node count7 (2–17)
Open (n= 9)7 (2–17)
Laparoscopic (n= 4)8 (3–14)
Robot-assisted (n= 7)6 (2–10)
Operating Room (OR) time, min279 (146–500)
Open (n= 9)218 (146–309)
Laparoscopic (n= 4)263 (208–338)
Robot-assisted (n= 7)357 (273–500)
Intraoperative blood loss, mL396 (100–1100)
Open (n= 9)556 (200–1100)
Laparoscopic (n= 4)143 (100–200)
Robot-assisted (n= 7)233 (100–500)
Total hospital stay in days7 (4–18)
Open (n= 9)7 (4–16)
Laparoscopic (n= 4)7 (5–9)
Robot-assisted (n= 7)7 (4–18)


There were no injuries to major vessels at the time of RNU with bladder cuff excision and modified RPLND; however, there were two bowel-related injuries during lysis of adhesions repaired intra-operatively with no further sequelae. All other complications occurred in the postoperative period and were classified according to the Clavien grading system as shown in Table 3. In all, nine patients developed postoperative complications. Of these, eight were minor complications (Clavien Grade I–II) including one patient who developed a prolonged postoperative ileus that required 2 weeks to resolve (Clavien grade II) and six patients requiring blood transfusions in the postoperative period (Clavien grade II). Importantly, only one patient had a major complication, which consisted of a postoperative chylous lymphatic leak requiring subsequent surgical exploration and lymphatic clipping/ligation (Clavien grade IIIb).

Table 3. Incidence of postoperative complications using Clavien grading system
CLAVIEN GRADINGSpecific complicationNumber of cases, N= 9
IIBlood transfusion6
Chylous leak managed conservatively1
IIIbChylous leak requiring surgical exploration1


Follow-up data were available for all patients. At last follow-up, there were no mortalities. Seventeen patients had no evidence of disease at last follow-up. Two patients had metastatic disease: one to regional lymph nodes, noted at the time of modified RPLND with concomitant RNU and bladder cuff excision, and one to the lungs, as detected on postoperative surveillance imaging. There were eight bladder tumour recurrences and two recurrences in the opposite renal pelvis and ureter. The median time to recurrence was 4 months. Patients with bladder recurrences underwent transurethral bladder tumour resections. One patient with a multifocal low grade Ta recurrence received a single dose of postoperative intravesical mitomycin immediately after the transurethral resection. Two patients with high grade T1 recurrences in the bladder were treated with an induction course of intravesical BCG therapy (once weekly × 6) after the transurethral resection. One patient, who had a recurrence in the contralateral renal pelvis in addition to multifocal high grade T1 bladder recurrences, underwent RNU of the opposite kidney along with an en bloc radical cystectomy and pelvic lymph node dissection. In total, six patients received postoperative systemic chemotherapy. Three of these patients were administered salvage chemotherapy: two for recurrences in the contralateral (solitary) kidney renal pelvis and one for metastatic progression to the lungs. Three patients received adjuvant systemic chemotherapy: two for locally advanced UUT-UC (pT3 disease) and one for lymph node metastasis as detected at the time of RNU and modified RPLND.


The present phase II clinical trial is the first to validate, in a prospective manner, the safety and feasibility of performing a modified RPLND at the time of RNU for UUT-UC, with our major complication rate (Clavien grade 3 or more) being only 1/19 patients. The European Organization for Research and Treatment of Cancer 30881 trial [16] found that lymphadenectomy did not affect the surgical complication rate but that trial compared radical nephrectomy and concomitant modified RPLND with radical nephrectomy alone. The present study was performed for urothelial cancer and establishes a baseline for the reported complication of a concomitant lymph node dissection conducted in this setting and in a prospective designed trial.

The standard treatment for UUT-UCC is RNU and bladder cuff removal. Recent data would suggest the potential clinical merit of an RPLND in patients with UUT-UCC [12,17,18]. There are, however, no data specifically addressing the added morbidity and peri-operative risk of performing a modified RPLND for UUT-UC. Extrapolating from data in patients with stage I testicular cancer undergoing primary RPLND provides some insight into the potential morbidity of this procedure [19]. Major complications from primary RPLND range from 1 to 5% and minor complications from 15 to 20%. They include vascular injuries to the aorta and/or the vena cava, injuries to solid organs such as the liver, kidney and pancreas, prolonged ileus, and lymphatic leaks [19–22].

The overall postoperative complication rate in the present series was 9/19 patients, with eight minor complications (Clavien Grade I–II) and just one major complication (Clavien Grade III or more). These rates are similar to previous open and laparoscopic RNU series conducted in the absence of a concomitant modified RPLND [23–25]. Of these complications, only chylous lymphatic leaks can be directly attributable to the modified RPLND component of the procedure.

In the present series, most complications were minor (Clavien Grade I-II), including one of the two cases of postoperative chylous ascites which resolved with conservative measures consisting of fat restriction and the administration of medium chain triglycerides. Our incidence of this complication is comparable with other reports in which the incidence of chylous lymphatic leaks after RPLND was between 2% and 11% [19–22]. One of our patients required surgical re-exploration for meticulous lymphatic ligation, which represents a major complication (Clavien grade IIIb). There were no statistical differences in the complication rate with regard to the surgical approach selected in an individual case. In addition, it is interesting to note that there were no differences in the mean of length of stay after surgery, irrespective of the surgical approach, which we suspect results from a similar return in bowel function and inability to be rapidly advanced to a regular diet after surgery, regardless of the approach.

There have been several retrospective studies addressing the role of RPLND in UUT-UC. Komatsu et al.[8] reported that RPLND offers a staging but not a prognostic benefit in UUT-UC. Roscigno et al.[18] conducted a large multicentre study of 412 patients in which they concluded that RPLND improved clinical staging in invasive UUT-UC. The study by Brausi et al.[11] also suggested a curative impact of RPLND in patients with infiltrative UUT-UC. Extrapolating from the scientific literature pertaining to bladder cancer, an RPLND provides not only pathological staging and prognostic information but also offers a curative potential and hence survival benefit [6,7,26,27] . Thus, we can hypothesize that a modified template RPLND at the time of RNU for UUT-UCC could offer a similar staging and therapeutic advantage to that clearly established for bladder urothelial carcinoma which was the premise for the present prospective clinical trial.

In a previous study from the Memorial Sloan Kettering Cancer Centre, no preoperative clinical variable, including radiological imaging, cytological, and/or biopsy findings, could reliably predict those patients with lymph node metastasis [28]. CT imaging of the abdomen and pelvis did not demonstrate clinically significant lymphadenopathy in 40% of patients with pathologically proven lymph node metastases. The authors therefore argue that a RPLND should be conducted in all patients with UUT-UC given the lack of accurate preoperative factors predicting lymph node metastasis. However, many studies have reported the correlation between the presence of lymph node metastasis with tumour stage and grade. Margulis et al.[29] reported that the incidence of lymph node metastasis sequentially increased with advancing stage (<1% for T0/Ta/Tis, 8% for T2, 17% for T3, and 46% for T4, P < 0.001) and grade (15% for high grade vs 2% for low grade, P < 0.001). Similar findings were reported by Kondo et al.[9] (0%, 5%, 24%, and 84% for Tis-1, T2, T3, and T4 tumours, respectively; and 0%, 11%, and 35% for G1, G2, and G3 tumours, respectively).

In the present series, 50% of the patients had superficial tumours and only four (20%) had invasive tumours (pT2-T3). One of our patients had a nodal metastasis at the time of modified RPLND but only a minute focus of high grade T1 UUT-UC was present in the final RNU pathological specimen. The low incidence of nodal metastasis in the present series could therefore result from the majority of our patients exhibiting low grade and non-invasive tumours.

The present study has some limitations. The sample size was small because it was a single institution study. It was thus underpowered to achieve a cancer-specific endpoint pertaining to the oncological merit of performing a modified RPLND in the context of UUT-UC. In addition, patients with locally advanced UUT-UC (clinical stage ≥ T3) were routinely recommended neoadjuvant systemic chemotherapy before surgical resection at our institution, resulting in an inherent selection bias. The present study still constitutes the first prospective clinical trial assessing the safety and feasibility of conducting a modified RPLND at the time of RNU.

In conclusion, the present, prospective phase II clinical trial shows that a modified RPLND at the time of RNU and bladder cuff excision is feasible with acceptable morbidity. The gauntlet is passed to national and international collaborative groups/societies to perform a large prospective clinical trial to adequately assess the potential oncological benefit of performing a modified RPLND at the time of RNU for UUT-UC.


Wade J. Sexton is a Study Investigator funded by Endo Pharmaceuticals.