Upper urinary tract tumour after radical cystectomy for transitional cell carcinoma of the bladder: an update on the risk factors, surveillance regimens and treatments


  • Kristin M. Sanderson,

    1. Urologic Oncology, Department of Urology, Keck School of Medicine, University of Southern California, USC/Norris Cancer Center, Los Angeles, CA, USA, and *Department of Urology of la Pitié-Salpêtrière Hospital, GHU Est, AP-HP, Faculté de Médecine Pierre et Marie Curie, University Paris VI, and CeRePP group, EA 3104, University Paris VII, France
    Search for more papers by this author
  • Morgan Rouprêt

    Corresponding author
    1. Urologic Oncology, Department of Urology, Keck School of Medicine, University of Southern California, USC/Norris Cancer Center, Los Angeles, CA, USA, and *Department of Urology of la Pitié-Salpêtrière Hospital, GHU Est, AP-HP, Faculté de Médecine Pierre et Marie Curie, University Paris VI, and CeRePP group, EA 3104, University Paris VII, France
    Search for more papers by this author

Morgan Rouprêt, Hôpital Pitié-Salpétrière, 47–83 Boulevard de l’hôpital, 75013 Paris, France.
e-mail: morgan.roupret@psl.aphp.fr


The mini reviews this month cover two somewhat unusual topics, but I feel certain that they will be of interest to the readership. Metachronous TCC in the upper urinary tract after radical cystectomy has been an important topic in relation to the follow-up, and this paper throws considerable light on the subject. The second paper is an excellent review of the literature on emphysematous cystitis.

Urothelial carcinoma is characterized by multiple, multifocal recurrences throughout the genitourinary tract; ≈ 3% of patients treated by radical cystectomy (RC) for invasive transitional cell carcinoma (TCC) of the bladder will subsequently develop a subsequent TCC in the upper urinary tract (UUT) urothelium. Metachronous upper UUT tumours (mUUT-TCC) typically occur as a late oncological event (>3 years after RC). The vast majority of mUUT-TCCs are detected only after the progression to tumour-related symptoms, e.g. haematuria, flank pain or pyelonephritis, despite strict adherence to surveillance protocols. Failure of imaging and cytology to detect most asymptomatic tumours has led to questions about the need for routine UUT surveillance. Some authors have advocated a more tailored approach to surveillance after RC, targeting high-risk patients and with limiting imaging in those patients at lowest risk of developing a subsequent UUT-TCC. mUUT-TCCs are most common in patients with TCC in the ureter or urethra, and with organ-confined bladder cancer. Although the prognosis is generally poor, long-term survival can be achieved in a subset of patients after radical nephroureterectomy (NU). Minimally invasive techniques, e.g. ureteroscopic and percutaneous resection, have been proposed as renal-sparing alternatives to radical surgery for patients with low-stage and -grade de novo UUT-TCC. However, oncological control of renal-sparing therapies in those with high-risk mUUT-TCC remains largely unconfirmed. Until oncological outcomes equivalent to the standard, radical NU, are reported in patients after RC, conservative treatment strategies should be avoided.


(metachronous) upper urinary tract


radical cystectomy


(radical) nephroureterectomy


frozen-section analysis


carcinoma in situ


lymph node dissection.


Urothelial carcinoma is thought to be associated with a pan-urothelial ‘field defect’ characterized by frequent, multifocal metachronous tumours. Isolated TCC of the upper urinary tract (UUT-TCC) in patients with a history of bladder cancer is consistent with this hypothesis. Two distinct patterns of oncological relapse emerge after radical cystectomy (RC), i.e. ‘early’ and ‘late’ tumour relapse, occurring after <3 or >3 years of disease-free survival. Early tumour relapse accounts for >80% of all tumour recurrences after RC and are located outside the urothelium, primarily within the bony pelvis or viscera [1,2]. Despite detecting the vast majority of early recurrences through routine surveillance, survival in this cohort remains poor, despite surgical intervention or adjuvant systemic therapy [1]. By contrast, late oncological recurrences are most commonly in the UUT and detected only after developing tumour-related symptoms (i.e. primarily gross haematuria and flank pain), despite routine surveillance [1–4]. Importantly, the potential to achieve prolonged survival after surgical resection for late TCC recurrence is significantly better than for early recurrences [2]. In this short review, we describe the main prognostic factors associated with metachronous UUT-TCCs after RC and update current surveillance and management strategies.


The goal of ongoing surveillance is to detect tumour recurrence before the development of widespread systemic disease, when the tumour is most amenable to adjuvant therapy. UUT surveillance is principally by imaging and urinary cytology, although the optimum duration, interval and method of surveillance have not yet been established (Table 1) [1–5]. Metachronous (m) UUT-TCCs detected in asymptomatic patients account for a mean (range) of 38 (0–56)% of all cases [1,3–8] (Table 2) [1,3–10]. The remaining mUUT- TCCs are identified only after development of tumour-related symptoms, e.g. haematuria (accounting for 60–80%), flank pain, pyelonephritis and weight loss [5,8,11]. Although some evidence indicates that tumours identified via radiographic studies are of a lower tumour stage than those presenting symptomatically, further analysis is required to determine whether this translates into improved survival [6].

Table 1.  Proposed tumour surveillance protocols following radical cystectomy in the contemporary series
ReferenceStage-specificProposed upper tract surveillance protocols
Slaton et al.[1]pT1 + pT2, N0IVU + urine cytology × 12 months
pT3N0 + pTxN1-3CT at 6 and 12 months, × 12 months thereafter; urine cytology × 12 months
Balaji et al.[3]NoneNo routine radiological surveillance; IVU for symptomatic patients
Kenworthy et al.[4]Risk group specificHigh risk (distal ureteric or prostatic TCC involvement); urine cytology × 6 months, IVU × 12 months
Low risk (all others); urine cytology × 12 months, IVU × 24 months
Huguet-Perez et al.[6]StandardIVU + urine cytology × 12 months
Solsana et al.[2]Updated standardCT + urine cytology × 12 months for first 5 years; IVU + urine cytology × 12 months thereafter
Sved et al.[5]Cytology-onlyUrine cytology × 6 months; IVU for symptomatic patients or positive cytology only
Table 2.  Characteristics of mUUT-TCC after RC for invasive bladder cancer, reported in contemporary series
ReferenceYearNumber of UUT-TCCs/ total cohortRate of mUUT-TCC, %Median time from RC to UUT-TCC, monthsUUT-TCCs at surveillance (asymptomatic*), n/N (%)UUT-TCCs presenting symptomatically, n/N (%)
  • *

    Includes detection by urine cytology or radiological studies;

  • †Percentage given for totals >20;

  • ‡Mean; NS, not specified.

Kenworthy et al.[4]199611/4302.6406/115/11
Slaton et al.[1]19999/3822.4255/94/9
Solsana et al.[10]199716/2157.458NSNS
Balaji et al.[3]199916/5293.0377/169/16
Huguet-Perez et al.[6]200126/5684.54611/26 (42)15/26 (58)
Yossepowitch et al.[9]200310/4832.1NSNSNS
Sved et al.[5]20045/2352.0401/54/5
Akkad et al.[7]20062/842.3620/22/2
Sanderson et al.[8]200627/10692.5405/23 (22)18/23 (78)
Overall (range) 122/39953.043 (25–62)35/92 (38)57/92 (62)

Urinary cytology remains the most frequent tool for screening for urothelial carcinoma progression within the urinary tract [12]. Although the specificity of this test is excellent (>95%), the overall sensitivity is relatively poor (50%) and highly dependent on the tumour grade and stage [13]. In addition, urinary cytology is operator-dependent and requires an experienced cytopathologist for reliable interpretation. When used in the surveillance of patients after RC, voided cytology is the primary mode of detection in <27% of all mUUT-TCCs [3–5,7] (Table 2). Once patients present with symptoms suggestive of UUT-TCC, urinary cytology becomes dramatically more reliable for confirming a urothelial tumour relapse. Indeed, in symptomatic patients cytology is suggestive of malignancy in 80–100% of patients with documented mUUT-TCC [3–5,14]. This improved accuracy among symptomatic patients might reflect a more advanced tumour stage and grade in these patients or perhaps a greater vigilance among cytopathologists alerted to the presence of suspicious symptoms.

Alternative molecular-based urine tests, such as nuclear matrix protein-22 and fluorescence in situ hybridization, have become increasingly common in bladder cancer surveillance, because of their greater accuracy in patients not having RC [12]. Although promising for future surveillance of patients after RC, the efficacy of these newer methods for patients with intestinal interposition remains largely untested and will require further assessment before widespread use.


Surveillance using IVU or CT detects 0–55% of all mUUT-TCCs [1,3–7]; although neither method is clearly better for detecting UUT tumours, CT is significantly better at detecting TCC metastases within the retroperitoneum or viscera [1,15]. Importantly, the location of the UUT-TCC directly affects the ability of CT to detect urothelial lesions; CT detects 78–94% of documented lesions in the renal pelvis and only 19–53% of tumours within the ureter [15,16]. Small tumours (<2 cm) are also much less likely to be detected by CT [17]. For preoperative tumour staging, the interpretation of CT was accurate in only 56% of mUUT-TCCs, and neither CT nor IVU could discriminate superficial (pTa, Pis, pT1) tumours from invasive pT2 disease [15,16]. MRI has not been shown to provide any significant advantage in imaging of UUT-TCC.

Ureteroscopy for diagnosis and staging

Ureteroscopic evaluation and biopsy has become indispensable in the diagnosis and staging of primary UUT-TCCs, particularly when diagnostic imaging and cytology are equivocal. However, endoscopic access in patients with urinary diversion and bilateral uretero-ileal anastomoses can be particularly challenging, often requiring a combined antegrade and retrograde approach to successfully negotiate the ureteric orifices.

Once the UUT is accessed, the biopsy specimen obtained ureteroscopically is highly accurate at predicting the final pathological tumour grade. Unfortunately, staging of UUT tumours, which requires the presence of lamina propria within the specimen, is less reliable because of the limited ability of ureteroscopic biopsy cups and stone baskets to retrieve adequate tissue samples [9,18]. However, even when lamina propria is identified within the specimen, nearly half of the tumours diagnosed as superficial on biopsy are found to be invasive (≥pT2) in the final nephroureterectomy (NU) specimen [19]. These data reinforce the limited ability of current methods to predict the true pathological stage of UUT tumours before RC.


Only 3% of all patients after RC will develop a mUUT-TCC as the primary site of urothelial progression (Table 2). The median (range) time from RC to diagnosis of a mUUT-TCC is ≈ 43 (25–62) months, with isolated reports of mUUT-TCC after up to 9 years of disease-free survival [1,2,4–7]. Identifying the pathological variables associated with a greater risk of developing a mUUT-TCC might help to better tailor surveillance regimens to those at greatest risk, while limiting unnecessary examination of low-risk patients.

pT stage

Patients with good prognostic factors at the time of RC, e.g. organ-confined (≤ pT2N0) disease, are more likely to develop a secondary urothelial tumour than patients with extravesical (>pT2) or node-positive bladder cancer [2]; 65–100% of mUUT-TCC occur in patients with organ-confined bladder cancer [3–7]. However, pathological stage is not an independent risk factor for mUUT-TCCs, but rather a predictor of prolonged survival after RC, and therefore a greater time at risk to develop a late oncological relapse [2,4,8].

TCC of the transvesical (intramural) ureter

Identifying a tumour within the RC specimen appears to be highly predictive of subsequent recurrence within the upper tracts. Kenworthy et al.[4] reported tumour involvement of the intramural ureter in 55% of their patients who subsequently developed a mUUT-TCC, vs only 8% who never did (P = 0.001). Also, anastomotic TCC recurrence was also independently associated with TCC within the intramural ureter, and this finding was independent of the final ureteric margin status (P < 0.001) [9].

Ureteric margin

The impact of a positive ureteric margin noted on frozen-section analysis (FSA) or in the final pathological specimen has been the subject of debate. High rates of false-negative results and a lack of association between FSA and subsequent UUT-related mortality have fuelled the debate [17]. The most recent evidence indicates that TCC involvement of the ureteric margin is associated with a greater risk of a secondary mUUT-TCC but does not influence the rate of recurrence at the uretero-ileal anastomosis, or overall survival [8,20]. Furthermore, achieving negative FSA margins via sequential resection of the ureters did not eliminate the risk of UUT urothelial tumour relapse [20]. Currently, there is little evidence to support the need for routine intraoperative FSA or the need to ‘clear’ all evidence of microscopic disease before uretero-ileal anastomosis.

Carcinoma in situ (CIS)

CIS is a well-established risk factor for UUT tumour relapse in patients treated for superficial bladder cancer (i.e. Ta, Tis and T1). Conversely, among patients with invasive bladder cancer undergoing RC there was no association between bladder CIS and mUUT-TCC in four contemporary series [3–5,8]. Solsona et al.[10] published the only contemporary RC series showing a significantly higher rate of mUUT-TCCs among patients with bladder cancer treated with RC for CIS-only disease than in those treated for muscle-invasive TCC (P < 0.01). These contradictory data might be explained by the confounding effect of survival disparity between comparison groups. The longer survival of the CIS-only cohort increases the time for detecting late oncological recurrences such as a mUUT-TCC, and results in more events than in those with muscle-invasive disease who have a shorter overall survival. As a result, the detection rate is artificially skewed toward the cohort with prolonged survival. Studies that control for variables that influence survival, e.g. pT stage, are more appropriate for assessing the impact of CIS on the rate of mUUT-TCC. When considered in this light, the bulk of data from well-controlled studies does not support a greater risk of mUUT-TTC in patients with bladder CIS.

Urethral urothelial carcinoma

A multivariate analysis by Sved et al.[5] confirmed prostatic urethral tumour to be an independent risk factor for mUUT-TCC (P = 0.04). More recently, a series confirmed that TCC of the male prostatic ducts or glands (pT1) and TCC of the female urethra were associated four to six times more mUUT-TTC, respectively [8]. Interestingly, TCC invasion of the prostatic stroma (pT4) did not predict any greater risk of mUUT-TCC in men, a result that might reflect the significantly lower survival in men with prostatic pT4 disease [8].

Because there were few patients and the data are limited, risk factors for mUUT-TCC have been difficult to delineate definitively, but the preponderance of evidence indicates that patients with organ-confined disease and ureteric or urethral TCC at the time of cystectomy are at the greatest risk.


Open Radical NU (RNU)

Open RNU, with or without lymph node dissection (LND), is the only treatment associated with extended survival among patients developing a mUUT-TCC after cystectomy [2,7,8]. Of those patients undergoing LND, node-positive rates are 32–83% (Table 3) [1,3–8]. The role of a systematic extended retroperitoneal LND in patients undergoing NU for mUUT-TCC is not well defined. Only three series provide data on the inclusion of LND at the time of NU, and there are too few patients for valid survival analyses (Table 3). However, given the compelling evidence to support the therapeutic, and even curative, benefit of LND at the time of RC for invasive bladder TCC, it is plausible that LND for UUT urothelial carcinoma could provide a similar benefit, particularly among patients with minimal nodal disease.

Table 3.  A review of contemporary series of UUT-TCC after RC; the use of RNU, LND and pathological findings from NU specimens
ReferencePatients treated with RNU, n/N (%)LND at time of NU, n/N (%)LN-positive disease, % or n/NAdvanced pT stage (pT3 or pT4), % or n/N
  1. NS, not specified by the author.

Sanderson et al.[8]24/27 (89)19/24 (80)3254
Huguet-Perez et al.[6]16/26 (61)NSNSNS
Kenworthy et al.[4]7/11NSNSNS
Akkad et al.[7]2/20/2NS1/2
Sved et al.[5]2/51/21/2NS
Slaton et al.[1]NSNSNSNS
Balaji et al.[3]8/166/85/64/6
Overall59/96 (61)26/36 (72)32–83%50–75%

Adjuvant chemotherapy

For node-positive patients and those with advanced tumour stage, the increased use of adjuvant chemotherapy might have contributed to the slightly better survival for primary UUT urothelial carcinoma during the last 20 years. Nevertheless, the final outcomes in advanced UUT-TCC treated with adjuvant chemotherapy remain unsatisfactory, with long-term survival achieved in <5% of patients [2].

Laparoscopic NU

The laparoscopic approach to RNU, with attention to reproducing the oncological standards achieved in open NU, has become a common method of managing UUT tumours. Reports indicate that equivalent oncological outcomes can be attained by open and laparoscopic NU. However, laparoscopic NU for treating UUT-TCCs after RC is more challenging because of the considerable adhesions, alterations in anatomy and the need to enter and reconstitute the intestinal portion of the uretero-ileal anastomosis. Perhaps as a result there are no reports of successful laparoscopic RNU in a patient after RC and urinary diversion, but these might emerge as an alternative approach in the future.

Conservative (renal-sparing) management

Endoscopic management of UUT-TCCs is rapidly gaining acceptance as a standard treatment for patients with low-grade and -stage tumours and concomitant renal insufficiency, bilateral tumours or a solitary kidney. In a comparison of primary UUT-TCC treated by either RNU, endoscopic or percutaneous resection, survival was predicted by tumour grade, stage and location, but not by the type of surgical procedure [11]. The widespread use of percutaneous resection, typically reserved for bulky, low-grade tumours, has been limited by concern about possible seeding and dissemination of tumour cells. Proponents cite renal preservation as an indication for percutaneous resection, but some reports suggest that the rate of successful renal preservation is as low as 30%. Renal loss after percutaneous procedures was attributed to progressive renal insufficiency, postoperative complications and the need for subsequent NU [21].

Because outcome analyses from endoscopic procedures are derived almost exclusively from primary UUT-TCC series, the results might not be relevant in high-risk patients after RC. Patients with mUUT-TCC have a worse overall survival than those with primary UUT-TCC. This disparity in survival could reflect the significantly more advanced stage and grade noted among mUUT-TCC tumours; >75% of mUUT-TCC identified in patients after RC are invasive and high-grade, compared with only 30% of tumours in those with de novo UUT-TCC [3,4,6–8,11,14,18]. However, even after controlling for tumour stage, patients with mUUT-TCC have a significantly worse prognosis than those with no history of invasive bladder cancer [22]. These data indicate that mTCC tumour is inherently more lethal than de novo urothelial carcinoma. Therefore, until further series report their long-term oncological outcomes after conservative management of mUUT-TCC in high-risk patients after RC, the widespread use of renal-sparing methods should be discouraged.


Although most mUUT-TCC series report disappointing 3-year survival rates after NU, of 0–25%, more recent series report long-term survival of >8 years [1,4,14]. Survival after a mUUT-TCC might be predicted by tumour stage, as suggested in the series in which median survival among patients with ≥ pT3 tumours was only 1.3 years, vs 3.4 years for patients with pTa or pT1 tumours [8]. In addition, tumour location affects survival; whereas patients with uretero-ileal anastomotic tumour recurrence have rapid progression to distant metastatic disease and death, those diagnosed with a non-anastomotic mUUT-TCC had a prolonged disease-free interval after radical surgery [9].

Another factor predictive of extended survival is patient eligibility to undergo radical surgery. Balaji et al.[3] reported that while the overall median survival after a diagnosis of the secondary mUUT-TCC was 10 months, for those patients eligible for RNU, the median survival was extended to 26 months. This difference can be explained both by differential performance status and the extent of the metachronous tumour.


mUUT-TCC after RC for invasive bladder cancer typically occurs after several years of disease-free survival. Despite rigorous surveillance, most such patients are identified only after developing symptoms. Those patients at greatest risk of developing a mUUT-TCC as a primary site of disease progression are those with ≤ pT2 bladder cancer and TCC within the urethra or the intravesical portion of the ureter. CT and IVU remain the standard imaging methods for detecting mUUT-TCCs, but are particularly poor for detecting small tumours and those within the ureter. Open RNU is the only treatment confirmed to provide extended survival in these high-risk patients. In the near future, endoscopic treatment might be a more realistic goal in selected patients, but currently remains unproven.


None declared.