Wassim Kassouf, Division of Urology, McGill University Health Center, 1650 Cedar Avenue, Rm L8-315, Montreal, Quebec, Canada H3G 1A4. e-mail: firstname.lastname@example.org
Study Type – Prognosis (cohort)
Level of Evidence 2a
What's known on the subject? and What does the study add?
Radical cystectomy with pelvic lymph node dissection is recognized as the standard of care for carcinoma invading bladder muscle and for refractory non-muscle-invasive bladder cancer. Owing to high recurrence and progression rates, a two-pronged strict surveillance regimen, consisting of both functional and oncological follow-up, has been advocated. It is also well recognized that more aggressive tumours with extravesical disease and node-positive disease recur more frequently and have worse outcomes.
This study adds to the scant body of literature available regarding surveillance strategies after radical cystectomy for bladder cancer. In the absence of any solid evidence supporting the role of strict surveillance regimens, this extensive examination of recurrence patterns in a large multi-institutional project lends further support to the continued use of risk-stratified follow-up and emphasizes the need for earlier strict surveillance in patients with extravesical and node-positive disease.
• To review our data on recurrence patterns after radical cystectomy (RC) for bladder cancer (BC).
• To establish appropriate surveillance protocols.
PATIENTS AND METHODS
• We collected and pooled data from a database of 2287 patients who had undergone RC for BC between 1998 and 2008 in eight different Canadian academic centres.
• Of the 2287 patients, 1890 had complete recurrence information and form the basis of the present study.
• A total of 825 patients (43.6%) developed recurrence.
• According to location, 48.6% of recurrent tumours were distant, 25.2% pelvic, 14.5% retroperitoneal and 11.8% to multiple regions such as pelvic and retroperitoneal or pelvic and distant.
• The median (range) time to recurrence for the entire population was 10.1 (1–192) months with 90 and 97% of all recurrences within 2 and 5 years of RC, respectively.
• According to stage, pTxN+ tumours were more likely to recur than ≥pT3N0 tumours and ≤pT2N0 tumours (5-yr RFS 25% vs. 44% vs. 66% respectively, P < 0.001). Similarly, pTxN+ tumours had a shorter median time to recurrence (9 months, range 1–72 months) than ≥pT3N0 tumours (10 months, range 1–70 months) or ≤pT2N0 tumours (14 months, range 1–192 months, P < 0.001).
• Differences in recurrence patterns after RC suggest the need for varied follow-up protocols for each group.
• We propose a stage-based protocol for surveillance of patients with BC treated with RC that captures most recurrences while limiting over-investigation.
Bladder cancer (BC) is the second most common urological cancer and the sixth most common overall cancer in Canada, with 6700 new cases per year. It has the fourth highest cancer incidence in men and the twelfth in females and represents up to 5.8% of new cancer cases and 3.3% of deaths in Canadian men . Radical cystectomy (RC) with pelvic lymph node dissection remains the gold standard in the treatment of carcinoma invading bladder muscle and refractory non-muscle invasive BC [2–6]. Owing to the high recurrence and progression rates of the disease, a two-pronged strict surveillance regimen consisting of both functional and oncological follow-up has been advocated [7,8]; however, in the absence of any solid evidence supporting their use, these surveillance strategies are currently based on consensus of expert opinions. The aim of the present study, therefore, was to review our data on the recurrence patterns after RC in order to establish appropriate stage-specific surveillance protocols for patients with localized and locally advanced BC.
MATERIALS AND METHODS
Retrospective data from eight academic centres across Canada were collected on 2287 patients who had undergone RC for BC between 1998 and 2008 as part of the Canadian Bladder Cancer Network. After institutional review board approval, clinical and pathological information was obtained from the following institutions: McGill University Health Center, Montreal, QC; Laval University, Quebec City, QC; University of Western Ontario, London, ON; University of Alberta, Edmonton, AB; University of Ottawa, Ottawa, ON; University of Montreal, Montreal, QC; Dalhousie University, Halifax, NS; and University of Manitoba, Winnipeg, MB. Recurrence information was available on 1890 patients.
A standardized template was distributed to all centres and data were entered and tabulated. Separate strict quality control procedures were done both internally and when all the data were collected to ensure optimum accuracy and concordance of data. Similar surgical indications were shared by all institutions and patients who were included were those with clinical evidence of muscle-invasive disease or patients with non-muscle invasive tumours that were either recurrent or refractory to transurethral resection and/or intravesical therapy.
Collected variables included clinical and pathological stage, grade, status of surgical margins, presence of lymphovascular invasion, concomitant carcinoma in situ, disease recurrence, site of recurrence, time to recurrence, and survival. Before surgery, all patients were clinically staged using CT of the abdomen and pelvis, with or without examination under anaesthesia, depending on physician preference. After surgery, all specimens were reviewed by staff pathologists with genito-urinary experience in each institution. Staging was done using 1997 TNM classification and grading, according to the 1998 WHO system.
In general, oncological and functional follow-up consisted of office visits, complete blood counts, serum chemistries, abdominal imaging, and chest radiography every 3–6 months for at least 3 years or until 1st January 2008. Further investigation such as bone scans were ordered when clinically indicated. Time to recurrence was calculated as the time interval from surgical intervention to first evidence of clinical recurrence or last follow-up in the absence of any recurrence. Patients who died without evidence of recurrence were censored at the time of death. Sites of recurrence were classified as local (pelvic), retroperitoneal, and/or distant via confirmation by either imaging and/or biopsy. Distant recurrences were further stratified into liver, lungs, bone and brain. All calculations were made using as a denominator all patients in whom recurrence information was available.
Associations between the different groups in frequency distributions of categorical variables were made using the Fisher and chi-squared tests. Kaplan–Meier analysis was used to determine recurrence-free rates. All P values were two-sided and a value of <0.05 was considered to indicate statistical significance. All analyses were performed using the SAS version 9.1.3 Service Pack 4.
The median (range) age of the population was 68 (26–90) years with a median (range) follow-up for patients alive of 29.1 (1–176) months, and a mean follow-up 35 months. Table 1 lists the clinical and pathological characteristics of the patient cohort who developed recurrence. Pelvic lymphadenectomy was performed in 91% of patients and 28% received adjuvant chemotherapy. Eight hundred and twenty five patients (43.6%) had recurrence with a median (range) time to recurrence of 10.1 (1–192) months. Approximately 90 and 97% of all recurrences happened by 2 and 5 years after RC, respectively, and 29% within the first 6 months (Fig. 1). The latest reported recurrence date was at 16 years. According to location, 48.6% of recurrences were distant with the remaining divided into: 25.2% pelvic, 14.5% retroperitoneal and 11.8% to multiple regions such as pelvic and retroperitoneal or pelvic and distant. Of those that were labelled as distant recurrences, the most common sites were lungs (42%), bone (36%), liver (27%) and brain (2% [Table 2]). Of patients who developed recurrence, 250 (30%) received salvage therapy (140 and 110 received systemic chemotherapy and radiation, respectively)
Table 1. Clinical and pathological characteristics in 825 patients who developed disease recurrence
Table 2. Time to recurrence stratified by stage and site of recurrence
Time to recurrence, months
When stratified according to stage, tumours with positive nodes (pTxN+) had worse outcomes than did extravesical node-negative tumours (≥pT3N0) or organ-confined node-negative tumours (≤pT2N0 [5-year recurrence-free survival of 25%, 44% and 66% respectively, P < 0.001]). Furthemore, pTxN+ tumours had a shorter median time to recurrence (9 months, range 1–72 months) than ≥pT3N0 tumours (10 months, range 1–70 months) or ≤pT2N0 tumours (14 months, range 1–192 months, P < 0.001 [Table 2]).
Subsequently, pulmonary recurrences were stratified according to pathological stage (Fig. 2). Following the same trend, ≤pT2N0 tumours had lower recurrence rates at 1 and 2 years post-RC (1.2% and 2.2%, respectively) compared with ≥pT3N0 (5.1% and 7.6%, respectively) and pTxN+ tumours (6.1% and 7.7%, respectively).
Alternatively, recurrences were then stratified according to two recurrence sites: abdominal and non-abdominal (Fig. 3A,B and C). Similarly, abdominal recurrence rates were higher in pTxN+ tumours at 1 and 2 years post-RC (19.2% and 25.4%, respectively) than in ≥pT3N0 tumours (13.2% and 18.8%, respectively) or in ≤pT2N0 tumours (4.1% and 7.0%, respectively). Finally, while distant non-abdominal recurrences were scarce in ≤pT2N0 tumours (3.1% and 4.6% at 1 and 2 years post-RC, respectively), they were significantly higher in ≥pT3N0 tumours (8.4% and 13.1% respectively) and in pTxN+ tumours (17.2% and 20.7%, respectively).
In the present study, we show that, in most patients, tumours recur within 2 years of RC, with almost half of these recurrences being distant. While the outcomes of these patients continue to be generally poor, tumours with higher pathological T stage and those with node-positive disease not only had the worst outcomes, but also the earliest median times to overall and distant recurrence. Based on recurrence patterns, a stage-based surveillance strategy was developed (Table 3).
Table 3. Recommended stage-specific surveillance protocol after RC
Urine washings/cytology once a year is optional. Vitamin B12 recommended when clinically indicated and upper tract imaging to assess the uretero-ileal anastomosis at 6–8 weeks from time of RC is recommended in all groups.
To date, there has been no consensus on how to follow up these patients appropriately after surgical intervention. The most recognizable early attempt at establishing surveillance guidelines after RC was suggested by Montie  who advocated chest X-ray and blood studies every 6 months for 2 years and then yearly for 3 more years, with CT of the abdomen and pelvis at 6, 12 and 24 months, irrespective of tumour stage. As is evidenced by the present study and others, there is variation in recurrence patterns between the different subgroups of patients and, because of this variation, using a universal surveillance protocol for all patients would result in unnecessary testing in some and lack of it in others. A stage-specific approach, such as that previously described after nephrectomy for RCC, therefore becomes a more feasible and realistic enterprise . Currently available guidelines include the National Comprehensive Cancer Network and the European Association of Urology recommendations, which acknowledge the need for risk-stratified surveillance but do not clearly delineate any protocols [10,11]. The ESMO (European Society for Medical Oncology) guidelines, by contrast, do not address a stage-specific surveillance regimen .
The mean (range) local recurrence rate after RC in contemporary series is 11.5 (3–18.6)%, similar to the rate in the present study of 12.3% [4,13–19]. Data also suggest that the odds of recurrence are significantly higher in patients with advanced pathological stage and presence of nodal spread [14,20–22]. Not only does this hold true in the present study, but it also extends to times of recurrence with more aggressive tumours having earlier local recurrences and worse overall outcomes. Furthermore, most local recurrences tend to happen within 2 years of surgery (91%), but with some events occurring as late as 16 years from the time of intervention. Similarly, distant recurrence rates also seem to be higher in those patients with node-positive or non-organ-confined tumours [14,20–22] with a mean (range) distal recurrence rate of 25 (14–37)% in contemporary series [4,13–19]. In the present cohort, almost half of the recurrences were to distant sites, mostly to lungs, liver and bone with >90% happening within 2 years of surgical intervention.
Recurrence site trends to the abdomen, which included the pelvis, retroperitoneum and liver, showed that in the first 18–24 months after surgery, events were more common in tumours with extravesical extension and positive nodal disease than in organ-confined node-negative tumours, thus highlighting the need for stricter earlier imaging in the former population. For this reason, for organ-confined node-negative tumours, we recommend triphasic CT to assess upper tracts, abdomen and pelvis at 12 months after surgery and that this be repeated at 2 years, 3 years and 5 years. This should be initiated earlier, at 6 months after surgery in non-organ-confined node-negative tumours, and repeated annually till the 5-year mark. Finally, in node-positive tumours, we recommend an earlier initial CT at 3 months and every 6 months thereafter for the first 2 years, then annually till the 5-year mark.
Recurrences to distant non-abdominal sites are very rare in patients with localized disease. In fact, in the present study population, there was ≈ 2% recurrence to lungs in the first 2 years after surgery. In patients with localized disease, therefore, we only recommend chest xray once a year. In those with extravesical node-negative disease and especially in those with node-positive disease, distal recurrence rates, particularly in the lungs, were substantially higher, especially within the first 2 years, which necessitates earlier stricter surveillance. We recommend chest xray every 6 months for the first 2 years then annually thereafter for extravesical node-negative tumours and at 3 months then every 6 months till 5 years for node-positive tumours. Bone investigation can be reserved for when clinically indicated.
With regard to urethral and upper tract recurrence, it remains controversial whether there is a need for routine surveillance in asymptomatic patients, one reason being that the mean recurrence rates to both these sites are low (4.6% and 3.4%, respectively) . The argument for performing routine urethral wash cytologies stems from the fact that the most important predictor of surveillance and successful management of urethral recurrences is the stage at diagnosis . However, data have shown no difference either in the rate of disease progression or in survival between patients monitored with urethral washings and those who were not [25,26]. We recommend very close monitoring, especially within the first 24 months, for symptoms related to a urethral event such as haematuria, bloody discharge, voiding abnormalities, discomfort or a palpable mass. There should be a low threshold for investigation by means of urethral washings, urine cytology, and/or urethroscopy in these cases. The low rate of upper tract recurrence (1–9%), as well as the fact that most patients become symptomatic, also puts into question the need for routine upper tract surveillance [23,27]. Meissner et al.  showed that the efficiency of intravenous pyelogram for detecting upper tract recurrence after RC was only 0.75%, and two other studies showed that all patients diagnosed with upper tract recurrence were symptomatic despite routine intravenous pyelogram surveillance, which makes the rationale for its use debatable [28–30]. While there are no studies to date showing that CT can detect recurrences in the upper tract earlier, because we recommend routine triphasic CT imaging of the abdomen and pelvis, examining the upper tracts is already included at negligible added cost and morbidity. Performing yearly urine cytologies is optional as only 45% of overall recurrences after RC will have a positive or suspicious cytology at the time of recurrence . There is also evidence suggesting that routine non-oncological surveillance is of importance to avoid further metabolic disturbances, debilitating surgical complications and renal deterioration [24,31]; therefore, we recommend, for all patients, regular office visits with blood evaluation consisting of a chemistry panel, complete blood count and liver function tests. Vitamin B12 monitoring should be reserved for those patients displaying symptoms or abnormality on the serum blood tests. Finally, we also advocate upper tract imaging or lasix renal scan to assess the uretero-ileal anastomosis at 6–8 weeks from time of RC in all groups. A repeat assessment of the uretero-ileal anastomosis at 6 months could also be considered.
More recently, some have challenged the role of routine oncological surveillance after RC altogether, claiming no significant improvement in survival over symptom-guided follow-up . While these data are thought-provoking, it should be mentioned that they are retrospective single-institutional data and should be viewed with caution. Furthermore, with ever-evolving imaging techniques, earlier and better detection of recurrences becomes possible. Extrapolating from validated oncological principles, it is fair to assume that the administration of salvage chemotherapy and/or radiation would be most efficacious when tumour burden is low and therefore detection of recurrences early may offer these patients better outcomes. Therefore, based on empirical data and, in the absence of any prospective randomized trials comparing symptom-guided with routine surveillance, it remains prudent and safer for patients to be routinely surveyed after surgery. Furthermore, the present study was recently challenged by Giannarini et al. , who showed improved survival in patients in whom recurrences were diagnosed by routine follow-up investigation compared with those with symptomatic recurrences. Our strategy is in line with that adopted by the MD Anderson group, who put together a stage-specific surveillance protocol taking into account the cost of postoperative imaging to the US healthcare system . The major difference lies in the abdominal imaging as they chose to use a combination of CT and upper tract imaging (loopogram, excretory urogram or ultrasonography) in patients with extravesical or node-positive tumours, and upper tract imaging without CT in organ-confined node-negative tumours. We opted for a simpler imaging strategy, with triphasic CT of the abdomen/pelvis, which can also survey the upper tracts. If a patient is found to have poor renal function and i.v. contrast is not possible, a loopogram can be added.
Patients with BC are often older patients, smokers with comorbidities who are also at risk for second malignancies. Indeed, recent retrospective data comparing patients with BC and matched-controls have shown that over a follow-up period of 16 years, up to 12.0% and 16.2% of male and female patients with BC develop non-urothelial second malignancies, most commonly lung cancer . While our algorithm does not expand to include surveillance for these secondary cancers, we do advocate that these patients be further followed long-term by their primary practitioners for these malignancies.
It is important to note that the main limitations of the present study are its retrospective nature and the lack of standardized surveillance protocols across different centres. Additionally, although we advocate strict oncological surveillance after RC, we do recognize the limitations of current imaging techniques, such as CT and MRI, in detecting asymptomatic recurrences and, therefore, future endeavours to further evaluate and incorporate urinary biomarkers and newer more accurate imaging and pathological tools are necessary for better risk-stratification and earlier detection of recurrent disease. Although, adjuvant chemotherapy was offered to only a small subset of patients (28%) in the present cohort, this may have potentially altered the recurrence patterns and influenced our surveillance strategies. Furthermore, while economic analysis studies have been carried out for many cancers, there are no published reports comparing various surveillance regimens after RC. In the absence of a thorough investigation prospectively assessing the survival outcomes and financial implications of a symptom-guided surveillance protocol vs a surveillance protocol, we find it more prudent to opt for the latter. Finally, recent data shows that there is lack of uniformity among urological oncologists in post-RC surveillance and lack of compliance to a predetermined follow-up schedule . The urology community should adopt a universally standard, easily applicable surveillance model to be able to validate it prospectively.
In conclusion, differences in recurrence patterns between the different subgroups after RC suggest the need for stage-based follow-up protocols. Tumours with higher pathological T stage and those with node-positive disease not only have worse outcomes, but also earlier median times to overall and distal recurrence. While organ-confined node-negative tumours recur less often at either abdominal or non-abdominal sites, more advanced tumours do so much more frequently, and there are frequently non-abdominal recurrences in node-positive disease. Earlier strict surveillance is therefore strongly advocated in patients with extravesical and node-positive disease.