Shahrokh F. Shariat, Department of Urology and Division of Medical Oncology, Weill Medical College of Cornell University, New York Presbyterian Hospital, 525 East 68th Street, Box 94, Starr 900, New York, NY 10065, USA. e-mail: email@example.com
Study Type – Therapy (case series)
Level of Evidence 4
What's known on the subject? and What does the study add?
Outcomes after disease recurrence in patients with urothelial carcinoma of the bladder treated with radical cystectomy are variable, but the majority of patients die from the disease within 2 years after disease recurrence. Knowledge about prognostic factors that may influence survival after disease recurrence is limited.
We found that outcomes after disease recurrence in patients with urothelial carcinoma of the bladder are significantly affected by common clinicopathological factors. In addition, a shorter time from surgery to disease recurrence is significantly associated with poor outcomes. These factors should be considered when scheduling salvage chemotherapy protocols/clinical trials.
• To describe the natural history following disease recurrence after radical cystectomy (RC) and to identify prognostic factors that influence cancer-specific survival with special focus on time from RC to disease recurrence.
• We identified 1545 patients from 16 international institutions who experienced disease recurrence after RC and bilateral lymphadenectomy. None of the patients received preoperative chemotherapy; 549 patients received adjuvant chemotherapy.
• A multivariable Cox regression model addressed time to cancer-specific mortality after disease recurrence.
• The median cancer-specific survival time after disease recurrence was 6.9 months (95% CI 6.3–7.4). Overall, 1254 of 1545 patients died from urothelial carcinoma of the bladder and 47 patients died from other causes. The actuarial cancer-specific survival estimate at 12 months after disease recurrence was 32%.
• On multivariable analysis, non-organ-confined tumour stages (hazard ratio [HR] 1.38, P= 0.002), lymph node metastasis (HR 1.25, P < 0.001), positive soft tissue surgical margin (HR 1.32, P= 0.002), female gender (HR 1.21, P= 0.003), advanced age (HR 1.16, P < 0.001) and a shorter interval from surgery to disease recurrence (P < 0.001) were significantly associated with cancer-specific mortality.
• The adjusted risk of death from cancer within 1 year after disease recurrence for patients who recurred 6, 12 and 24 months after surgery was 70%, 64% and 60%, respectively.
• Over two-thirds of patients who experience disease recurrence of urothelial carcinoma of the bladder after RC die within 12 months.
• Common clinicopathological factors are strongly associated with cancer-specific mortality. A shorter time from surgery to disease recurrence is significantly associated with poor outcomes.
• Accurate risk stratification could help in patient counselling and decision-making regarding salvage treatment.
Urothelial carcinoma of the bladder (UCB) is a common malignancy with an estimated 55 600 new cases and 14 880 deaths in the USA in 2012 . Radical cystectomy (RC) with bilateral lymphadenectomy (LND) and urinary diversion with or without preoperative chemotherapy is the gold standard treatment for patients with muscle-invasive and high risk non-muscle-invasive UCB [2–4]. Despite advances in surgical techniques and improvement of systemic chemotherapies, up to 50% of patients with muscle-invasive UCB experience disease recurrence within 5 years after surgery and the majority of these patients eventually succumb to UCB [2,3,5]. The natural history of UCB from RC and LND to disease recurrence has been intensively investigated and several studies have identified risk factors to help in outcome prognostication and individualized risk assessment [6–9]. In contrast, only a few studies have evaluated the natural history of UCB after disease recurrence, and knowledge about prognostic factors that may influence survival after disease recurrence is very limited [10–12]. Clinical course after disease recurrence is variable, with 80% of patients dying from UCB within 1–2 years after disease recurrence and 20% surviving beyond 2 years post-recurrence [2,3,5,11–13]. Improved understanding of the natural history of recurrent UCB could help in patient counselling and clinical trial design.
Using a large international cohort, we attempted to describe the natural history of patients who experienced disease recurrence after RC and LND and we identify risk factors for cancer-specific mortality. We hypothesize that standard pathological features help predict outcomes in patients with UCB even after disease recurrence. Moreover, we hypothesize that a shorter time to disease recurrence after RC would be associated with an unfavourable clinical course.
All studies were performed with the approval and oversight of the Institutional Review Board at each institution, with all participating sites providing the necessary data-sharing agreements prior to initiation.
In total, 1818 patients from 16 international centres experienced disease recurrence from UCB (median time of 10 months) after RC and bilateral LND between 1979 and 2010. The database was closed for follow-up in 2011; prior to final analysis the database was frozen. Patients who were missing follow-up data (n= 242) or did not undergo LND (n= 31) were excluded, leaving a final cohort of 1545 patients for analysis. All patients had pathological documentation of UCB and recurrence of UCB, with no evidence of distant metastases at the time of RC. All patients underwent surgery according to criteria consistent with guideline recommendations . All patients underwent an RC with bilateral pelvic LND and urinary diversion as described elsewhere . No patient received preoperative chemotherapy or radiotherapy. Adjuvant chemotherapy was administered to 549 patients (36%) at the clinicians' discretion based on tumour stage and overall health status.
All surgical specimens were processed according to standard pathological procedures as previously described . Genitourinary pathologists assigned tumour grade according to the 1973 WHO grading system. Pathological stage was reassigned according to the 2002 American Joint Committee on Cancer TNM staging system. The presence of concomitant carcinoma in situ (CIS) was defined as the presence of CIS in conjunction with another tumour other than CIS. Pelvic lymph nodes (LNs) were examined grossly, and all lymphoid tissue was submitted for histological examination. The extent of LN dissection was at the surgeon's discretion. Positive soft tissue surgical margin (STSM) was defined as the presence of tumour at inked areas of soft tissue on the RC specimen . Urethral or ureteral margins were not considered as STSMs. Lymphovascular invasion (LVI) was defined as the unequivocal presence of tumour cells within an endothelium-lined space without underlying muscular walls [15,16].
Follow-up was performed according to institutional protocols. Patients were generally seen postoperatively at least every 3–4 months for the first year, semi-annually for the second year and annually thereafter. Follow-up visits consisted of a physical examination and serum chemistry evaluation, including liver function tests and alkaline phosphatase. Diagnostic imaging of the upper tracts (e.g. ultrasonography and/or intravenous pyelography, CT abdomen/pelvis with intravenous contrast) and chest radiography were performed at least annually or when clinically indicated. Additional radiographic evaluations, such as bone scan and/or CT, were performed at the discretion of the treating physician when indicated. Detection of cancer in the ureter, renal pelvicalyceal system and/or urethra was coded as a second (metachronous) primary and not as local or distant recurrence. When patients died, the cause of death was determined by the treating physicians, by chart review corroborated by death certificates, or by death certificates alone . Peri-operative mortality (i.e. death within 30 days of surgery) was censored at time of death for bladder-cancer-specific survival analyses.
The aim of this analysis was to describe the relationship between risk of cancer-specific mortality after disease recurrence and time from RC to disease recurrence. To assess this relationship, we constructed a Cox proportional hazards regression model. The time to death from UCB was calculated as months from disease recurrence to death. The Cox model included time from RC to disease recurrence, age, gender, pathological stage categorized as non-muscle-invasive (pT0, pTa, pTis, pT1), muscle-invasive (pT2) and non-organ-confined disease (pT3–4), pathological grade categorized as G0–G1, G2 or G3, LN status, concomitant CIS, LVI and STSM. Months from RC to disease recurrence was added to the model with restricted cubic splines with knots at the tertiles to account for nonlinearity. As a sensitivity analysis the model was created using varying knot locations but this had little effect on the model results. Competing risk regression was not utilized because so few patients died from other causes. Adjuvant chemotherapy was not included as a covariate in the Cox regression model because many patients recurred prior to 3 months after RC and potentially did not have the opportunity to receive adjuvant chemotherapy. Additionally, adjuvant chemotherapy is not expected to improve survival after disease recurrence . As a sensitivity analysis, we repeated our previous analysis including adjuvant therapy and omitting patients with less than 3 months of follow-up. Kaplan–Meier curves were constructed to illustrate cancer-specific survival after disease recurrence. All reported P values are two-sided and a P value of 0.05 was considered to be statistically significant. All statistical analyses were performed using Stata 11.0 (Stata Corp., College Station, TX, USA).
ASSOCIATION OF DISEASE RECURRENCE WITH CLINICOPATHOLOGICAL CHARACTERISTICS AND OUTCOMES
The clinicopathological characteristics of the 1545 patients with disease recurrence are summarized in Table 1. The median age was 67 years (interquartile range 60–74). Median time from RC to disease recurrence was 10 months (interquartile range 5–20). The majority of patients with disease recurrence had advanced and biologically aggressive tumour features, such as pT3–4 (64%) and G2–3 disease (97%), LVI (51%) and LN metastasis (43%).
Table 1. Descriptive characteristics of 1545 patients with urothelial carcinoma of the bladder treated with radical cystectomy and bilateral lymphadenectomy who experienced disease recurrence
IQR, interquartile range.
Age (years, median, IQR)
Pathological tumour stage
Pathological tumour grade
No grading (pT0 patients)
Lymph node metastasis
Concomitant carcinoma in situ
Positive soft tissue surgical margin
Months from surgery to recurrence (IQR)
In total, 1254 of 1545 patients with disease recurrence died from UCB, while 47 patients died of other causes. The median time from disease recurrence to death was 10 months. The probability of cancer-specific survival after disease recurrence is displayed in Fig. 1. Actuarial cancer-specific survival estimates at 6, 12 and 24 months after disease recurrence were 54% (95% CI 51–56), 32% (95% CI 29–34) and 17% (95% CI 15–19), respectively.
ASSOCIATION OF TIME TO DISEASE RECURRENCE WITH SURVIVAL AFTER DISEASE RECURRENCE
Figure 2 illustrates the risk of cancer-specific mortality within 1 year of disease recurrence by months from RC and LND to disease recurrence, adjusted to an average patient with UCB in our cohort: a 67-year-old male patient with a non-organ-confined stage G3 tumour without concomitant CIS and negative STSM, LVI positive, and without LN metastasis. The risk of death from UCB was found to be highest if the time to disease recurrence was less than 12 months after RC. The probability of cancer-specific mortality at 1 year after disease recurrence for the average patient was 75%, 70% and 64% if disease recurred 3, 6 or 12 months respectively after RC. After 12 months after RC, the risk of cancer-specific mortality continues to fall as time from RC to disease recurrence increases; however, the risk only slightly declines, compared with the distinct decline in risk between 0 and 12 months.
Figure 3A,B show the estimated 6-month risk of cancer-specific mortality after disease recurrence stratified by pathological stage or LN status and time from RC to disease recurrence adjusted for all factors in the model. The risk of death from UCB was influenced adversely by advancing stage and decreased time to disease recurrence (Fig. 3A): patients with non-organ-confined disease had an increased risk of cancer-specific mortality at any time compared with patients with organ-confined UCB, but the increase in time from RC to disease recurrence led to a linear decrease in the cancer-specific mortality risk in every group. The risk of death from UCB was also influenced by LN metastasis. Patients with LN metastasis had poorer outcomes than patients without LN metastasis at any time after disease recurrence (Fig. 3B).
RISK FACTORS FOR OUTCOMES AFTER DISEASE RECURRENCE
On multivariable Cox regression analysis that adjusted for the effects of standard clinicopathological characteristics, higher tumour stages (hazard ratio [HR] pT3–4 vs pT0, Ta, Tis, T1 1.38, P= 0.002), LN metastasis (HR 1.25, P < 0.001), positive STSM (HR 1.32, P= 0.002), female gender (HR 1.21, P= 0.003), higher age (HR 1.16, P < 0.001) and a shorter time from surgery to disease recurrence (P < 0.001) were significantly associated with cancer-specific mortality (Table 2).
Table 2. Multivariable Cox proportional hazards regression model predicting cancer-specific mortality in 1545 patients treated with radical cystectomy and bilateral lymphadenectomy for urothelial carcinoma of the bladder
Hazard ratio and confidence interval are not given due to nonlinear modelling. See Fig. 2 for the relationship between time to disease recurrence and risk of death from UCB after disease recurrence.
INFLUENCE OF ADJUVANT CHEMOTHERAPY ON OUTCOMES AFTER DISEASE RECURRENCE
In the sensitivity analysis including adjuvant chemotherapy and excluding those patients with less than 3 months of follow-up (151 died of the disease, seven died of other causes and 21 were censored), adjuvant therapy was not significantly associated with death from the disease after disease recurrence (HR 1.02, 95% CI 0.896–1.18, P= 0.7). Including adjuvant chemotherapy did not change the hazard ratios of the other covariates in the model, suggesting that adjuvant chemotherapy is not a confounding factor in this analysis.
Patients with UCB who experience disease recurrence after RC and LND have very poor outcomes. Despite advances in surgical techniques and systemic chemotherapies [2,3,5], the majority of patients (68%) die from UCB within 1 year after disease recurrence and only few patients (17%) survive beyond 2 years after disease recurrence. Modern salvage chemotherapies may prolong survival if a patient responds, but cure is very rare [13,18]. This underscores the lethal nature of UCB once the disease recurs and becomes systemic [2,3,5]. Interestingly, over 10% of patients were still alive 3 years after disease recurrence demonstrating the highly variable natural history of UCB. Accurate prediction of clinical outcomes after disease recurrence could help in patient counselling and clinical trial design and analysis. However, to the best of our knowledge, only a few smaller studies have analysed risk factors for outcomes in patients with UCB who experience disease recurrence after RC [10–12].
We confirmed that time from RC to disease recurrence is a strong predictor of survival after disease recurrence . The shorter the time from surgery to disease recurrence, the shorter is the survival time. Interestingly, survival after disease recurrence disproportionately improved in patients experiencing disease recurrence after 12 months. After 12 months, the time to death remains associated with the time to disease recurrence, but the rate of decrease in survival becomes proportionally smaller. Moreover, the relationship between time to recurrence and rate of mortality is enhanced in patients with advanced stage or LN positive UCB. In other words, while even after 12 months it remains true that the shorter the time to disease recurrence the faster the mortality, the change in rate becomes less significant if the time to disease recurrence is over 12 months. Similar to previous studies, we confirmed that, in patients who experience disease recurrence, the median time from surgery to disease recurrence was less than 12 months [11,12]. Mitra et al.  reported that a time to disease recurrence <12 months is associated with worse outcomes in patients with UCB. Finally, previous studies did not investigate cut-offs other than 12 months or time to disease recurrence as a continuous variable.
We found that established pathological features such as non-organ-confined pathological stage, LN metastasis and positive STSM as well as the clinical factors such as advanced age and female gender were significant predictors of cancer-specific mortality after disease recurrence. All these factors have previously been shown to be strong predictors of survival outcomes in patients treated with RC and LND [2,3,5–7,9,14,19–21]. It seems reasonable that the more biologically aggressive the disease, the faster the disease may recur. Pathological and nodal stage are still the most well established predictors for outcome prognostication [3,6,7,9,22,23]. In this study we show that these factors cannot only help in outcome prediction after RC, but should also be taken into consideration for patient counselling and risk stratification after disease recurrence. Bajorin et al.  have demonstrated that patients' performance status and location of metastases impact survival after urothelial carcinoma recurrence. Unfortunately, these data were not available and therefore our analyses remained unadjusted for these factors which might have influenced outcomes. However, since outcomes are variable and outcome prediction based on clinicopathological factors alone is of limited reliability, our study underscores the urgent need for new biomarkers allowing a better prognostication of clinical outcomes and prediction of response to therapy [24,25]. This could help in patient counselling and selection of therapies, as well as in timely administration of optimal systemic therapy in patients with UCB experiencing disease recurrence . Although no UCB blood or tissue biomarker is yet widely accepted, they still hold the potential to improve UCB staging and outcome prognostication [24–26].
Our study is not devoid of limitations. First are limitations inherent to the multicentre and retrospective design over a long time period including a lack of data regarding prior treatments, delay between diagnosis and surgery, patient preferences and comorbidities. We also could not adjust for the number of surgeons at each institution as well as surgeons' preferences, experience or techniques. However, all surgeons operated at tertiary care centres with experience in UCB. Another limitation includes possible inter-observer variability between pathologists; a central pathology review was not performed. Differences in the indication for and protocols of adjuvant chemotherapy were not controlled for in our study. Moreover, patients with early disease recurrence and/or those with a worse performance status after RC might not have the opportunity to receive adjuvant chemotherapy. In addition, recurring patients after adjuvant chemotherapy may have platinum-refractory disease, which is associated with poor survival and might represent another bias . Finally, data on salvage therapies and patients performance score after disease recurrence were missing and therefore remained unadjusted for in this study. However, previous studies have shown that salvage chemotherapy might prolong survival by a few months but almost never cures patients [13,18]. There is no doubt that our findings need to be validated in robust prospective studies.
In conclusion, over two-thirds of patients die from UCB within 12 months of disease recurrence after RC. Time from surgery to disease recurrence is a strong predictor of cancer-specific mortality, particularly if disease recurrence occurs within the first 12 months after RC. In addition, non-organ-confined pathological stage, LN metastasis, positive STSM, advanced age and female gender are strongly associated with cancer-specific mortality after recurrence. These factors should be considered for patient counselling as well as for the design and analysis of salvage chemotherapy protocols/clinical trials.
We thank the members of the Bladder Cancer Research Consortium (BCRC): Evi Comploj, Siamak Daneshmand, Yves Fradet, Jonathan I. Izawa, Wassim Kassouf, Seth P. Lerner, Roman Mayr, Francesco Montorsi, Arthur I. Sagalowsky, Mark P. Schoenberg, Eila Skinner, Viktor Soukup, Robert S. Svatek and Bjoern G. Volkmer. Dr. Michael Rink and Dr. David A. Green are supported by the Frederick J. and Theresa Dow Wallace Fund of the New York Community Trust.