Intravesical gemcitabine in combination with mitomycin C as salvage treatment in recurrent non-muscle-invasive bladder cancer

Transcription

1 Intravesical gemcitabine in combination with mitomycin C as salvage treatment in recurrent non-muscle-invasive bladder cancer Patrick A. Cockerill, John J. Knoedler, Igor Frank, Robert Tarrell and Robert J. Karnes Department of Urology, Mayo Clinic, Rochester, MN, USA Objectives To evaluate oncological outcomes after combination intravesical therapy with gemcitabine (GC) and mitomycin C (MMC) in the setting of recurrent non-muscle-invasive bladder cancer (NMIBC) after failure of previous intravesical therapy. Patients and Methods We retrospectively identified patients with recurrent NMIBC after previous intravesical therapy, who refused or were not candidates for cystectomy, between 2005 and GC and MMC were sequentially instilled weekly for 6 8 weeks. Data were collected regarding patient demographics, bladder cancer history, and number and type of intravesical therapies before GC/MMC. Outcomes evaluated included time to recurrence and/or progression after GC/MMC. Recurrence-free outcomes were estimated using the Kaplan Meier method, and Cox proportional hazards regression models were used to test the association of clinicopathological features with outcomes. Results In all, 27 patients were identified, 23 with high-risk disease (high-grade or carcinoma in situ) and four with intermediaterisk disease (multifocal or recurrent low-grade). All patients received prior intravesical therapy, and 17 patients (63%) received multiple courses. Twenty-four patients were treated with BCG. The median (range) disease-free survival of all patients was 15.2 ( ) months. Seventeen patients (63%) developed recurrent bladder cancer, a median of 15.2 months after therapy. One patient progressed to muscle-invasive disease 5 months after treatment, and one developed metastatic disease 22 months after treatment. Three patients went on to cystectomy. Ten patients (37%) had no evidence of disease at last follow-up, with a median follow-up of 22.1 months. Conclusion The combination of intravesical GC and MMC could offer durable recurrence-free survival to some patients with recurrent NMIBC who are not candidates for, or refuse, cystectomy. Keywords gemcitabine, mitomycin C, non-muscle-invasive bladder cancer Introduction In 2012 there were an estimated new cases of bladder cancer and deaths [1]. Although radical cystectomy remains the gold standard for muscle-invasive tumours, 70% of new cases are non-muscle-invasive tumours [2]. Of these tumours, 70% are Ta, 25% are T1, and 5% are carcinoma in situ (CIS) [2]. The European Organization for Research and Treatment of Cancer (EORTC) assesses recurrence and progression risk with a combination of multiplicity, size, recurrence, T stage, presence of CIS and grade, dividing patients into low-, intermediate- and high-risk categories, with recurrence rates of 30%, 46 63% and 78%, respectively [3]. Adjuvant intravesical therapy is employed to decrease the risk of recurrence and progression after transurethral resection of non-muscle-invasive bladder cancer (NMIBC). A single postoperative instillation of chemotherapy is associated with an 40% decrease in the odds of recurrence compared with transurethral resection alone [4]. For intermediate- and highrisk tumours, an additional 6 weeks of courses of chemotherapy or immunotherapy are warranted to further decrease the risks of recurrence and progression. Primary options for treatment include mitomycin C (MMC) and BCG. Additional available agents based on the AUA guidelines include interferon, thiotepa, and the intercalating agents doxorubicin, valrubicin and epirubicin [2]. The only US Food and Drug Administration-approved regimen for BCGrefractory CIS is valrubicin, which achieves a complete response rate of 18% (recurrence-free at 3 months follow-up) [5]. Emerging evidence suggests that gemcitabine (GC) alone might also be efficacious in the setting of BCG failure [6 8]. Despite intravesical therapy, some patients develop tumour recurrence. The AUA guidelines recommend consideration of BJU Int 2016; 117: wileyonlinelibrary.com BJU International 2015 BJU International doi: /bju Published by John Wiley & Sons Ltd.

2 Intravesical gemcitabine and mitomycin C in recurrent NMIBC cystectomy as therapy in patients with recurrent tumours after one induction course of intravesical therapy [2]. Further intravesical therapy is an option, and this becomes relevant in clinical scenarios where the patient either refuses cystectomy or is not medically fit for the operation. The guidelines acknowledge that there is no standard option for intravesical therapy in this scenario and that further investigation is needed to define appropriate therapies. Intravesical GC alone has been evaluated in several trials and multiple observational studies, in the postoperative setting, as primary intravesical therapy and as salvage therapy after BCG [6]. In the setting of BCG failure, Sternberg et al. [7] evaluated 69 patients, 27 of whom experienced a complete response, with 46 eventually experiencing recurrence. On the other hand, in the setting of BCG failure, the role of MMC alone has not been fully elucidated. In a randomized trial comparing MMC with BCG in high-risk bladder cancer, 21 patients crossed over to MMC after failing BCG, but only four remained recurrence-free after treatment [8]. The combination of GC and MMC has been evaluated in two previous cohorts, the largest of which treated 47 patients from three institutions, 14 (30%) of whom remained recurrence-free at a median follow-up of 26 months [9,10]. The concept of combination therapy parallels the treatment of systemic urothelial carcinoma, and many other cancers, where multi-drug regimens are the standard of care. A population of tumour cells is heterogenous and might have different genetic abnormalities. Using multiple drugs with multiple different mechanisms provides the chance for maximal kill of tumour cells and decreases the chance of developing cells resistant to therapy [11]. Thus, through their separate mechanisms of action, GC and MMC might treat more tumour cells than either drug alone and therefore reduce the chance of recurrence. GC is a pyrimidine analogue that incorporates into actively replicating DNA and thereby prevents further synthesis, whereas MMC cross-links DNA moieties to prevent synthesis. In addition, MMC is a vesicant to the urothelium, which could increase permeability to subsequent GC administration through its irritating action. Whether their mechanisms are cumulative or synergistic on the same cancer cell, or whether MMC alkylates DNA in cells that did not take up and therefore would not respond to GC remains to be explored. After seeing promising preliminary data from another institution using combination therapy, we adopted this approach at our institution after verifying the feasibility through our pharmacy. While the combination of GC/MMC has been previously evaluated, this has been through small institutional series and has not been reproduced [9,10]. In addition, the previous studies included treatment-na ıve patients who had not received previous intravesical therapy and therefore did not fully evaluate the role of GC/MMC as salvage treatment for previous intravesical failures. The objective of the present study was to evaluate outcomes with GC/MMC in patients with recurrent NMIBC who have failed prior intravesical therapy. Patients and Methods We retrospectively evaluated 27 patients who received GC and MMC intravesically between 2005 and 2011 at our institution. Since 2005, we have offered GC/MMC in our clinical practice as an option to patients with recurrent NMIBC, who have failed previous intravesical therapies, and who either refused cystectomy or were not medically acceptable candidates for cystectomy as deemed by the primary treating urologist. Specifically, all patients had to have failed previous intravesical therapy. For the present study we then retrospectively identified and analysed these patients after institutional review board approval. Before undergoing GC/MMC, all patients were clear of visible disease. If disease was found on office white light cystoscopy, transurethral resection under anaesthesia was performed before proceeding. Gemcitabine was instilled into the bladder for 90 min, at a dose of mg in 50 ml of sterile water. The GC was then drained via a catheter. Immediately afterwards, MMC was instilled into the bladder for 90 min, at a dose of 40 mg in 20 ml of sterile water. This was repeated weekly for 6 8 weeks. No standardized maintenance intravesical therapy was utilized. Therapy was delayed for active UTI or myelosuppression at the treating physician s discretion. The patients were then followed with abdominal imaging, urine cytologies and cystoscopic examination using white light. Although individual surveillance was at the treating physician s discretion, follow-up at our institution has been recommended every 3 months for the first 2 years after surgery, every 6 months for the next 2 years, and annually thereafter in patients without evidence of disease recurrence. In the event of disease recurrence, the surveillance schedule was reset to every 3 months for 2 years. Evaluation of suspected recurrence with cystoscopy and transurethral resection under anaesthesia was carried out at the physician s discretion. Data were gathered about baseline patient characteristics, bladder cancer history, previous intravesical therapies, as well as the date and number of doses of GC/MMC therapy. After treatment with GC/MMC, patients were considered to have recurrence of intravesical neoplasm if there was pathological recurrence (i.e. biopsy-proven recurrence), cystoscopic evidence that was considered definitive by the treating physician or distant recurrence documented on imaging (distant recurrence was not necessarily biopsy-proven if radiographically consistent). Positive cytology alone without cystoscopic or pathological data and upper tract tumours was BJU International 2015 BJU International 457

3 Cockerill et al. coded as recurrence. Suspicious findings on abdominal imaging underwent image-guided biopsy, and, if positive, were coded as recurrence. The time to recurrence was calculated as the interval from initiation of GC/MMC therapy (i.e. the first instillation) to the date of recurrence. A summary of patient characteristics is provided in Table 1. In addition, recurrence-free outcomes were estimated using the Kaplan Meier method. Patients were censored at last follow-up or response if the endpoint of interest had not been attained. Cox proportional hazards regression models were used to test the association of clinicopathological features with outcome. All tests were two-sided, with P 0.05 considered to indicate statistical significance. Statistical analysis was done using SASR, version 9.2. (SAS Institute, Cary, NC, USA). Results Twenty-seven patients were identified, including 24 men and three women. The mean (interquartile range [IQR]) age at diagnosis was 68 ( ) years and the mean (IQR) age at treatment was 72 (65 80) years. The patients had a mean (range) of 3.6 (1 10) recurrences of intravesical neoplasm before treatment with GC/MMC. Twenty-five of 27 (93%) patients had a history of grade 2 or 3 tumours, and 17/27 (63%) patients had a history of CIS. Four of 27 patients (15%) were at intermediate risk for recurrence and progression (multifocal or recurrent low-grade) and 23 (85%) were at high risk for recurrence (high-grade with or without CIS) (Table 1). All 27 patients received intravesical therapies before receiving GC/MMC including BCG, MMC or BCG in combination with interferon-alpha. A single postoperative instillation of intravesical therapy was not considered to be a course of therapy. Twenty-four patients (89%) underwent BCG therapy before GC/MMC, and 10 (37%) additionally underwent BCG in combination with interferon. Ten patients (37%) received only one course of therapy, nine patients (33%) received two courses of therapy, and eight patients (30%) received three or more courses of therapy. The patients received GC/MMC therapy a median (IQR) of 40 (18 67) months after their original diagnosis of bladder cancer. Patients received induction therapy with six to eight doses of GC/MMC. Five patients received an induction course of six doses, four received seven doses and 14 received eight doses. Patients received fewer than eight doses secondary to side-effects from therapy. Maintenance was not standardized among the cohort. In our cohort, only three patients received maintenance courses of therapy after induction GC/MMC, and this was at the discretion of the treating physician. These three patients received maintenance GC alone for three to five instillations. In addition, there was one patient who received a second induction course of GC/MMC 8 months after initial treatment secondary to a suspicious cytology. This patient subsequently had negative cytologies and negative cystoscopies at the last follow-up 36 months later. There were eight patients who reported side-effects/adverse events. The most common was irritative voiding and bladder spasms, which occurred in six of them. Anaemia occurred in two patients, thought to be secondary to systemic absorption of GC. One patient developed acute renal failure during therapy. Four patients received incomplete courses of therapy, one for acute renal failure and three secondary to irritative voiding symptoms (Table 2). Of the 27 patients who underwent salvage GC/MMC therapy, 17 experienced intravesical recurrence (63%), with a median (range) time to recurrence of 15.2 (1.7 32) months. A Kaplan Meier estimate is shown in Fig. 1. Of the 17 patients who experienced disease recurrence, 13 recurred with NMIBC, one with muscle-invasive bladder cancer and one with upper tract T3 disease, with two patients lacking Table 1 Patient characteristics. Characteristic No. of patients Total 27 Male 24 Female 3 Median age at diagnosis, years 69 Median age at treatment, years 74 N % Risk category before GC/MMC Low 0 0 Intermediate 4 15 High Median no. of previous treatments Type of previous treatment BCG BCG plus interferon Highest T stage Ta 7 26 T CIS alone 7 26 CIS + Ta 5 19 CIS + T Highest grade Table 2 Reports of adverse effects. Adverse effect N (%) Irritative voiding/bladder spasms 6 (22) Anemia 2 (7) Acute renal failure 1 (4) 458 BJU International 2015 BJU International

4 Intravesical gemcitabine and mitomycin C in recurrent NMIBC Fig. 1 Recurrence free survival. Recurrence-free, % Survival,% (no. at risk) RFS 100(27) Time after GC, months 72(16) 49(10) 22(2) pathological data. Of the 13 patients who recurred with NMIBC, three had intermediate-risk disease before treatment and recurred with intermediate-risk disease, and two had high-risk disease before treatment and recurred with intermediate risk disease. The remaining eight had high-risk disease before treatment and recurred with high-risk disease (Table 3). We performed univariate Cox proportional hazards regression analysis to evaluate the impact of pre-therapy variables on recurrence. Pre-therapy variables analysed included gender, number of previous recurrences ( 3 vs >3), presence vs absence of CIS before therapy, high grade before therapy (grade 3 vs grade 1 or 2), number of courses of intravesical therapy before GC/MMC ( 1 vs >1), age at treatment with GC/MMC, and time from initial diagnosis to combination therapy. There were no significant predictors of response in our cohort (Table 4). Ten patients (37%) had no evidence of recurrence at a median (range) follow-up of 22 ( ) months (Table 5). Over the course of the study, eight patients died, including one patient who died of metastatic urothelial cancer 14 months after concluding GC/MMC therapy. Discussion At our institution, the salvage protocol of intravesical GC in combination with MMC is used in patients refusing cystectomy or who are medically unfit for cystectomy. The present cohort of 27 patients represents recurrent disease after failed intravesical therapy, at intermediate to high risk of recurrence. These characteristics (high-grade and BCG failure) have proved to be predictors of progression [12], and failure to respond to an induction course of BCG has been associated with increased risk of progression and death [13]. Despite these poor prognostic indicators, we noted that 37% Table 3 Summary of patient characteristics. Patient Highest T stage Highest grade No. of recurrences Previous BCG No. of previous intravesical courses Time to last follow-up, months Time to recurrence, months Stage/grade at recurrence 1 Ta 3 1 Yes Ta 1 2 Yes T1 2 3 Yes CIS 3 2 Yes CIS 3 2 Yes CIS 3 3 Yes Ta + CIS 3 1 Yes Ta + CIS 3 2 Yes Ta + CIS 3 3 Yes T1 + CIS 3 9 Yes Ta 3 3 Yes 1 4 T1G3 12 Ta 1 2 Yes 2 12 TaG1 13 Ta 3 5 Yes 2 10 TaG2 14 Ta 2 9 Yes 1 26 TaG1 15 Ta 2 7 Yes 3 3 TaG1 16 T1 3 5 Yes 3 5 T2G3 17 T1 2 3 Yes 1 22 TaG1 18 CIS 3 2 Yes 2 27 CIS 19 CIS 3 2 No 1 26 T1G3 20 CIS 3 1 Yes 1 2 Unknown 21 CIS 3 4 Yes 7 32 CIS 22 Ta + CIS 3 4 No 1 15 CIS 23 Ta + CIS 3 4 Yes 2 18 Upper Tract T3 24 T1 + CIS 3 1 No 1 3 CIS 25 T1 + CIS 3 3 Yes 3 13 CIS 26 T1 + CIS 3 3 Yes 2 15 T1G3 + CIS 27 T1 + CIS 3 10 Yes 2 24 Unknown BJU International 2015 BJU International 459

5 Cockerill et al. Table 4 Univariate analysis: predictors of response to GC/MMC. HR 95% CI P value Gender Age Time from diagnosis to GC/MMC >3 recurrences before GC/MMC Absence of CIS before GC/MMC Presence of grade 3 NMIBC before GC/MMC No. of previous intravesical courses CI, confidence interval. Table 5 Outcomes after GC/MMC. Outcomes Overall median recurrence-free survival, months 15.2 Number of recurrences 17 (63.0%) Median time to recurrence, months 15.2 Number recurrence-free at last follow-up 10 (37.0%) Median follow-up in recurrence-free, months 22 of patients had no recurrence at a median follow-up of 22 months. In the 63% of patients who recurred, the median time to recurrence was 15.2 months. Only one patient experienced disease progression while receiving GC/MMC. We were interested to see if any pre-treatment characteristics could predict response to GC/MMC, and thus we performed a univariate analysis to evaluate predictors of response to GC/ MMC. No characteristic was predictive of response, including gender, age, time from diagnosis to treatment with GC/MMC, number of recurrences ( 3 vs >3) before GC/MMC, presence vs absence of CIS before GC/MMC, high grade (grade 3 vs grade 1 or 2) before GC/MMC, and number of previous intravesical courses ( 1 vs >1) before GC/MMC. The inability of the univariate analysis to find predictors of outcomes is secondary to our small patient cohort, which is a problem inherent to this patient population. Intravesical therapy with BCG is the mainstay of adjuvant treatment for NMIBC. The AUA guidelines recommend an induction course of intravesical chemotherapy or BCG after surgical treatment of NMIBC, and support a maintenance regimen of BCG to augment its efficacy [2]. Patients may undergo additional courses of BCG, although efficacy declines with repeated courses and it has been suggested that alternative treatment should be sought after two BCG failures [14]. Additional options include further intravesical chemotherapy, immunotherapy, thermochemotherapy and photodynamic therapy, although many of these are still investigational [15]. There has been increasing interest in new intravesical therapies for these patients, to avoid the morbidity and quality of life changes associated with cystectomy. To our knowledge, no study has specifically evaluated MMC alone in the setting of previous BCG or intravesical therapy failure. MMC has shown efficacy in treating NMIBC, although it proved inferior to BCG in the setting of disease at high risk of recurrence [16]. Strategies to augment the effect of MMC have been evaluated, including chemohypothermia, which has been reported to reduce recurrence rates by up to 60% compared with MMC alone, and electromotive delivery [17,18]. However, specifically in the BCG failure setting, in a randomized trial comparing MMC with BCG in high-risk bladder cancer, 21 patients crossed over to MMC after failing BCG and only four remained recurrence-free, suggesting limited utility of MMC alone in this patient population [8]. Gemcitabine has proven activity in advanced urothelial carcinoma, prompting interest for use in NMIBC. In a phase I study, Dalbagni et al. [19] showed safety of up to mg of GC instilled intravesically. Side-effects included irritative voiding symptoms, haematuria, UTI, hand foot syndrome, myelosuppression, weakness, nausea and vomiting. The most common of these was irritative voiding, with the others being rare in their small cohort. The subsequent phase II trial included 30 patients with CIS, T1 or high-grade Ta, all of whom previously received intravesical therapy [20]. The authors administered mg of GC twice a week for 3 weeks, and noted a 50% complete response to therapy (no disease at 3 months). The 1-year recurrence-free survival in patients with a complete response was 21%, and two patients had continued recurrence-free survival after 23 and 29 months. Since initial studies affirmed the safety of GC intravesically, multiple studies have been carried out using it as a single dose after surgery, as induction intravesical chemotherapy or as maintenance intravesical chemotherapy. The studies, although heterogeneous, have supported intravesical GC s safety and activity in NMIBC [6]. There have been few observational studies in patients with NMIBC refractory to BCG. Gunelli et al. [21] noted 95% complete response at 6 months in 40 patients with BCG-refractory NMIBC, with continued complete response in 80% of patients at 1 year and 66% of patients at 2.5 years [21]. Importantly, Gunelli et al. s study included only patients with Ta or T1 disease, and excluded patients with CIS. Additional studies of GC in BCGrefractory NMIBC reported 33 60% recurrence-free rates [22 24]. However, none of these studies specifically evaluated patients with CIS or high-risk disease, instead focusing on BCG-refractory disease, which included many intermediaterisk tumours. In the study by Perdona et al. [23], one-third of the patients had CIS, whereas the other studies excluded patients with CIS. There have been three manuscripts, evaluating two patient cohorts, using GC in combination with MMC for refractory NMIBC. Breyer et al. [25] administered GC/MMC to BJU International 2015 BJU International

6 Intravesical gemcitabine and mitomycin C in recurrent NMIBC patients and noted recurrence-free survival in six of them (60%) at a median follow-up of 14 months. Maymi [9] reported a 50% recurrence-free survival in 27 patients treated with GC/MMC at 18 months [9]. Lightfoot et al. [10] reported on a multi-institution cohort of 47 patients receiving GC/MCC. They noted a response in 14 of 47 patients (30%), who remained recurrence-free at a median follow-up of 26 months. However, it is important to note that their study included 17 patients who were either BCG-na ıve or had never received any intravesical therapy, by contrast with our study, where all patients had failed previous therapy. However, the authors did perform a subgroup analysis and noted no significant difference in recurrence-free survival between groups. We believe the present study offers promising results, which are similar to those from Lightfoot s study utilizing GC/MMC, as we found a recurrence-free rate of 37% at 22 months, compared with 30% at 22 months. However, with a cohort consisting of patients who have failed courses of intravesical therapy, we feel the present data further reinforce the utility of GC/MMC in this group. We believe that when these studies are considered in aggregate, GC/MMC shows promise in treating patients with recurrent NMIBC after failed intravesical therapy, and could be useful as an additional treatment regimen in these patients. Limitations of the present study include its retrospective nature. The patient cohort is heterogeneous, in that patients with different types of previous bladder pathology were included in the analysis, and it is small, which limits the conclusions we can make about the general patient population. Median follow-up was 22 months, which is relatively short for NMIBC. Patients who underwent the salvage protocol of GC/MMC either were not candidates for cystectomy or were interested in alternatives for bladder preservation. Because of this, the data include patients with different recurrence risks and they are analysed as one cohort. Finally, there is no referent cohort for comparison. However, we believe our salvage protocol shows promise, and our data corroborates this, in patients with recurrent NMIBC who have failed other types of intravesical therapy. In conclusion, patients with recurrent NMIBC who have failed previous intravesical therapy continue to be a challenge to manage. Treatment with radical cystectomy has been challenged in recent years with the advent of new intravesical therapies. The combination of GC and MMC has shown promise in observational studies, and the present study corroborates these findings and shows good results in a subset of patients. Caution must be used in interpretation, however, secondary to the observational and retrospective nature of the study. Randomized controlled trials are needed to further delineate the exact role GC and MMC will play in therapy. However, initial results suggest that despite having recurrent high-grade NMIBC, some patients might experience durable recurrence-free survival. Conflict of Interest None disclosed. References 1 American Cancer Society (ACS). Cancer Facts & Figures Atlanta, Georgia: American Cancer Society (ACS), American Urologic Association. Guideline for the management of nonmuscle invasive bladder cancer: (Stages Ta, T1, and Tis). J Urol 2007; 178: Babjuk M, Burger M, Zigeuner R et al. EAU guidelines on non-muscleinvasive urothelial carcinoma of the bladder, the 2011 update. Eur Urol 2011; 59: Sylvester RJ, Oosterlinck W, van der Meijden AP. A single immediate postoperative instillation of chemotherapy decreases the risk of recurrence in patients with stage Ta T1 bladder cancer: a meta-analysis of published results of randomized clinical trials. J Urol 2004; 171: Dinney CP, Greenberg RE, Steinberg GD. Intravesical valrubicin in patients with bladder carcinoma in situ and contraindication to or failure after bacillus Calmette-Geurin. Urol Oncol 2013; 31: Shelly MD, Jones G, Cleves A, Wilt TJ, Mason MD, Knaystone HG. Intravesical gemcitabine therapy for non-muscle invasive bladder cancer (NMIBC): a systematic review. BJU Int 2012; 109: Sternberg IA, Dalbagni G, Chen LY, Donat SM, Bochner BH, Herr HW. Intravesical gemcitabine for high risk, nonmuscle invasive bladder cancer after bacillus Calmette-Geurin treatment failure. J Urol 2013; 190: Malmstrom P, Wijkstrom H, Lundholm C, Wester K, Busch C, Norlen BJ. 5-year followup of a randomized prospective study comparing Mitomycin C and bacillus Calmette-Guerin in patients with superficial bladder carcinoma. J Urol 1999; 161: Maymi JL. Intravesical sequential gemcitabine-mitomycin chemotherapy as salvage treatment for patients with refractory superficial bladder cancer. J Urol 2006; 175(Suppl.): Lightfoot AJ, Breyer BN, Rosevear HM, Erickson BA, Konety BR, O Donnell MA. Multi-institutional analysis of sequential intravesical gemcitabine and mitomycin C chemotherapy for non-muscle invasive bladder cancer. Urol Oncol 2014; 32: Chu E, DeVita V. Principles of medical oncology. In Devita V, Hellman S, Rosenberg S eds, Cancer: Principles & Practice of Oncology, 7th edn, Vol. I. Chapt 14. Philadelphia: Lippincott Williams & Wilkins, 2005: Shirakawa H, Kikuchi E, Tanaka N et al. Prognostic significance of Bacillus Calmette-Guerin failure classification in non-muscle-invasive bladder cancer. BJU Int 2012; 110: Lerner SP, Tangen CM, Sucharew H, Wood D, Crawford ED. Failure to achieve a complete response to induction BCG therapy is associated with increased risk of disease worsening and death in patients with high risk non-muscle invasive bladder cancer. Urol Oncol 2009; 27: Catalona WJ, Hudson MA, Gillen DP, Andriole GL, Ratliff TL. Risks and benefits of repeated courses of intravesical bacillus Calmette-Guerin therapy for superficial bladder cancer. J Urol 1987; 137: Yates DR, Roupret M. Failure of bacille calmette-guerin in patients with high risk non-muscle-invasive bladder cancer unsuitable for radical cystectomy: and update of available treatment options. BJU Int 2010; 106: Shelley MD, Court JB, Knayston H, Wilt TJ, Coles B, Mason M. Intravesical bacillus Calmette-Guerin versus mitomycin C for Ta and T1 bladder cancer. Cochrane Database Syst Rev 2003; (3): CD Colombo R, Salonia A, Leib Z, Pavone-Macaluso M, Engelstein D. Longterm outcomes of a randomized controlled trial comparing thermochemotherapy with mitomycin-c alone as adjuvant treatment for non-muscle-invasive bladder cancer (NMIBC). BJU Int 2011; 107: BJU International 2015 BJU International 461

7 Cockerill et al. 18 Di Stasi SM, Giannantoni A, Stephen RL et al. Intravesical electromotive mitomycin C versus passive transport mitomycin C for high risk superficial bladder cancer: a prospective randomized study. J Urol 2003; 170: Dalbagni G, Russo P, Sheinfeld J et al. Phase I trial of intravesical gemcitabine in bacillus Calmette-Guerin-refractory transitional-cell carcinoma of the bladder. J Clin Oncol 2002; 20: Dalbagni G, Russo P, Bochner B et al. Phase II trial of intravesical gemcitabine in bacille Calmette-Guerin-refractory transitional cell carcinoma of the bladder. J Clin Oncol 2006; 24: Gunelli R, Bercovich E, Nanni O et al. Activity of endovesical gemcitabine in BCG-refractory bladder cancer patients: a translational study. Br J Cancer 2007; 97: Mohanty NK, Nayak RL, Vasudeva P, Arora RP. Intravesical gemcitabine in management of BCG refractory superficial TCC of urinary bladder- our experience. Urol Oncol 2008; 26: Perdona S, Di Lorenzo G, Cantiello F et al. Is gemcitabine an option in BCG refractory non muscle invasive bladder cancer? A single arm prospective trial. Anticancer Drugs 2010; 21: Gacci M, Bartoletti R, Nerozzi S et al. Intravesical gemcitabine in BCGrefractory T1G3 transitional cell carcinoma of the bladder: a pilot study. Urol Int 2006; 76: Breyer BN, Whitson JM, Carroll PR, Konety BR. Sequential intravesical gemcitabine and mitomycin C chemotherapy regimen in patients with non-muscle invasive bladder cancer. Urol Oncol 2010; 28: Correspondence: Robert J. Karnes, Department of Urology, Mayo Clinic, 200 First Street Southwest, Rochester, MN 55905, USA. karnes.r@mayo.edu Abbreviations: CIS, carcinoma in situ; GC, gemcitabine; IQR, interquartile range; MMC, mitomycin C; NMIBC, nonmuscle-invasive bladder cancer. 462 BJU International 2015 BJU International