Maintenance Bacillus Calmette-Guerin in High-Risk Nonmuscle-Invasive Bladder Cancer

Transcription

1 710 Maintenance Bacillus Calmette-Guerin in High-Risk Nonmuscle-Invasive Bladder Cancer How Much Is Enough? Marc Decobert, PhD Helène LaRue, PhD François Harel, MSc François Meyer, MD Yves Fradet, MD Louis Lacombe, MD Centre de Recherche en Cancerologie de l Universite Laval, CHUQ/L Hôtel-Dieu de Quebec, Quebec, Canada. See editorial on pages 674-6, this issue. The first 2 authors contributed equally to this work. The authors thank Drs. Claude Lemay, Alain Naud, and Mireille Gregoire (CHUQ, L Hôtel-Dieu de Quebec, Quebec, QC), Drs. Jean-Guy Vezina and Rejean Roy (Hôpital de L Enfant-Jesus, Quebec, QC), Dr. Isabelle Boisvert (Hôpital du Grand-Portage, Rivière-du Loup, QC), Dr. François Benard (Hôpital Saint-Luc, Montreal, QC), Dr. Michael Jewett (Toronto Hospital, Toronto, ON), Dr. Martin Gleave (Vancouver Hospital, Vancouver, BC), Drs. Gary Steinhoff, John Kinahan, Eric Freedman, and Peter Pommerville (Victoria General Hospital, Victoria, BC), Dr. Joseph Chin (London Health Science Center, London, ON) for their participation in this study by referring patients, and Pierre Simard for data management. Sample collection from high-risk bladder cancer patients for this study was supported by University/Industry grant #UI from the Canadian Institutes of Health Research and Mayne Pharma Canada (formerly Faulding Canada). Address for reprints: Louis Lacombe, MD, Centre de Recherche en Cancerologie, L Hôtel-Dieu de Quebec, CHUQ, 11 côte du Palais, Quebec, PQ, Canada G1R 2J6; Fax: (418) ; louis.lacombe@crhdq.ulaval.ca Received August 24, 2007; revision received January 28, 2008; accepted January 29, BACKGROUND. Intravesical bacillus Calmette-Guerin (BCG) immunotherapy is effective in preventing recurrence and progression in nonmuscle-invasive bladder cancer but the dosing schedule and duration of treatment remain empirical. The outcome of BCG therapy was prospectively evaluated on patients according to the number of maintenance cycles received. METHODS. Between 1997 and 2002, 111 patients with nonmuscle-invasive bladder cancer at high risk of recurrence and progression underwent transurethral resection followed by intravesical instillations of BCG. After an induction treatment of 6 weekly instillations, patients were scheduled to receive a 3-weekly maintenance treatment at the 3rd, 6th, 12th, 18th, 24th, 30th, and 36th month. At each visit a clinical assessment was obtained. RESULTS. Over a median clinical follow-up of 31 months, 43% of patients experienced recurrence and 8.1% progressed to muscle-invasive disease or metastasis. Only 1 patient received all scheduled instillations. Patients who received at least 3 maintenance BCG cycles had a significantly reduced risk of recurrence (hazard ratio [HR] , P , adjusted for gender, age, and stage) compared with patients receiving only induction BCG. Twelve months after the end of maintenance, the estimated Kaplan-Meier recurrence-free survival was 89% for patients who received at least 3 maintenance BCG cycles, 67% for those who received 2 maintenance BCG cycles, and 41% for those who received only the induction BCG or 1 maintenance cycle (P ). CONCLUSIONS. The results of this study suggest that a minimum of 3 cycles of maintenance BCG is required to significantly reduce the recurrence rate. It also suggests that more cycles may result in further improvements but the benefits may be outweighed by increasing side effects in some patients. Cancer 2008;113: Ó 2008 American Cancer Society. KEYWORDS: bladder cancer, bacillus Calmette-Guerin (BCG), immunotherapy, recurrence, maintenance BCG. Bladder cancer represents the sixth most common cancer in Canada 1 and is nonmuscle-invasive (stage pta, pt1, or ptis) in approximately 70% of cases. 2 Nonmuscle-invasive transitional cell carcinomas (nmtcc) recurs in up to 70% of cases within 60 months. 2 Carcinoma in situ (ptis) is at higher risk of progression to muscleinvasive disease or metastasis, as more than 50% of patients treated with transurethral resection (TUR) alone suffer progression. 3 Standard treatment for high-risk nmtcc consists of TUR, removing all visible lesions, followed by intravesical therapy. A second look resection is performed for pt1 lesions to remove any residual tumor (37% of cases) and for further staging. 4 Persistence of pt1 ª 2008 American Cancer Society DOI /cncr Published online 9 June 2008 in Wiley InterScience (

2 Maintenance BCG in Bladder CA/Decobert et al 711 disease is associated with a very high rate of progression to muscle-invasive disease within 5 years. 5 Once appropriate staging is complete, intravesical chemotherapy or immunotherapy will follow. Although intravesical chemotherapy improves recurrence-free survival compared with TUR alone, 6 intravesical instillation of bacillus Calmette-Guerin (BCG) is recognized as the best treatment for high-risk nmtcc, combined with TUR, to prevent or delay recurrences and progression and to cure Tis. 2,7-11 Although its efficacy is clearly acknowledged, BCG treatment remains largely empirical in terms of dosing and duration. Knowing when to stop BCG or pursue a more aggressive strategy is one of the hardest decisions facing urologists managing these highrisk patients. Persistence of T1 disease 3 to 6 months post-bcg therapy has shown an association with an 82% chance of progression, with a median time of 8.4 months, compared with 25%, with a median time to progression greater than 5 years, for patients with non-t1 lesions or no disease, suggesting that these patients require additional therapy. 12 A maintenance regimen has been proposed, adding instillations cycles after the initial 6-instillation cycle. 10 A prospective study from the Southwest Oncology Group (SWOG) including 550 patients randomized to a weekly BCG regimen for 6 weeks, followed or not by 3 weekly instillations at the third and sixth month and every 6 months thereafter for up to 3 years, added support to the use of maintenance therapy. In patients with Tis, maintenance BCG resulted in an improvement in 5-year disease-free survival, from 41% in the observation group to 60% in the maintenance group (P <.04). 13 However, in that study only a small percentage of patients were able to complete the 3-year maintenance dosing schedule. This low level of compliance makes it difficult to perform a randomized study of the minimal number of BCG maintenance cycles necessary for an optimal effect. In recent years we prospectively recruited a cohort of patients treated with maintenance BCG for high-risk nmtcc to analyze various prognostic markers. The present study took advantage of this cohort to evaluate the modalities of maintenance BCG. MATERIALS AND METHODS From 1997 to 2002, 156 patients with high-risk nmtcc from 8 institutions in Canada, all eligible for BCG therapy, were offered to participate in this prospective study; 131 accepted and signed an informed consent form. Characteristics of high-risk tumors were stage pt1, grade 3, multiplicity (>3), diffuse Tis, more than 3 cm in size, or the presence of recurrent tumors within 6 months of inclusion TUR. Previous BCG therapy more than 1 year before inclusion and previous use of chemotherapeutic agents like Mitomycin C were allowed provided they were discontinued 6 months before inclusion. Previous chemotherapy for invasive bladder cancer with no residual invasive disease was allowed provided it was discontinued 6 months before inclusion. Patients previously treated with local radiation therapy to the pelvis for any cancer or planning to receive concomitant chemotherapy were excluded. Fourteen were excluded because pathologic staging showed muscle-invasive tumors, 5 for insufficient follow-up, and 1 for protocol violation. At least 3 weeks after TUR, patients received the initial 6 weekly intravesical administrations of 120 mg BCG (Pacis Shire Pharmaceuticals) in 50 ml of physiological bacteriostatic-free saline solution. The solution was instilled in an emptied bladder through a catheter. Patients were instructed to try retaining the solution for 2 hours, changing position every 15 minutes. Maintenance therapy was given weekly for 3 weeks, starting the week after control cystoscopies at the 3rd, 6th, 12th, 18th, 24th, 30th, and 36th month. In cases of moderate to severe side effects, dosage reduction to 60 or 30 mg was allowed. Control cystoscopies were pursued annually after the first 3 years. At each visit clinical assessment including a questionnaire on side effects and urine cytology was obtained. Visible recurrences or suspicious lesions were removed by TUR. The TNM classification was used for staging. 14 Complete response was defined as normal cystoscopy and negative cytology. For Tis a biopsy was necessary. Tumor recurrence or progression was histologically confirmed or based on the surgeon s opinion in cases of electrofulguration or laser treatment. To avoid any doubt as to the authenticity of recurrences, as recurrences detected at the first cystoscopy might be tumors missed at TUR, only recurrences detected at the second cystoscopy (sixth month) and after were considered. The primary endpoint was recurrence-free survival (RFS). To demonstrate a relationship between the number of BCG cycles and RFS, the latter was not evaluated from the time of TUR because the length of BCG treatment itself could affect the validity of statistical tests. Premature interruption of scheduled BCG cycles due to early recurrence might have introduced a potential bias in the analysis, namely, the inability to determine whether the low number of cycles resulted in recurrence or whether recurrence resulted in the low number of cycles. Therefore, RFS was defined as the period elapsed between the last BCG instillation and

3 712 CANCER August 15, 2008 / Volume 113 / Number 4 TABLE 1 Baseline Characteristics Overall and by Number of Maintenance Bacillus Calmette-Guerin (BCG) Cycles Maintenance BCG Cycles Overall (N5111) None (n540) 1 (n524) 2 (n516) 3 or More (n531) Variable Levels n/n % n/n % n/n % n/n % n/n % P Sex Men 91/111 (82.0) 32/40 (80.0) 18/24 (75.0) 14/16 (87.5) 27/31 (87.1).62 Women 20/111 (18.0) 8/40 (20.0) 6/24 (25.0) 2/16 (12.5) 4/31 (12.9) Age, y <70 58/111 (52.3) 19/40 (47.5) 9/24 (37.5) 9/16 (56.3) 21/31 (67.7) /111 (47.7) 21/40 (52.5) 15/24 (62.5) 7/16 (43.8) 10/31 (32.3) Mean SD Smoker Nonsmoker 20/104 (19.2) 7/37 (18.9) 3/22 (13.6) 4/16 (25.0) 6/29 (20.7).97 Current smoker 20/104 (19.2) 7/37 (18.9) 5/22 (22.7) 2/16 (12.5) 6/29 (20.7) Exsmoker 64/104 (61.5) 23/37 (62.2) 14/22 (63.6) 10/16 (62.5) 17/29 (58.6) No. of tumors 1 38/96 (39.6) 13/37 (35.1) 9/21 (42.9) 4/11 (36.4) 12/27 (44.4).87 >1 58/96 (60.4) 24/37 (64.9) 12/21 (57.1) 7/11 (63.6) 15/27 (55.6) Initial stage TA 54/111 (48.6) 19/40 (47.5) 10/24 (41.7) 6/16 (37.5) 19/31 (61.3).58 T1 36/111 (32.4) 13/40 (32.5) 9/24 (37.5) 5/16 (31.3) 9/31 (29.0) TIS 21/111 (18.9) 8/40 (20.0) 5/24 (20.8) 5/16 (31.3) 3/31 (9.7) Grade 1 11/111 (9.9) 4/40 (10.0) 0/24 (0.0) 3/16 (18.8) 4/31 (12.9) /111 (47.7) 17/40 (42.5) 12/24 (50.0) 6/16 (37.5) 18/31 (58.1) 3 47/111 (42.3) 19/40 (47.5) 12/24 (50.0) 7/16 (43.8) 9/31 (29.0) either recurrence or the last study visit. 15 The secondary endpoint was progression-free survival (PFS), defined similarly. Patients were followed past a first recurrence to detect any progression. The association between the number of BCG cycles and the endpoints was evaluated using survival analysis techniques. 16 Kaplan-Meier survival curves were constructed and compared using the log-rank test. The hazard ratio (HR) and 95% confidence intervals (CI) according to the number of BCG cycles were computed. To facilitate clinical interpretation, survival curves were also constructed using a standard definition of RFS after TUR. Multivariate Cox proportional hazards models were used to adjust for potential confounding effect of gender, age, stage, and grade. Comparisons of baseline characteristics were done using 1-way analysis of variance and Pearson s chisquare test. Statistical analysis was performed using SAS v. 9.1 (SAS Institute, Cary, NC). Statistical significance was determined at P <.05. All tests were 2- sided. RESULTS Many patients from this cohort at high risk of recurrence had tumors with properties usually associated with a bad prognosis: 42% were grade 3, 32% were pt1, and 60% were multiple (Table 1). For analysis purpose, and because of the limited number of patients, the cohort was divided into 4 groups, based on the number of BCG cycles: no maintenance, 1, 2, and 3 or more maintenance cycles received before the first recurrence. Disease progression, patient refusal, or severe side effects motivated ending the treatment. Only 1 patient tolerated the complete treatment schedule of 8 cycles (27 instillations). No significant differences were found between groups on any of the baseline characteristics except for age, where patients receiving 3 or more maintenance BCG were younger (P 5.041). Over a median clinical follow-up of 31 months, 48 of the 111 patients (43.2%) experienced recurrence. Thirty-seven (77.1%) of these first recurrences were histologically confirmed. Five patients had invasive tumors, whereas 19 patients had pta, 8 had pt1, and 5 had ptis tumors; 6 patients had grade 1, 17 had grade 2, and 14 had grade 3 tumors. On the basis of univariate Cox proportional hazards model, gender and number of maintenance BCG cycles were predictors of RFS (Table 2). Women were at higher risk of recurrence (HR , P 5.028). Compared with no maintenance, 1 maintenance BCG cycle had no significant effect on recurrence risk, whereas 2 maintenance cycles produced a 50% decrease in recurrence risk, which did not reach statistical significance (HR , P 5.14). Three or more maintenance BCG cycles significantly reduced the recurrence risk by 80% (HR , P ). Controlling for age, gender, and stage, multivariate analysis confirmed those observations for at least 3 maintenance BCG cycles (HR , P ). Further subdivision of patients showed a risk associated with

4 Maintenance BCG in Bladder CA/Decobert et al 713 TABLE 2 Cox Proportional Hazards Regression Analysis to Predict Time to Recurrence After Last Bacillus Calmette-Guerin (BCG) Instillation Univariate Multivariate* Predictor Levels Frequency (%) Event (%) HR 95% CI P HR 95% CI P Sex Men 91 (82.0) 35 (38.5) Women 20 (18.0) 13 (65.0) 2.05 [1.08; 3.87] [0.85; 3.27].17 Age, y <70 58 (52.3) 21 (36.2) (47.7) 27 (50.9) 1.51 [0.85; 2.67] [0.73; 2.41].35 Smoking status None 20 (19.2) 11 (55.0) 1.0 Current 20 (19.2) 9 (45.0) 0.72 [ 0.3; 1.74].46 Ex-smoker 64 (61.5) 25 (39.1) 0.66 [0.32; 1.34].25 No of tumors 1 38 (39.6) 18 (47.4) 1.0 >1 58 (60.4) 24 (41.4) 0.99 [0.54; 1.83].97 Initial stage TA 54 (48.6) 21 (38.9) T1 36 (32.4) 17 (47.2) 1.32 [0.69; 2.49] [0.68; 2.49].42 TIS 21 (18.9) 10 (47.6) 1.15 [0.54; 2.45] [0.50; 2.26].88 Grade 1 11 ( 9.9) 2 (18.2) (47.7) 23 (43.4) 3.33 [0.78; 14.1] (42.3) 23 (48.9) 2.94 [0.69; 12.5].14 Risk level T1-G3 20 (18.0) 11 (55.0) 1.27 [0.65; 2.5].48 Other 91 (82.0) 37 (40.7) 1.0 Maintenance BCG cycles None 40 (36.0) 24 (60.0) (21.6) 15 (62.5) 1.67 [0.88; 3.21] [0.91; 3.35] (14.4) 5 (31.3) 0.49 [0.19; 1.28] [0.19; 1.30].15 3 and more 31 (27.9) 4 (12.9) 0.21 [0.07; 0.59] [0.08; 0.66].0064 HR indicates hazard ratio; CI, confidence interval. * Adjusted for gender, age, and stage. 3 maintenance cycles of (adjusted P 5.15) and with four or more maintenance cycles of (adjusted P 5.017). However, there was no significant difference between these 2 groups (P 5.98) and for that reason we maintained our grouping of 3 or more maintenance cycles. To calculate Kaplan-Meier RFS estimates, none and 1 maintenance BCG were grouped together. The Kaplan-Meier curves (Fig. 1) showed a significant difference between groups in RFS after the last BCG instillation (overall log-rank test, P ); however, the reduction of risk of recurrence attributed to 2 maintenance BCG failed to reach statistical significance (P 5.054). Twelve months after the last BCG instillation the estimated frequency of recurrence was of 11.5% (95% CI, 0% 24%) for patients who received at least 3 maintenance BCG cycles, 33.3% (95% CI, 9% 57%) after 2 maintenance cycles, and 53.5% (95% CI, 40% 66%) for patients who received 1 maintenance cycle or only the 6 initial instillations. We also observed that among the 31 patients who received at least 3 maintenance BCG cycles, the 4 patients who suffered recurrence received 3 (n 5 1) or 4 (n 5 3) maintenance BCG cycles, and that no recurrence occurred among the 12 patients who received 5 maintenance BCG cycles or more. FIGURE 1. Kaplan-Meier survival analysis of recurrence-free survival according to the number of cycles of maintenance bacillus Calmette-Guerin (BCG), from time of last BCG treatment. To facilitate clinical interpretation, Figure 2 shows the same data analyzed using RFS after TUR. Thirty-six months after TUR the estimated frequency of recurrence was of 11% (95% CI, 0% 23%) for patients who received at least 3 cycles of maintenance BCG, 33.3% (95% CI, 9% 57%) after 2 maintenance cycles, and 71.5% (95% CI, 58% 84%) for

5 714 CANCER August 15, 2008 / Volume 113 / Number 4 FIGURE 2. Kaplan-Meier survival analysis of recurrence-free survival according to the number of cycles of maintenance bacillus Calmette-Guerin (BCG), from time of transurethral resection (TUR). patients who received 1 maintenance cycle or only the 6 initial instillations. Recurrence rates were similar for 20 T1G3 and non-t1g3 tumors (P 5.48) (Table 2). Progression was observed in 9 patients (8.1%), 8 developing muscle-invasive disease, and 1 a metastasis. Five tumor progressions appeared at the first recurrence; 3 of them occurring less than 7 months after TUR. The 4 other progressions occurred at 18 and 33 months (second recurrence), 44 months (fifth recurrence), and 49 months (fourth recurrence) after TUR. Progression was slightly more frequent among patients who received none or 1 maintenance BCG than among patients who received 2 maintenance BCG cycles or more (7/ % vs 2/ %, P 5.20). During the study period, 3 of the 9 patients with progressive bladder cancer died from their cancer; 4 other patients died from other cancers and 2 from heart disease. Noncompliance to the whole schedule of treatment was mostly related to morbidity (Table 3). Local side effects were more frequently reported and their frequency significantly related to the number of BCG cycles received. Bladder irritation and burning sensation were reported by about 20% of patients receiving no maintenance and by more than 50% of patients after 3 maintenance BCG cycles (Table 3). Systemic side effects were less frequent and their frequency not significantly related to the number of BCG cycles. DISCUSSION Intravesical immunotherapy with BCG has been used for almost 30 years in the treatment of nmtcc after TUR or to treat Tis. It is effective in delaying or preventing recurrence and progression, although its outcome is still unpredictable. Various trials have suggested that maintenance therapy could improve the outcome of BCG treatment but the optimal maintenance scheme still has to be defined. 9 Andius and Holmang 17 showed that multiple instillation cycles improved RFS compared with a single 6-week induction in 236 Ta/T1 bladder cancer patients, but suggested that maintenance therapy may not be necessary for pta and lower-grade tumors. Van der Meijden et al 18 concluded that maintenance therapy should mainly be applied to high- and intermediaterisk-tumors. Saint et al 19 similarly reported that maintenance therapy for nmtcc yielded a significant recurrence rate reduction but adherence was hindered by adverse reactions, leading 81% of patients to voluntary interruption. Han and Pan, 11 in a metaanalysis of 25 trials (4767 patients), concluded that TABLE 3 Side Effects Overall and by Number of Maintenance Bacillus Calmette-Guerin (BCG) Cycles Maintenance BCG Cycles Overall (N5111) None (n540) 1 (n524) 2 (n516) 3 or More (n531) Adverse Side Effects No. % No. % No. % No. % No. % P None 46 (41.4) 24 (60.0) 10 (41.7) 5 (31.3) 7 (22.6).012 Local Bladder irritation 39 (35.1) 7 (17.5) 7 (29.2) 8 (50.0) 17 (54.8).0054 Burning 41 (36.9) 8 (20.0) 10 (41.7) 7 (43.8) 16 (51.6).040 Hematuria 37 (33.3) 9 (22.5) 6 (25.0) 6 (37.5) 16 (51.6).054 Systemic Dyspnea 5 (4.5) 2 (5.0) 1 (4.2) 1 (6.3) 1 (3.2).97 Chills 17 (15.3) 2 (5.0) 5 (20.8) 3 (18.8) 7 (22.6).15 Nausea 3 (2.7) 0 (0.0) 0 (0.0) 1 (6.3) 2 (6.5).24 Fever (>38.58C) 9 (8.1) 2 (5.0) 3 (12.5) 2 (12.5) 2 (6.5).64

6 Maintenance BCG in Bladder CA/Decobert et al 715 intravesical BCG with maintenance treatment should be offered as the treatment of choice to patients with papillary carcinomas. In the present study we used a cohort of prospectively recruited patients with high-risk nmtccs who received various numbers of maintenance BCG cycles to evaluate the impact of the number of BCG cycles administered. We are well aware that this is not a randomized study and that the number of patients is limited. However, we believe that carrying out a randomized study would be very difficult, for various reasons, the most important being the low level of compliance of patients to maintenance BCG. Considering these limitations we nonetheless believe that the analysis we carried out provides reliable information to help optimize maintenance BCG therapy. To increase the reliability of our analysis we used a strict definition of recurrence, considering only those discovered at the second cystoscopy and later. Moreover, as detailed in Materials and Methods, RFS time was evaluated from the end of BCG treatment to minimize the bias introduced by the arrest of BCG therapy due to early recurrence. Under these conditions the present study showed that BCG maintenance therapy brought a delayed first recurrence and a trend in lowering progression rate in high-risk nmtccs. We showed that 2 maintenance BCG cycles was sufficient to have an impact on RFS, whereas 3 or more led to a significant reduction of the risk of recurrence (Fig. 1, Table 2), with an RFS of 82.6% 3 years after the last BCG instillation compared with only 29.0% for patients receiving only induction BCG. Although the approximately 5-fold risk reduction brought by 3 maintenance cycles did not reach significance, we observed no significant difference between 3 cycles and 4 or more cycles, suggesting that 3 demonstrated a near maximal effect. Moreover, patients who received 5 treatments or more had no recurrence. Although these observations do not point to an optimal number of maintenance cycles to recommend, we believe they suggest that patients should be encouraged to tolerate at least 3 cycles of maintenance and to continue further instillations if well tolerated. An earlier meta-analysis, regrouping results from 24 studies, demonstrated that maintenance BCG was associated with a reduced risk of progression. 9 Tumor progression was very low in our cohort (8.1%) and could not be significantly associated with maintenance treatment, although we observed 10.9% progression among patients receiving only the induction or 1 maintenance BCG compared with 4.3% for patients with 2 or more maintenance BCG. Our 90.8% progression-free status at 3 years and low progression occurrence for T1G3 tumors (14%) may reflect a different nature of tumors treated in general urology practice compared with the higher progression rate observed in referral cancer centers. Local side effects such as cystitis, hematuria, and pollakiuria, and systemic side effects mainly related to host reactivity against infection affect BCG-treated patients. It was reported that side effects tend to be less frequent during maintenance. 18 However, we observed an increase in local side effects with an increased number of BCG treatments. Because of early recurrence or intolerable side effects, only 1 patient received all scheduled maintenance BCG cycles. Conclusion This study shows that maintenance BCG is effective in increasing the RFS of patients with high-risk nmtcc. However, the low percentage of patients tolerating the entire planned schedule of treatment shows that despite its efficacy BCG should still be administered with care. Our data also suggest that maintenance BCG shows some efficacy even with a reduced number of cycles, as only 2 maintenance BCG cycles notably reduced the risk of recurrence and 3 or more lowered the risk by approximately 80% with no significant difference between 3 cycles and 4 or more cycles. Nonetheless, the absence of recurrences in the small group of patients who received 5 cycles and more suggest that patients should be encouraged to continue the treatment beyond 3 maintenance cycles, if well tolerated. Although this study was not randomized, its observations could be helpful to counsel bladder cancer patients treated with BCG. REFERENCES 1. McLaughlin JR, Dryer D, Mao Y, et al. Canadian Cancer Statistics Toronto, Canada: Canadian Cancer Society/ National Cancer Institute of Canada, Amling CL. Diagnosis and management of superficial bladder cancer. Curr Probl Cancer 2001;25: Lamm DL. Carcinoma in situ. Urol Clin North Am. 1992; 19: Rigaud J, Karam G, Braud G, Glemain P, Buzelin JM, Bouchot O. [T1 bladder tumors: value of a second endoscopic resection.] Prog Urol. 2002;12: Herr HW, Donat SM. A re-staging transurethral resection predicts early progression of superficial bladder cancer. BJU Int. 2006;97: Tolley DA, Parmar MK, Grigor KM, et al. The effect of intravesical mitomycin C on recurrence of newly diagnosed superficial bladder cancer: a further report with 7 years of follow up. J Urol. 1996;155:

7 716 CANCER August 15, 2008 / Volume 113 / Number 4 7. Altay B, Girgin C, Kefi A, Cikili N. The best management of superficial bladder tumours: comparing TUR alone versus TUR combined with intravesical chemotherapy modalities? [In Process Citation.] Int Urol Nephrol. 2000;32: Lamm DL, Blumenstein BA, Crawford ED, et al. A randomized trial of intravesical doxorubicin and immunotherapy with bacille Calmette-Guerin for transitional-cell carcinoma of the bladder. N Engl J Med. 1991;325: Sylvester RJ, Van der Meijden AP, Lamm DL. Intravesical bacillus Calmette-Guerin reduces the risk of progression in patients with superficial bladder cancer: a meta-analysis of the published results of randomized clinical trials. J Urol. 2002;168: Lamm DL. BCG in perspective: advances in the treatment of superficial bladder cancer. Eur Urol. 1995;27(suppl 1): Han RF, Pan JG. Can intravesical bacillus Calmette-Guerin reduce recurrence in patients with superficial bladder cancer? A meta-analysis of randomized trials. Urology. 2006;67: Herr HW. Progression of stage T1 bladder tumors after intravesical bacillus Calmette-Guerin. J Urol. 1991;145: Lamm DL, Blumenstein BA, Crissman JD, et al. Maintenance bacillus Calmette-Guerin immunotherapy for recurrent TA, T1 and carcinoma in situ transitional cell carcinoma of the bladder: a randomized Southwest Oncology Group Study. J Urol. 2000;163: Sobin LH, Fleming ID. TNM Classification of Malignant Tumors, fifth edition (1997). Union Internationale Contre le Cancer and the American Joint Committee on Cancer. Cancer. 1997;80: Redmond C, Fisher B, Wieand HS. The methodologic dilemma in retrospectively correlating the amount of chemotherapy received in adjuvant therapy protocols with disease-free survival. Cancer Treat Rep. 1983;67: Collett D. Modelling Survival Data in Medical Research, 2nd ed. London: Chapman and Hall; Andius P, Holmang S. Bacillus Calmette-Guerin therapy in stage Ta/T1 bladder cancer: prognostic factors for time to recurrence and progression. BJU Int. 2004;93: Van der Meijden AP, Sylvester RJ, Oosterlinck W, Hoeltl W, Bono AV. Maintenance bacillus Calmette-Guerin for Ta T1 bladder tumors is not associated with increased toxicity: results from a European Organisation for Research and Treatment of Cancer Genito-Urinary Group Phase III Trial. Eur Urol. 2003;44: Saint F, Irani J, Patard JJ, et al. Tolerability of bacille Calmette-Guerin maintenance therapy for superficial bladder cancer. Urology. 2001;57: