Repeat Induction

On failing one course of BCG, a second BCG induction course can result in a significant response rate. Brake and colleagues examined 126 patients with T1 disease who underwent induction BCG, 37 (29%) of whom recurred at a median of 16 months. The 24 subjects with non-muscle invasive recurrence underwent a second 6-week induction course of BCG, with a 79% response rate (19 of 24) at a median of 24 months, displaying the benefit of repeat induction at time of recurrence. However, 13% of patients (17 of 126) developed progression to MIBC, either after one course (N = 13) or two courses (N = 4) of BCG. There was no placebo group in this study, so it is impossible to know the true effect of second BCG induction as compared with TUR alone. In addition, patients with residual tumor after repeat induction went on to cystectomy; thus, it is unknown if they would have achieved a delayed response, as is seen with primary BCG induction.

Maintenance

Another form of additional BCG is maintenance therapy, as reported in the pivotal randomized trial conducted by the Southwest Oncology Group (SWOG) in 2000. Patients with BCG-naive NMIBC were randomized to receive intravesical and percutaneous BCG for a 6-week induction course, or induction plus 3 weeks of maintenance at 3, 6, 12, 18, 24, 30, and 36 months. Five-year recurrence-free survival was significantly higher in the maintenance group (60% vs 41%), although there was no significant difference in overall survival. A subset analysis of patients with CIS also showed an improved response rate with maintenance. However, only 16% of patients randomized to maintenance received all 3 years of therapy because of a combination of recurrence and adverse events.

A more recent follow-up review of the SWOG BCG maintenance trial evaluated potential risk factors for recurrence and progression. Early and late (>12 months) recurrence were associated with increased risk of death, compared with no recurrence, but the timing of recurrence was not found to significantly correlate with survival. Of the 394 patients who were disease-free after induction BCG and thus randomized, 185 ultimately recurred and 37 went on to cystectomy, with a 5-year survival rate of 56% for greater than or equal to T2 disease, and 95% for less than or equal to T2 disease. Patients in the maintenance arm were more likely to undergo early cystectomy, at a median time of 11 months versus 24 months, although the rates of cystectomy for MIBC were similar (45% vs 49%). These data provide further evidence that those who recur after BCG, whether early or late, have worse survival than patients who have complete responses (CRs).

Cystectomy

It is difficult to determine exactly when and how to treat a patient who fails BCG. Guzzo and colleagues examined a cohort of 184 patients with cT1 or CIS who underwent cystectomy (64% with prior BCG). Nineteen percent (35 of 184) of patients were upstaged to pT2, and when compared with a group of cT2 who underwent cystectomy (and found also to be pT2), recurrence-free survival rates were lower at 3 and 5 years (64% and 61% vs 83% and 74%). Prior BCG therapy with subsequent upstaging was found to be a risk factor for long-term recurrence compared with patients with known MIBC, highlighting that BCG failure portends poor outcomes, even despite cystectomy.

Delaying radical cystectomy has been shown to increase the risk of death, which can be disconcerting to those receiving BCG, because several weeks of treatment typically elapse before response to therapy is assessed. Herr and Sogani showed that in BCG-failure patients, cystectomy within 2 years of initial BCG provided improved 15-year disease-specific survival compared with patients who had cystectomy at least 2 years after BCG (69% vs 26%; P = .003). Reviewing two historical cohorts in which different treatment paradigms were used, Raj and colleagues compared early (at T1 recurrence) with late cystectomy (at progression to MIBC) in BCG-failure patients. Although retrospective and not randomized, it showed a trend toward improved progression-free and cancer-specific survival in the early cystectomy cohort.

Conversely, Badalato and colleagues reviewed a more contemporary series of 349 patients with high-grade T1 bladder cancer, and showed that immediate cystectomy (within 90 days of diagnosis and without further TUR or intravesical therapy) did not offer a benefit in cancer-specific survival, as compared with conservative therapy. In the more recent subgroup (2000–2010), only 56.7% went on to cystectomy after attempted conservative treatment. Although retrospective in nature, this analysis suggests that delaying cystectomy may be a feasible option to consider.

The American Urological Association guidelines currently recommend radical cystectomy as first-line therapy in patients who fail BCG. The European Association of Urology guidelines state that a patient has failed BCG when a patient develops MIBC, or when high-grade NMIBC is present at both 3 and 6 months. They also recommend cystectomy as the next line of treatment in early BCG failures (in patients who can tolerate it) because of the increased risk of developing MIBC. National Comprehensive Cancer Network guidelines also recommend early cystectomy for patients with recurrence after BCG, although these guidelines do not clearly define BCG failure. Early cystectomy is widely recommended because it is superior in the oncologic control of BCG-refractory bladder cancer. However, in the patient who is not fit for cystectomy, refuses cystectomy, or has a late or low-grade recurrence after BCG, there are multiple agents for salvage therapy, which are examined in the remainder of this article.

Mitomycin C

Mitomycin C (MMC) is an alkylating agent that causes DNA cross-linking and inhibition of DNA synthesis ( Table 1 ). It is widely used in NMIBC as an adjunct to TUR in preventing tumor recurrence, and can be used postoperatively and with maintenance. However, it can cause significant chemical cystitis, and has not proved effective in the setting of BCG failure. A randomized prospective study of BCG versus MMC showed a recurrence-free rate of only 19% (4 of 21) in the group that crossed over to MMC (median follow-up, 64 months). These low success rates have been attributed in part to inadequate drug delivery and tissue penetration, leading to assistive devices that use hyperthermia and electric current, allowing for more effective uptake of drug into the urothelial cells. Both modalities have been used in patients after BCG failure with improved results.

Thermochemotherapy

Thermochemotherapy (TCT) uses local microwave-hyperthermia to the bladder, by way of a specialized catheter, to a temperature of 42C ± 2°C (Synergo, Medical Enterprises, Amsterdam, the Netherlands). The heat improves drug penetration into the urothelium, and is also cytotoxic, damaging DNA and inducing apoptosis. Nativ and colleagues reported their retrospective view of 111 patients with prior BCG failure who had undergone TCT-MMC treatment with six weekly treatments of 20 mg MMC with two consecutive 30-minute cycles of TCT, followed by maintenance therapy as warranted. Recurrence-free rates were 85% (1 year) and 56% (2 years), and the average time to recurrence was 16 months. Patients were stratified by BCG failure categories (refractory, recurrent, resistant, intolerant), and there was no significant difference in response rates. The 2-year recurrence rate for the BCG-refractory group was 56% versus 44% in the other groups. Completion of maintenance was the only factor found to correlate with lower recurrence rates. Forty-five percent of patients experienced adverse events, mostly local urinary toxicities, and 5.4% patients withdrew because of significant side effects. There is no comparison group in this retrospective review; however, a significant 2-year recurrence-free rate was seen in a high-risk group of patients.

Electromotive MMC

Another mechanism of enhancing intravesical drug efficacy is electromotive delivery. Electromotive drug administration (EMDA) involves enhancing membrane permeability and drug transport by way of electric current. MMC is not ionic at urinary pH states, and thus requires a salt-containing solution to aid with transport across urothelial membranes. Physionizer 30 (Physion, Mirandola, Italy) allows for a gradual increase in the current by a transurethral catheter, to a maximum of 23 to 25 mA, with a total treatment time of 30 minutes. EMDA-MMC has shown superiority over MMC in a randomized trial of 108 patients with CIS (solitary or concurrent T1) with response rates of 58% (EMDA-MMC) and 31% (MMC) at 6 months, and median time to recurrence of 35 versus 19.5 months. A third arm in this study received induction BCG with recurrence rates similar to that of EMDA-MMC. Long-term follow-up of 82 months showed similar equivalence of EMDA-MMC to BCG, and superiority to MMC alone with regards to recurrence, but no significant difference with overall survival or disease-specific survival.

The previously mentioned trial did not include BCG-failure patients, and so Sockett and colleagues reported on 13 BCG-failure patients who received a 6-week course of EMDA-MMC, with a 31% recurrence-free rate at 15-month follow-up. EMDA-MMC has also been evaluated in conjunction with BCG in a randomized trial of 212 patients, and was shown to be more effective in reducing recurrence, progression, and all-cause and cancer-specific mortality than BCG alone. Additionally, it has been shown to be safe and effective immediately before transurethral resection of bladder tumor (TURBT) in a BCG-naive population.

MMC’s activity in preventing tumor recurrence seems to be enhanced by TCT and EDMA technology, and is relatively well-tolerated. However, these tools have not been widely studied in the BCG-failure population, aside from Colombo and colleagues prospective trial, nor have the two modalities been directly compared with each other. There are currently no known ongoing trials evaluating enhanced MMC after BCG failure, and these treatments will likely be targeted more toward incident NMIBC.

BCG and Interferon

Because BCG acts by eliciting an immune response, the immunostimulant interferon (IFN)-α2b was introduced as another intravesical agent, showing some short-term efficacy alone. O’Donnell and colleagues published their experience in using BCG-IFN in patients with BCG failure, hypothesizing that patients failed because of an incomplete immune response, and that the addition of IFN would enhance the immunotherapeutic effect. Forty patients were given 6 to 8 weeks of induction BCG (one-third strength) and IFN, with further BCG dose reduction for patients with treatment intolerance. CR rates of 63% and 53% were seen at 12 and 24 months, respectively, and at 30-month median follow-up, 22 (55%) of 40 of patients were disease free. Initial induction treatments were well-tolerated, but treatment discontinuation was encountered with progressive maintenance treatments, and 40% of patients ended therapy early because of intolerance.

Given these promising results, a multicenter phase II trial enrolled 1106 patients to receive BCG-IFN, inclusive of all types of NMIBC, regardless of prior treatment status. BCG-naive patients received full-strength BCG-IFN, whereas reduced strength BCG was given to BCG-recurrent or BCG-intolerant subjects, with maintenance given to responders as tolerated. Of the 467 BCG-failure patients enrolled, 39% had received at least two courses of prior BCG, and 47% had recurred or persisted at 6 months, indicating a BCG-refractory state, whereas the remaining 53% were BCG-relapsing. At 24 months, 45% of prior BCG-failure patients were disease-free, and on multivariate analysis, patients with at least two prior courses of BCG had a significantly worse outcome than those with none or one courses.

Rosevear and colleagues re-examined the 231 subjects in the previously mentioned phase II trial who had CIS immediately before enrollment (63% CIS alone, 37% concurrent with Ta or T1), 52% of whom were BCG failures. Response rates at 24 months were 23% (failed ≥2 courses BCG); 57% (failed one course); and 60% (BCG-naive). On multivariate analysis, having two or more courses of BCG was a significant risk factor for BCG-IFN failure, as compared with less than or equal to one course (hazard ratio, 2.003), as was BCG refractory disease, compared with intermediate BCG recurrence (>12 months) with a hazard ratio of 1.547. BCG-IFN seems to be an effective and safe salvage therapy, displaying acceptable response rates at 2 years. However, individuals who have failed multiple courses of BCG and those with early failure are less responsive to BCG-IFN combination therapy, and should consider early cystectomy if deemed appropriate.

Gemcitabine

Gemcitabine (Gemzar) is another antineoplastic agent used systemically in many cancers and has proved effective in combination with cisplatin (with or without paclitaxel) for treating metastatic urothelial cancer, and in the neoadjuvant setting before cystectomy. Gemcitabine is a nucleoside analog and is incorporated into DNA, halting DNA synthesis, thus leading to cell death. It also irreversibly inhibits ribonucleotide reductase, preventing deoxyribonucleotide production, subsequent DNA synthesis, and repair, thereby inducing apoptosis.

A phase II trial by Dalbagni and colleagues at Memorial Sloan Kettering Cancer Center evaluated 30 patients who had failed BCG and were refusing cystectomy; 23 had CIS only. Gemcitabine was administered twice weekly for two 3-week periods with 1 week of rest in between, at a dose of 2000 mg (the highest soluble dose). At 3-month assessment, 50% (15 of 30) of patients achieved a CR, whereas 23% (7 of 30) had a partial response, defined as a negative cystoscopy and biopsy, but positive cytology. Of the 15 initial complete responders, 12 recurred at a median of 3.6 months, with one case of progression. Half of all nonresponders underwent cystectomy, whereas the other half (N = 4) had not shown disease progression at a median follow-up of 19 months. Ultimately, 10% of all patients were recurrence-free at 1 year. Eight grade 3 toxicities were reported, 75% of which were dysuria or frequency, one requiring cystectomy despite no evidence of disease.

This phase II trial enrolled a very high-risk population, with most patients being truly BCG-refractory after two prior induction cycles. A significant initial response rate was seen (50%); however, the response was not durable in most patients. Maintenance therapy was not evaluated in this trial, and many patients went on to cystectomy, making it difficult to determine the long-term rates of recurrence and progression.

Gacci and colleagues reported on a group of nine patients who had failed two BCG induction courses. This cohort received 6 weeks of intravesical gemcitabine, followed by a maintenance schedule of 3 weekly instillations at 3, 6, 12, 18, and 24 months. Again, it was well-tolerated, but six (67%) of nine patients recurred, anywhere from 4 to 19 months after onset of treatment. Two of those patients recurred with cT2 disease, and subsequently underwent cystectomy, and were alive without recurrence at last follow-up. Three patients remained recurrence-free at 13, 17, and 21 months of follow-up. Although this is a small pilot study, it confirms the tolerability of gemcitabine, and raises the possibility of using maintenance therapy for initial responders.

A larger multicenter phase II trial was conducted by Bartoletti and colleagues using intravesical gemcitabine, treating 116 patients with Ta, T1, or CIS, 40 (35%) of whom had failed BCG treatment, using the definition of persistent disease at 6 months postinduction. A total of 17.5% of patients reported adverse effects, requiring cessation of treatment in two cases. In the absence of maintenance therapy, at 1-year follow-up, 25.4% recurred at a median of 7 months. The subset analysis of the BCG failure group indicated that recurrence rates may be slightly higher, although it did not reach statistical significance. Overall, gemcitabine was more effective in patients with superficial disease and with no prior history of intravesical treatment.

In 2010, Addeo and colleagues reported results of a randomized phase III trial, comparing intravesical gemcitabine with MMC for recurrent NMIBC. The gemcitabine treatment course consisted of a 6-week induction course, followed by 10 monthly maintenance treatments for initial responders. Of 120 enrolled patients, 109 were ultimately randomized with 54 and 55 patients in each treatment arm, respectively. A total of 91 patients had prior BCG treatment, although definitions of failure were not provided. At a median follow-up of 36 months, disease-free survival was superior in the gemcitabine arm, which was a consistent finding in the pretrial grade 3 histology group. Tumor progression occurred in 10 (18%) of 55 MMC patients, and in 6 (11%) of 54 of gemcitabine patients. With regards to adverse effects, MMC had a significantly higher rate (72.2% vs 38.8%), particularly with chemical cystitis (21.1% vs 5.5%) and dysuria frequency (20% vs 9.2%). Treatment delay occurred 10% of the time with MMC, and only 5% of the time for gemcitabine. In this phase III trial with many high-risk BCG-failure patients, gemcitabine proved slightly better at preventing recurrence, and had fewer adverse effects compared with MMC.

Recently, a national phase II nonrandomized study of monthly maintenance for intravesical gemcitabine (SWOG S0353) was completed. Forty-seven patients who had failed two or more courses of BCG were treated, including patients with CIS (60%). At 3-month evaluation, 21 (45%) were disease-free, and 13 of those were continuously disease-free at the 12 months (28% of all evaluable patients). Three grade 3 toxicities were reported, whereas 30 patients had grade 1 to 2 toxicities, most of which were dysuria and urinary frequency. These data for gemcitabine maintenance therapy in a high-risk BCG-failure group indicate that a sizable portion of patients respond to therapy with a tolerable side effect profile, consistent with prior studies. However, it remains to be seen whether or not maintenance therapy adds benefit in preventing long-term progression and recurrence.

Although its efficacy has been illustrated in systemic disease, gemcitabine’s role as an intravesical agent is still being evaluated. Phase I and II trials have shown it to be safe and somewhat effective in the treatment of recurrent NMIBC. The study populations have been somewhat heterogeneous, but have included many patients who have previously failed two induction courses of BCG and are thus at highest risk of recurrence and progression. Initial response rates of up to 50% have been reported and durable responses (disease free at ≥1 year) in 28% of patients, with the addition of maintenance therapy.

Valrubicin

Valrubicin (Valstar) is the only intravesical agent currently approved by the Food and Drug Administration for the treatment of BCG-refractory NMIBC. Valrubicin was developed as a lipid-soluble semisynthetic analog of doxorubicin (adriamycin), with modifications allowing for more rapid uptake into cells, and less potential for cardiac toxicity. It interferes with nucleoside incorporation into nucleic acids, causing chromosomal damage. It is also converted in the cytoplasm to N -trifluoroacetyladriamycin, which then binds to topoisomerase II, inhibiting DNA repair and replication, RNA and protein synthesis, and ultimately results in cell cycle arrest during the G2 phase.

Greenberg and colleagues reported the first experience with intravesical valrubicin in 1997, in a trial of 32 patients, 22 (69%) of whom had CIS after BCG treatment. Most patients experienced adverse effects, mostly limited to grade 1 to 2 urinary toxicity. In short-term follow-up, a CR rate of 41% was seen. A phase I trial enrolling 22 patients evaluated the safety of immediate post-TUR administration of valrubicin. This modality was found to be well-tolerated, with minimal systemic absorption, significant only in a patient who experienced bladder perforation. Response rates were not reported.

This led to a multi-institutional study reported by Steinberg and colleagues enrolling 90 patients with CIS who had received at least one prior course of BCG. Six weekly doses of 800-mg valrubicin were administered intravesically, based on prior dose escalation studies. A CR rate of 21% (19 of 90) was observed, defined as no disease at 3- or 6-month follow-up. For these 19 patients, median time to failure or last evaluation was more than 18 months, with seven patients (8%) still disease-free at a mean of 30 months of follow-up. Forty-four patients had cystectomy, with median time to cystectomy of 24 months for the nonresponders. Six cystectomy patients had greater than or equal to pT3a disease, and two had pN+ disease. Four patients have died of bladder cancer, none of whom was a complete responder. Treatment was well-tolerated, with local urinary toxicity as the primary complaint. Based on these data, valrubicin was approved by the Food and Drug Administration for intravesical treatment of BCG-refractory bladder cancer in patients who cannot tolerate the morbidity and mortality associated with cystectomy.

Docetaxel

Taxanes are chemotherapeutic agents that stabilize microtubules, preventing cell division and subsequently cause M-phase cell cycle arrest. They are used widely in oncology, and have shown efficacy in metastatic bladder cancer. Intravesical docetaxel (Taxotere) was first studied in a phase I trial in which 18 patients with NMIBC who had failed BCG or BCG-IFN received 6 weekly instillations. Eight patients (44%) reported grade 1 to 2 toxicity, including hematuria, urgency, and dysuria, whereas none experienced grade 3 to 4 toxicity, and a maximum tolerated dose was not reached. All patients completed six treatments, and there was no drug detected on serum measurements. Results of the original cohort of patients have been recently reported with a median follow-up of 48 months. No delayed toxicity was reported, and 4 (22%) of 18 had a durable CR without maintenance or any other therapy, whereas 3 (17%) of 18 had a durable partial response, meaning a non-muscle invasive recurrence treated by local resection.

Paclitaxel

Because intravesical docetaxel was shown to be well-tolerated in patients with BCG-refractory NMIBC another taxane was investigated in a phase I trial of nab- paclitaxel (Abraxane), a nanoparticle albumin-bound formulation that facilitates drug delivery into cells. The nab technology also allows for higher concentrations of taxane to be administered, which is important to note, because the maximum tolerated dose was never reached with docetaxel. Eighteen patients with NMIBC were enrolled in this phase I dose escalation study in 2008 to 2009, all of whom had had at least one course of BCG induction, with a mean of 2.8 prior courses. Twelve of the eighteen had also received BCG-IFN therapy, and three had received MMC, whereas three had received other experimental intravesical agents. No grade 2 or higher toxicities were noted, whereas 10 of 18 patients experienced grade 1 local toxicity, none of which required cessation of treatment. There was only one instance of detectable serum levels of paclitaxel, which resolved within a week. At first follow-up assessment, 5 (28%) of 18 patients demonstrated a CR. The efficacy of nab -paclitaxel is currently being evaluated in an ongoing phase II trial ( clinicaltrials.gov ID NCT00583349 ).

An alternative drug delivery platform is hyaluronic acid (HA), a glycosaminoglycan abundant in human connective tissues. Binding HA to paclitaxel was shown to increase solubility and to increase in vitro antitumor activity, without systemic absorption in animal models. Bassi and colleagues published a phase I study of intravesical paclitaxel-HA (ONCOFID-P-B), enrolling 16 patients with BCG-refractory CIS, with 15 of 16 receiving all 6 weekly instillations. A total of seven patients reported 11 adverse events, three of which were serious, including exacerbation of pre-existing atrial fibrillation, resultant heart failure, and significant hematuria after a bladder biopsy 40 days after drug administration. At posttreatment assessment, 9 (60%) of 16 exhibited a CR. This phase I study indicates that a different formulation of paclitaxel may also be safe to administer intravesically in patients with BCG-refractory CIS; it is not clear if the significant adverse events were treatment-related. A phase II study using paclitaxel-HA for patients with G1-G2 Ta urothelial cancer is currently enrolling patients in Europe, but it excludes patients with CIS or T1 lesions, and does not require prior BCG failure, making it less applicable for patients seeking salvage therapy after BCG failure. It is hoped that it will shed more light on the overall efficacy of intravesical taxanes.

Mycobacterial Cell Wall Extract

BCG has been proved as the most effective agent in high-risk NMIBC, but carries the risk of BCG sepsis, because it is a live attenuated vaccine. To retain antitumor efficacy but reduce toxicity, the concept of using mycobacterial cell wall extract came about, and was first reported in 2001 as intravesical treatment in patients with CIS. However, the emulsion contained thimerosal, which has its own toxicities. Subsequently, a mycobacterial cell wall-DNA complex (MCC) was developed, and a multi-institutional open label study with 55 patients was conducted. MCC was administered weekly for 6 weeks, and then for 3 weeks at Weeks 12 and 24 as maintenance therapy. Twenty-eight patients in this cohort had had prior BCG therapy, whereas eight patients had no history of intravesical therapy. In the intention-to-treat analysis, response rates of 27% (4 mg) and 46% (8 mg) were seen at 12 weeks with identical rates at 26 weeks. At 18 months, disease-free rates of 23% (4 mg) and 29% (8 mg) were noted. Serious adverse events occurred at rates of 32% and 33% in the 4-mg and 8-mg of MCC groups, respectively, one of which necessitated treatment cessation.

Currently, a phase III trial is ongoing ( clinicaltrials.gov ID: NCT00406068 ) using MCC in patients with BCG-refractory disease, or recurrent high-grade NMIBC less than or equal to 2 years after initial BCG success. Eight milligrams of MCC was administered as a 6-week induction course, followed by 3 weekly instillations at 3, 6, 12, 18, and 24 months as maintenance. Interim results of 129 patients in this multi-institutional trial show a 1-year disease-free survival rate of 25%, with recurrence at a median of 177 days. Eighty-six percent of patients reported a treatment-related adverse event, mostly urinary. Two patients had to cease treatment based on adverse effects (asthenia, vomiting). These interim results show some efficacy of MCC in patients who have failed BCG, although not as effective as repeat BCG induction. Again, there is no standard control agent for this population against which to compare these agents, but MCC does seem to elicit a response in a select group of patients. MCC was recently being evaluated in a phase III open-label, randomized trial as compared with MMC for BCG-failure patients ( clinicaltrials.gov ID: NCT01200992 ). However, the trial was closed early because of poor accrual. Ultimately, MCC may be better suited as an alternative to BCG in primary treatment of NMIBC, provided it has similar efficacy and less toxicity.