INTRODUCTION — Approximately 70 percent of new urothelial (formerly called transitional cell) bladder cancer cases are classified as non-muscle invasive [1]. Non-muscle invasive bladder cancer includes Ta, T1 (submucosal invasive) tumors, and Tis (carcinoma in situ [CIS]), which account for approximately 70, 20, and 10 percent of non-muscle invasive cancers, respectively.
The rate of recurrence of non-muscle invasive bladder cancer surpasses that of all other cancers [2], and the majority of patients will experience a recurrence. Management of recurrent disease is, therefore, a critical concern in patients with non-muscle invasive bladder cancer. Determining optimal therapy, however, is complicated by the heterogeneity of disease in these patients.
Even with optimal treatment, patients with non-muscle invasive disease are at high risk of recurrence with further non-muscle invasive disease or of progression to more advanced disease. The management of recurrent or persistent non-muscle invasive disease is discussed in this topic.
The initial management and follow-up of patients with non-muscle invasive bladder cancer is discussed separately. (See "Treatment of primary non-muscle invasive urothelial bladder cancer".)
RISK OF RECURRENCE — Data on the risk of recurrence and progression in patients with non-muscle invasive bladder cancer are derived from large series that included patients with both primary and recurrent disease. These issues are discussed separately. (See "Treatment of primary non-muscle invasive urothelial bladder cancer", section on 'Risk stratification'.)
DEFINITION OF RECURRENCE STATES — Patients who have had one or more recurrences constitute a heterogeneous population, and assessment of risk is critical for proper management. Recurrence can occur in different clinical scenarios, which must be distinguished (table 1).
Recurrence without prior adjuvant intravesical therapy — Recurrence can occur after transurethral resection of bladder tumor (TURBT) in patients who have not received prior adjuvant intravesical therapy. Recurrence of low-grade papillary non-muscle invasive bladder cancer in this setting is the most common type of recurrence, since high-grade disease generally has been treated with intravesical Bacillus Calmette-Guerin (BCG) unless there is a contraindication. (See "Treatment of primary non-muscle invasive urothelial bladder cancer", section on 'Risk stratification'.)
Recurrent low-grade non-muscle invasive bladder cancer constitutes intermediate-risk disease in most cases; however, it can constitute high-risk disease if the recurrence is both large (>3 cm) and multifocal.
Whether the patient previously received a single perioperative dose of intravesical chemotherapy is not usually considered important in defining this disease state.
Recurrence after prior adjuvant intravesical therapy — Recurrence can also occur in patients who have received prior adjuvant intravesical chemotherapy or BCG therapy. In this situation, it is critical to distinguish intermediate- from high-risk, and prior intravesical chemotherapy from intravesical BCG in choosing the next line of therapy.
Standard definitions of post-BCG states have evolved to guide both clinical decision-making and clinical trial design. These definitions have developed primarily through consensus building between the US Food and Drug Administration (FDA) and the American Urological Association (AUA) [3-5].
●"BCG failure" refers to patients with either intermediate- or high-risk non-muscle invasive bladder cancer who received a single round of induction BCG without maintenance therapy [3].
●"BCG unresponsive" has replaced the older term "BCG-refractory." BCG-unresponsive non-muscle invasive bladder cancer encompasses patients with one of the following in the absence of urothelial carcinoma of the prostatic urethra or upper tract [3-5]:
•Persistent or recurrent high-grade Ta/Tis urothelial carcinoma of the bladder after completion of at least induction (≥5 doses) and one round of maintenance (or second induction; ≥2 doses) intravesical BCG.
•Persistent or recurrent high-grade T1 bladder cancer after induction BCG (≥5 doses) without further maintenance therapy.
•Initial complete response (no disease six months after diagnosis) followed by subsequent high-grade recurrence within 6 (Ta/T1) or 12 months (carcinoma in situ [CIS] with or without Ta/T1) after the last BCG dose.
•Patients with a relapse more than 6 (Ta/T1) or 12 months (CIS) after the last dose of BCG are considered to have a late relapse and would not be considered to have BCG unresponsive disease.
Time to relapse is incorporated into this definition of a post-BCG recurrence because it has been shown to correlate with the subsequent response to another challenge with intravesical BCG [6-8]. The importance of time to relapse was originally described in a study that found that BCG plus interferon alfa-2b was as effective in patients who had failed prior BCG as it was in BCG-naïve patients if the last dose of BCG was administered ≥12 months previously. In a retrospective subset analysis of 334 patients (98 with CIS and 236 with papillary disease only) who had all received at least one prior course of BCG, recurrence within six months adversely affected subsequent response to additional BCG, but recurrence between 6 and 12 months was no different than recurrence after 12 months. This held true for both CIS and papillary disease. The six-month cut-off has, therefore, been incorporated into the definition of BCG-unresponsive bladder cancer.
A low-grade papillary (Ta) recurrence does not qualify as BCG unresponsive and is generally not considered an indication to discontinue BCG therapy in a patient being treated for high-risk disease [9].
There is no standardization of disease states following intravesical chemotherapy since BCG is considered the more effective therapy.
DIAGNOSIS OF RECURRENCE — Recurrent or persistent non-muscle invasive bladder cancer is usually diagnosed by surveillance cystoscopy or urine cytology during surveillance after initial treatment. (See "Treatment of primary non-muscle invasive urothelial bladder cancer", section on 'Posttreatment surveillance'.)
Technical improvements are being evaluated to improve the efficiency of cystoscopy in detecting recurrent or persistent disease. Fluorescence (blue light) cystoscopy and narrow band imaging (NBI) are being used more widely in the initial evaluation of patients with possible non-muscle invasive bladder cancer. However, the role of these techniques is less well established for the detection and resection of recurrent non-muscle invasive bladder cancer and after BCG therapy [10,11]. (See "Treatment of primary non-muscle invasive urothelial bladder cancer", section on 'Enhanced imaging techniques'.)
Various urine biomarkers have demonstrated a substantially higher sensitivity than urine cytology but with a significantly lower specificity. Despite the promise of urine biomarkers, cystoscopy remains the procedure of choice for surveillance in previously treated patients. (See "Urine biomarkers for the detection of urothelial (transitional cell) carcinoma of the bladder".)
Evaluation of the upper urinary tracts (kidneys and ureters) as well as the prostatic urethra is also critical in the surveillance of patients with prior non-muscle invasive bladder cancer. Although this is obvious for patients with positive cytology but no visible bladder lesion, even patients with a recurrence in the bladder can have concomitant extravesical sites of recurrence, which must be identified before initiating salvage therapy. (See "Treatment of primary non-muscle invasive urothelial bladder cancer", section on 'Posttreatment surveillance' and "Malignancies of the renal pelvis and ureter".)
The importance of a detailed evaluation of the entire urothelial surface is illustrated by a study on the site of recurrence in 110 patients who had recurrent urothelial carcinoma after receiving adjuvant intravesical Bacillus Calmette-Guerin (BCG) for non-muscle invasive bladder cancer [12]. Recurrence limited to the bladder was detected in 53 cases (48 percent), while 14 (13 percent) had only upper urinary tract and/or prostatic urethral involvement, and 43 (39 percent) had both bladder and other urothelial sites of disease recurrence.
TREATMENT OF RECURRENT DISEASE — For patients with recurrent or persistent non-muscle invasive bladder cancer, a balance must be struck between the risk of progression to muscle invasive disease or distant metastasis and the risk of toxicity from treatment.
Transurethral resection of bladder tumor (TURBT) — The primary management of all recurrent non-muscle invasive bladder cancer is TURBT (for papillary tumors) or biopsy (for carcinoma in situ [CIS]).
TURBT alone — TURBT for a persistent or recurrent lesion does not differ markedly from TURBT at the initial diagnosis. (See "Treatment of primary non-muscle invasive urothelial bladder cancer", section on 'Transurethral resection'.)
Office fulguration of recurrent disease may be considered as an alternative in carefully selected patients with fewer than five, small (<0.5 cm), low-grade-appearing lesions in the context of a history of prior low-grade Ta tumors and negative cytology [13]. For patients with an established pattern of low-grade recurrences, especially if older or with significant co-morbidities, an initial period of surveillance is reasonable, with intervention only necessary if the tumor size or number increases, or the patient develops symptoms [14].
Persistent high-grade T1 bladder cancer at the time of re-TURBT after initial diagnosis of high-grade T1 disease, based on complete TUR, does not necessarily constitute recurrent disease. However, it is worth emphasizing that the risk of progression in this setting is up to 80 percent [15], so patients meeting these criteria should be considered for immediate cystectomy. (See 'Radical cystectomy' below.)
In general, there should usually be no need for repeat TURBT in patients with recurrent high-grade T1 disease after intravesical therapy because these patients should undergo radical cystectomy. If the patient refuses cystectomy or is unfit for more aggressive surgery and does not undergo radiotherapy, then re-TURBT should be considered to stage the patient optimally, minimize the residual disease, and potentially optimize subsequent courses of intravesical therapy.
Otherwise, the greatest challenge of TURBT in the setting of recurrence is the detection of CIS, either alone or in addition to papillary disease. In both instances, the CIS may provide the impetus to proceed with radical cystectomy versus additional intravesical therapy. Prior TURBT and prior intravesical therapy may increase the likelihood of false positive lesions on cystoscopy.
Single-dose perioperative chemotherapy — Single-dose perioperative intravesical chemotherapy is most efficacious for low-grade lesions at the time of initial diagnosis. The benefit of a single dose of perioperative chemotherapy in association with TURBT for recurrence is unclear [16]. Approximately one-third of the patients with urothelial carcinoma included in the S0337 trial, which demonstrated the decreased incidence of recurrence with gemcitabine in the post-TURBT setting, had recurrent disease rather than an initial occurrence [17]. (See "Treatment of primary non-muscle invasive urothelial bladder cancer", section on 'Single postoperative instillation'.)
A 2016 meta-analysis offers some information about the role of perioperative chemotherapy in the context of recurrent disease [18]. Instillation did not reduce the rate of recurrence in patients with a prior history of more than one recurrence per year. Furthermore, single-dose chemotherapy in patients with a European Organisation for Research and Treatment of Cancer (EORTC) recurrence score ≥5 was of no benefit.
The parameters that determine the EORTC recurrence and progressions scores are reproduced in the table (table 2). As an example, a patient with multifocal disease (two to seven tumors) who has recurred within less than one year has a recurrence score of 7.
No prior adjuvant intravesical therapy — First-line adjuvant intravesical Bacillus Calmette-Guerin (BCG) therapy is the standard of care for the first occurrence of high-risk disease, and subsequent intravesical therapy for recurrent or persistent non-muscle invasive bladder cancer represents salvage therapy. (See 'Salvage intravesical therapy' below.)
For intermediate-risk disease, however, many patients may not have had prior intravesical therapy, and the disease recurrence may be the impetus for adjuvant intravesical BCG or chemotherapy. Intermediate-risk non-muscle invasive bladder cancer represents a spectrum of disease, and it should not be automatic that all patients with recurrent low-grade disease need intravesical therapy. A balance must be struck between the risk of progression, on the one hand, and the risk of toxicity from treatment on the other hand.
For recurrent intermediate-risk non-muscle invasive bladder cancer, the risk of progression depends on the size of tumor, the number of sites of involvement, and the rate of recurrence (EORTC Tables) [19]. The International Bladder Cancer Group developed a consensus statement that offers a risk-adapted treatment algorithm for different patient subsets within the context of intermediate-risk non-muscle invasive bladder cancer [20]. Based on the number of risk factors that a patient with intermediate-risk disease has, they are treated as low risk (0 risk factors), intermediate-risk (1 to 2 risk factors), or high risk (3 to 4) (table 2). Those treated as low risk do not require anything beyond TURBT and optional single perioperative dose of chemotherapy, even though they are classified as having intermediate-risk disease. These include the infrequently recurring, solitary, small Ta lesions. Those treated as intermediate-risk should receive adjuvant intravesical chemotherapy or BCG including one year maintenance therapy, and those treated as high risk should receive intravesical BCG with three years of maintenance therapy.
Salvage intravesical therapy — Salvage treatment strategies for recurrent non-muscle invasive bladder cancer, including both intravesical therapy and radical cystectomy, are summarized in the table according to the different disease states (table 3).
Optimal first-line therapy will decrease the need for salvage therapy. Optimal first-line therapy includes key components, such as complete TURBT, assurance that muscularis propria is sampled in the TURBT specimen, repeat TURBT for high-grade T1 disease, use of maintenance BCG, and adequate evaluation of the upper tracts. Immediate cystectomy should be offered to the highest risk patients (eg, high-grade T1 with lymphovascular invasion [LVI] and concomitant CIS).
Furthermore, it is important not to discontinue first-line therapy too soon and label a patient as a treatment failure. This is particularly true for patients with CIS, who can have a delayed response after induction BCG and should not be deemed treatment failures until after completion of the first round of maintenance BCG. The rate of complete response at three and six months in the Southwest Oncology Group (SWOG) 8507 trial was 55 and 84 percent, respectively, for patients receiving maintenance therapy [21].
While not explicitly part of the European Association of Urology (EAU) risk classification [22], high-grade T1 bladder cancer in the context of treatment-refractory disease warrants the designation "very high-risk" disease when compared with Ta and Cis, which are classified as high-risk disease. This is reflected in the definition of BCG-unresponsive non-muscle invasive bladder cancer, which includes high-grade T1 disease after failing induction BCG alone but includes Ta and Cis only after failing induction plus the first round of maintenance BCG or a second course of induction BCG.
The very high risk associated with recurrent/persistent high-grade T1 tumor is demonstrated by two large retrospective series. In a Spanish series of 191 high-risk patients treated with TURBT and BCG, those with high-grade T1 disease at baseline and less than a complete response at three months had an 80 percent risk of subsequent progression [23]. Similarly, in a cohort of 195 patients with non-muscle invasive bladder cancer, the rate of progression (including metastasis or the development of "uncontrolled local disease") was 82 percent for patients with T1 disease at three months versus 25 percent for patients without T1 disease [24].
If a patient has BCG-unresponsive, high-risk non-muscle invasive bladder cancer, outcomes after intravesical chemotherapy with a variety of agents in the second-line setting have been uniformly poor, with two-year recurrence-free survival (RFS) rates typically below 20 percent [25].
Intravesical BCG — While further courses of intravesical therapy are not recommended for BCG-unresponsive high-grade T1 non-muscle invasive bladder cancer due to the heightened risk of progression, additional intravesical therapy can be considered for intermediate-risk and high-grade Ta/Tis disease after failure of BCG therapy.
The decision to proceed with cystectomy versus bladder-conserving therapy is guided by the balance between quality of life impairment, and morbidity and mortality of major surgery on the one hand, and the risk of recurrence, progression, and mortality with additional intravesical therapy on the other hand, all in the context of patient choice. There are only limited data on trimodality therapy in this setting, and this is currently the subject of a trial being conducted by the Radiation Therapy Oncology Group (RTOG).
Although not explicitly endorsed by guidelines, a second course of intravesical therapy is often preferred by patients due to the relatively low risk of immediate progression from high-grade Ta and CIS, compared with the risk associated with radical cystectomy. In a retrospective series, the rates of progression to muscle invasive disease and metastasis after repeat cycles of intravesical therapy in a heterogenous non-muscle invasive bladder cancer patient cohort were 7 and 5 percent after the first cycle, 11 and 14 percent after the second cycle, and 30 and 50 percent after the third cycle, respectively [26]. This evidence reinforces the approach that cystectomy is generally preferable to a third course of intravesical therapy in patients who are candidates for surgery.
Gemcitabine — Gemcitabine has had a similar level of activity compared with other salvage intravesical agents in several small studies [27-29]. Three relatively small studies have compared gemcitabine with BCG in differing populations, two of which were in patients who had failed with previous BCG therapy, and they have been summarized in a systematic review [30].
In a phase II trial, 80 subjects were randomized to gemcitabine or retreatment with BCG [27]. At two years, RFS was significantly better with gemcitabine than BCG (19 versus 3 percent). The rate of requiring radical cystectomy was similar in the two treatment arms (33 versus 38 percent).
In another phase II study, 58 patients with recurrent disease were treated with intravesical gemcitabine (weekly for six doses, then monthly to 12 months) [28]. All patients met the current definition of BCG-unresponsive disease. At three months and two years, 47 and 21 percent of patients, respectively, were free of recurrence.
In a phase III trial, 109 patients who had recurred following therapy with BCG were randomly assigned to either gemcitabine or mitomycin [31]. At a median follow-up of 36 months, fewer patients had recurred with gemcitabine compared with mitomycin (28 versus 39 percent).
Salvage intravesical regimens
Docetaxel — The efficacy of intravesical docetaxel was reported in a series of 54 patients, which included 18 from a phase I trial [32]. All but seven patients were high risk. All patients received weekly instillations for six weeks, and in the phase II trial, responders were also given monthly maintenance therapy for up to 12 months. The complete response rate was 59 percent, and the one- and three-year RFS rates were 40 and 25 percent, respectively.
Sequential gemcitabine and docetaxel — Based on retrospective studies, the intravesical administration of gemcitabine and docetaxel appears to be an effective therapy for patients with BCG-unresponsive disease and is associated with minimal toxicity [33]. Induction treatment is administered weekly for six weeks as follows: Gemcitabine (1 g in 50 mL of sterile water) is instilled into the bladder for 60 to 90 minutes, followed by sequential instillation of docetaxel (37.5 mg in 50 mL normal saline) for 60 to 120 minutes. Patients who respond to induction therapy are typically treated with monthly maintenance therapy for up to 24 months.
The efficacy of this combination was demonstrated in several retrospective studies [33-35]. As an example, one multicenter study included 276 patients with a non-muscle invasive bladder cancer recurrence after prior BCG therapy treated with gemcitabine and docetaxel [33]. One- and two-year RFS rates were 60 and 46 percent, respectively. High-grade RFS rates were 65 and 52 percent, respectively. Among the entire study population, 21 patients (8 percent) progressed to muscle invasive bladder cancer, and 43 (16 percent) underwent subsequent cystectomy. Among the 105 BCG-unresponsive patients (38 percent), two-year high-grade RFS rates were 50 percent for those with CIS and 58 percent for those with papillary disease only. Treatment was well tolerated, with only 9 percent of patients receiving a modified dosing schedule due to toxicity and 3 percent of patients not completing induction therapy.
Salvage chemohyperthermia — Salvage chemohyperthermia using either mitomycin or epirubicin has shown promising results in two retrospective studies; a prospective trial closed early due to poor accrual (NCT01094964).
In one series, 111 patients received induction and maintenance chemohyperthermia with mitomycin after failing BCG therapy. The one- and two-year RFS rates were 85 and 56 percent; 3 percent of patients progressed to muscle invasive bladder cancer [36]. In another series, 160 patients were treated with chemohyperthermia using either mitomycin or epirubicin [37]. The one- and two-year RFS rates were 60 and 47 percent, and 4 percent of patients progressed to muscle invasive disease.
Valrubicin — Valrubicin, a semisynthetic analogue of doxorubicin, is an intravesical agent approved by the US Food and Drug Administration (FDA) specifically for patients with BCG-unresponsive disease, but studies have demonstrated limited complete response rates (18 percent) [31,38]. Valrubicin has not been directly compared with other salvage intravesical agents in randomized trials.
The activity of intravesical valrubicin was established in a phase II single-arm study of 90 patients with CIS who had failed prior BCG therapy with or without subsequent other intravesical therapies [38]. There was an 18 percent complete response rate at six months and an 8 percent disease-free rate at 30 months. An additional trial in 80 patients who were refractory to or intolerant of BCG also observed an 18 percent complete response rate, with 10 and 4 percent of patients recurrence free at one and two years, respectively [39].
Electromotive administration of mitomycin — Electromotive administration (EMDA) of mitomycin sequentially with BCG has not been adequately evaluated in the salvage setting. The potential role of this modality as initial therapy is discussed separately. (See "Treatment of primary non-muscle invasive urothelial bladder cancer", section on 'Alternative administration approaches'.)
Experimental therapies — A number of other intravesical agents are under active investigation in patients with recurrent non-muscle invasive bladder cancer, but are not yet available for routine clinical practice:
●N-803 – N-803, an IL-15 superagonist antibody cytokine fusion protein, is effective in patients with BCG-unresponsive disease, with complete response rates of up to 71 percent in patients with CIS [40,41].
N-803 was evaluated in combination with intravesical BCG in an open-label, nonrandomized phase II/III trial (QUILT 3.032) of 161 patients with heavily treated BCG-unresponsive non-muscle invasive bladder cancer (84 patients with CIS with or without Ta/T1 (papillary) disease, and 77 patients with papillary disease alone) [40,41]. Patients were treated intravesically with 50 mg of intravesical BCG plus 400 micrograms of N-803 weekly for six doses, followed by maintenance therapy for up to two years.
•CIS – At median follow-up of 24 months, patients with CIS had a complete response rate of 71 percent (58 of 82 evaluable patients). Among those achieving a complete response, the probability of complete response durability lasting 24 months or longer was 53 percent, and the median duration of complete response was 26 months [41].
•Papillary disease – Among patients with papillary disease, at median follow-up of 21 months, the 24-month disease-free survival (DFS) and bladder cancer specific overall survival were 48 and 99 percent, respectively. In addition, 72 of 77 patients (94 percent) avoided radical cystectomy [41].
Treatment was well-tolerated since N-803 activity is localized to the bladder, with treatment and immune-related adverse event rates of 1 and 0 percent, respectively. Grade ≥3 toxicities included dysuria, pollakiuria, hematuria, myalgias, arthralgias, urinary tract infection, decreased urine flow, bacteremia, sepsis, and encephalopathy (all <1 percent) [41].
●Nadofaragene firadenovec – Nadofaragene firadenovec, a novel recombinant adenovirus vector encoding the interferon alfa-2b gene, has efficacy in patients with BCG-unresponsive disease, with complete response rates of up to 53 percent [42,43].
This agent was evaluated in a single-arm phase III trial of 151 patients with BCG-unresponsive non-muscle invasive bladder cancer [43]. Among the 103 patients with CIS, complete responses were achieved in 55 patients within three months of treatment (53 percent); additionally, the 12-month high-grade RFS rate was 24 percent. Among the 48 patients with papillary disease, the 3- and 12-month high-grade RFS rates were 73 and 44 percent, respectively.
●Oportuzumab monatox – Oportuzumab monatox (VB4-845), a recombinant fusion protein that targets tumor cells expressing the epithelial cell adhesion molecule (EpCAM), has efficacy in patients with BCG-unresponsive disease, with complete response rates of up to approximately 40 percent [44-46].
Oportuzumab monatox is composed of a humanized single-chain variable fragment of an anti-EpCAM antibody conjugated to a truncated Pseudomonas exotoxin A payload. In an open-label phase II trial of 46 patients with BCG-unresponsive CIS, the 3- and 12-month complete response rates were 40 and 16 percent, respectively [45]. Similar results were seen in preliminary results of a single-arm phase III trial (VISTA) of 133 patients with BCG-unresponsive non-muscle invasive bladder cancer [46,47]. Among 82 patients with BCG-unresponsive CIS, the 3- and 12-month complete response rates were 39 and 17 percent.
●CG0070 – CG0070 is a modified adenovirus under the control of the E2F-1 promoter that selectively targets high-grade bladder cancer cells with deficient expression of the retinoblastoma (Rb) tumor suppressor gene. CG0070 also carries the gene for granulocyte-macrophage colony-stimulating factor (GM-CSF), which is released at the time of virus-induced cell lysis and augments immune response to tumor antigens. In a single-arm phase II trial of 45 patients with BCG-unresponsive non-muscle invasive bladder cancer, the six-month complete response rate was 58 percent in patients with CIS and 50 percent in patients with CIS with or without Ta/T1 disease [48].
Immune checkpoint inhibition
Pembrolizumab — The programmed cell death protein 1 (PD-1) inhibitor pembrolizumab is approved by the FDA for patients with high-risk, BCG-unresponsive CIS of the bladder who are ineligible for or decline radical cystectomy [49].
Clinicians who offer this approach should discuss the benefits and risks of treatment as well as the method of administration. Pembrolizumab has modest durable complete response rates compared with other available agents. Pembrolizumab also results in grade ≥3 toxicities in approximately one-third of patients and requires intravenous administration in an infusion center by a medical oncologist. Further data are needed to determine how to incorporate pembrolizumab into standard clinical practice in the context of other salvage therapy options. (See 'Salvage intravesical therapy' above.)
The efficacy of pembrolizumab was demonstrated in a single-arm phase II trial (KEYNOTE-057) that included 96 patients with BCG-unresponsive CIS, with or without papillary disease [50]. Pembrolizumab was administered at 200 mg intravenously every three weeks for up to two years in the absence of disease progression, recurrence, or treatment-related toxicity. At median follow-up of 36 months, the complete response rate was 41 percent (39 patients) at three months. Among these 39 patients, 18 patients (46 percent, or 19 percent of the total study population) remained in complete response for 12 months or longer. No patients had progression to muscle invasive bladder cancer or metastatic disease while on study. Among the 40 patients with persistent or recurrent disease following pembrolizumab who underwent cystectomy, three had muscle invasive (T2) disease on pathology. The grade ≥3 toxicity rate was 13 percent, including hyponatremia (3 percent), arthralgias (2 percent), dermatitis, malaise, and pruritus (1 percent each).
Atezolizumab — The programmed death ligand 1 (PD-L1) inhibitor atezolizumab has clinical efficacy in patients with BCG-unresponsive non-muscle invasive cancer of the bladder either with or without CIS, based on a single-arm phase II trial (SWOG 1605). In preliminary results of this study, among 75 patients with CIS of the bladder, atezolizumab demonstrated a complete response rate of 27 percent at six months [51], with durable responses noted among the 20 patients who achieved a complete response (median response duration of 15 months, one-year complete response rate of 49 percent) [52]. Additionally, among the 55 patients without CIS of the bladder (ie, Ta and/or T1 disease only), atezolizumab demonstrated an 18-month event-free survival of 47 percent [52]. Atezolizumab remains investigational for routine treatment of this patient population.
The efficacy of immunotherapy in patients with metastatic urothelial carcinoma is discussed separately. (See "Treatment of metastatic urothelial cancer of the bladder and urinary tract", section on 'Second-line therapy (immunotherapy)'.)
Radical cystectomy — Radical cystectomy is only rarely required in patients with intermediate-risk disease that is uncontrollable with the combination of intravesical therapies and transurethral bladder tumor resections. Radical cystectomy is, however, considered the standard of care for any patient with BCG-unresponsive high-grade non-muscle invasive bladder cancer [22].
Even in the best surgical hands, radical cystectomy entails high morbidity (up to 60 percent) [53] and risk of mortality (2.7 to 2.9 percent) [54]. Many patients are medically unfit for cystectomy, and others refuse the surgery. In all cases, cystectomy has a significant impact on quality of life. The optimal timing of cystectomy due to failure of intravesical therapy is, therefore, controversial. The potential benefits of preserving the bladder must be balanced against the risks of progression to muscle invasive and even metastatic disease, with a resultant decreased survival.
●For patients who recur with high-risk disease beyond 12 months after the last dose of intravesical therapy, data from one study suggested that another course of intravesical therapy (preferentially BCG) should be considered [6]. More recent data from the same group suggest that six months is a better cut-off than 12 months [7,8]. This rule is also often applied to patients with high-grade T1 tumors at the time of recurrence, but as described above, the risk of progression is particularly high in these patients, and cystectomy should always be favored. If an additional course of BCG is administered, cystectomy should be performed promptly if it fails.
●For patients with high-grade Ta tumors or CIS recurring within six months of the last dose of BCG, cystectomy is the most definitive therapy and likely has the highest cure rate, but it also represents overtreatment in many cases. Although the guidelines call for cystectomy in this setting, it is generally considered safe in clinical practice to attempt one additional course of intravesical therapy before proceeding to cystectomy.
A major objective of bladder preservation strategies is to avoid muscle invasive disease. Multiple patient series and a systematic review suggest that patients with progression to muscle invasive bladder cancer after BCG have a worse prognosis compared with patients with stage-matched primary muscle invasive bladder cancer [55].
In an analysis based upon a comparison with a historical cohort, preemptive radical cystectomy was associated with better disease-specific survival than postponing cystectomy until the appearance of T2 disease [56]. In patients who progress to muscle invasive disease, the bladder cancer specific death rate at 10 years has been reported to be approximately 50 percent [57,58].
Multiple series suggest that survival is worse in patients who are treated with continued bladder-sparing strategies compared with earlier cystectomy [59-62]. In a Spanish study, patients with high-risk non-muscle invasive bladder cancer progressed to muscle invasive disease in 3.3 percent of cases at six months and 8 percent at 12 months [61]. The mortality from bladder cancer was 0.4 percent at six months, 1 percent at one year, and 2.5 percent at 1.5 years. The authors concluded that the risk of death from surgery matched the risk of death from the bladder cancer at 1.5 years. Other studies have also shown improved cancer specific survival with earlier radical cystectomy compared with delayed intervention [59,62].
For patients undergoing radical cystectomy for recurrent or persistent disease, the value of the extended pelvic lymph node dissection has not been adequately studied. Specific recommendations are, therefore, impossible to make. These patients have also not been included in the prospective randomized trials. The risk of occult lymph node metastasis is 10 to 15 percent for clinical T1 disease [63-66]. Considering that 40 to 50 percent of patient with high grade T1 disease who fail BCG are upstaged at radical cystectomy, it is reasonable to perform an extended lymphadenectomy in this setting.
Radiation therapy — Radiation is sometimes administered for high-grade T1 disease. A Medical Research Council (MRC) trial indicated no advantage of external beam radiation alone compared with intravesical BCG in the first-line setting [67]. The group in Erlangen, however, has achieved good results with radiation in conjunction with radiosensitizing chemotherapy in the BCG naïve setting [68].
Definitive trimodal therapy should be considered as an option for BCG-unresponsive high-grade T1 bladder cancer for patients who refuse or are unfit for cystectomy. An ongoing clinical trial is evaluating the efficacy of trimodal therapy in patients who have failed or are intolerant of intravesical chemotherapy or BCG and are considered candidates for radical cystectomy (Radiation Therapy Oncology Group [RTOG] 0926; NCT00981656).
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Bladder cancer".)
SUMMARY AND RECOMMENDATIONS
●Adherence to therapy for non-muscle invasive bladder cancer – Non-muscle invasive bladder cancer is associated with a high risk of recurrence or progression to more advanced disease. However, optimal first-line therapy reduces the rate of recurrence and the need for salvage therapy. Adherence to standard treatment guidelines is therefore critical. (See "Treatment of primary non-muscle invasive urothelial bladder cancer".)
●Recurrent non-muscle invasive bladder cancer – Recurrent non-muscle invasive bladder cancer represents a heterogenous disease mix with a wide spectrum of risk for additional recurrence and progression. Risk-adapted therapy is therefore recommended (table 2 and table 3). Prior intravesical therapies also influence treatment decisions. (See 'Definition of recurrence states' above.)
●Treatment of recurrent disease based on risk-adapted therapy – Recurrent low-grade Ta disease is considered intermediate risk in most cases, but this risk can be stratified based upon four adverse prognostic features (table 2 and table 3):
•No adverse features – Patients with intermediate-risk non-muscle invasive bladder cancer who lack adverse prognostic features do not need adjuvant intravesical therapy and can be treated instead with transurethral resection of bladder tumor (TURBT) and a single dose of perioperative chemotherapy. (See 'Transurethral resection of bladder tumor (TURBT)' above and 'Single-dose perioperative chemotherapy' above.)
•One to two adverse features – Patients with intermediate-risk non-muscle invasive bladder cancer with one to two adverse prognostic features should receive either intravesical chemotherapy or intravesical Bacillus Calmette-Guerin (BCG; with 12 months maintenance). Such patients can be treated with multiple rounds of intravesical therapy before considering cystectomy. If BCG was administered first line, then intravesical chemotherapy can be used second line, and vice versa (See 'Salvage intravesical therapy' above.)
•Three to four adverse features – Patients with intermediate-risk non-muscle invasive bladder cancer with three to four adverse prognostic features should receive intravesical BCG with three years of maintenance therapy. (See 'Salvage intravesical therapy' above.)
●BCG-unresponsive CIS
•Salvage intravesical therapy – For patients with BCG-unresponsive carcinoma in situ (CIS) who are ineligible for or decline radical cystectomy, we offer salvage intravesicular therapies rather than other systemic agents. Treatment options include gemcitabine and docetaxel (either as single agents or in combination), chemohyperthermia, and valrubicin. (See 'Salvage intravesical therapy' above and 'Salvage intravesical regimens' above.)
•Pembrolizumab – Immunotherapy with the programmed cell death protein 1 (PD-1) inhibitor pembrolizumab is an option for patients with high-risk, BCG-unresponsive CIS of the bladder who are ineligible for or decline radical cystectomy. The decision to use pembrolizumab is dependent on patient preference and a discussion of the risks of treatment and the need for intravenous administration. (See 'Immune checkpoint inhibition' above.)
•Experimental agents – Additional experimental intravesical approaches (eg, nadofaragene firadenovec, oportuzumab monatox, N-803) are being evaluated. (See 'Experimental therapies' above.)
●Indications for radical cystectomy for recurrent disease
•Although radical cystectomy is recommended for recurrent high-grade Ta disease and CIS after prior BCG therapy, a second cycle of intravesical therapy is often administered. Radical cystectomy is preferred over a third cycle of intravesical therapy in any patient who is eligible for this surgery because the risk of progression rises with each additional cycle of intravesical therapy. (See 'Radical cystectomy' above.)
•Radical cystectomy should be performed in any patient with recurrent/persistent high-grade T1 at the three-month time point. BCG-unresponsive, recurrent, high-grade T1 bladder cancer is considered to be very high risk for additional recurrence and progression, especially if associated with lymphovascular invasion (LVI), CIS or large volume and is an indication for radical cystectomy. Until the Radiation Therapy Oncology Group (RTOG) clinical trial matures, trimodal therapy is an appropriate alternative in patients who refuse or are unfit for cystectomy. (See 'Radical cystectomy' above and 'Radiation therapy' above.)
