As outlined by IOM standards, all financial relationships of expert panel members that might result in actual, potential, or perceived conflicts of interest were individually reported. Disclosures were made prior to the onset of manuscript development and updated on an annual basis. In addition, panel members were asked to articulate any actual or potential conflicts at all key decision points during guideline development, so that participants would understand all possible influences, biases, and/or the diversity of perspectives on the panel. Although some degree of relationships with outside interests are to be expected among experts, panel candidates with significant financial connections that may compromise their ability to fairly weigh evidence (either actual or perceived) were not eligible to participate in guideline development.

Recognizing that guideline panel members are among the leading experts on the subject matter under consideration and guideline recommendations should have the benefit of their expertise, any identified potential conflicts of interests were managed as outlined in SITC’s disclosure and conflict of interest resolution policies. As noted in these policies, panel members disclosing a real or perceived potential conflict of interest may be permitted to participate in consideration and decision-making of a matter related to that conflict, but only if deemed appropriate after discussion and agreement by the expert panel.

The financial support for the development of this guideline was provided solely by SITC. No commercial funding was received.

Panel recommendations are based on literature evidence, where possible, and clinical experience, where appropriate.10 Consensus for the recommendations herein was generated by open communication and scientific debate in small- and whole-group settings, surveying and responses to clinical questionnaires, as well as formal voting in consensus meetings.

For transparency, a draft of this CPG was made publicly available for comment during the development process and prior to publication. All comments were evaluated and considered for inclusion into the final manuscript according to the IOM standard.

The evidence- and consensus-based recommendations of the panel were refined throughout the development process in order to obtain the highest possible agreement among the experts, however, the minimum threshold was defined as 75% approval among the voting members. Evidence supporting panel recommendations was graded according to the Oxford Centre for Evidence-Based Medicine (OCEBM) Levels of Evidence Working Group “The Oxford Levels of Evidence 2” (2016 version). A summary of the OCEBM grading scale may be found below (table 1). The level of evidence (LE) for a given recommendation is expressed in parentheses following the recommendation (eg, LE: 1). Recommendations without an associated LE were based on expert consensus.

BCG is a live, attenuated strain of Mycobacterium bovis that is administered intravesically as a therapy for NMIBC. BCG has been an important option for the management of NMIBC for more than four decades. BCG treatment depends on the risk category of NMIBC. Low-risk disease is defined as single, primary, low-grade tumors (G1, Ta).53 For patients with low-risk NMIBC, the SOC treatment is TUR followed by intravesical chemotherapy, without the use of BCG or other immunotherapies.54 The International Bladder Cancer Group (IBCG) defines intermediate-risk bladder cancer as multiple and/or recurrent low-grade Ta tumors, with specific factors including number of tumors (> one), size of tumors (>3 cm), timing of recurrence (within 1 year), rate of recurrence (> one per year), and prior treatment.55 High-risk NMIBC is defined as any T1, high-grade (G3) or CIS disease.53

In SWOG-8795 (also identified as INT-0094 or EST-1888), 447 patients with NMIBC were administered BCG or mitomycin C as adjuvant therapy following surgical resection. Among the study population, 377 patients had Ta/T1 NMIBC, and median follow-up was 2.5 years. In patients with Ta/T1 NMIBC, mitomycin C treatment led to a recurrence-free survival (RFS) rate of 43%, while those treated with BCG exhibited an RFS rate of 54% (HR 1.41; 95% CI 1.06 to 1.88; p=0.017). It was also noted that local and systemic adverse events (AEs) of grade 1 or 2 were more common in patients treated with BCG (p=0.003).56

During the Nijmegen study, 437 patients were treated with TICE-BCG, the Rijksinstituut voor Volksgezondheid en Milieuhygiene (RIVM) strain of BCG, or mitomycin C as adjuvant therapy following surgical resection of NMIBC (for papillary lesions, n=387) or biopsy (for CIS, n=50). Despite the fact that the BCG regimen used was suboptimal (ie, no maintenance therapy was given) and the cohort in the mitomycin C arm received monthly maintenance, at 2-year follow-up, the estimated rate of DFS for patients with papillary lesions was similar across groups, at 65% (95% CI 60% to 70%) for mitomycin C treatment, 54% (95% CI 49% to 59%) for TICE-BCG treatment, and 62% (95% CI 57% to 67%) for RIVM-BCG treatment. The differences between treatment arms were not statistically significant.57

An analysis of six pooled phase II clinical trials in which 119 patients with CIS received TICE-BCG as first-line therapy (6 weekly instillations for induction, 12 monthly instillations for maintenance) found an ORR of 75.6%, and complete response (CR) rate of 45.4% at a median follow-up of 47 months. The median PFS was estimated at ≥48 months.58 On the basis of the clinical trials discussed above, in August 1998, the FDA approved the use of TICE-BCG for the first-line and adjuvant treatment of CIS, and for the adjuvant treatment of Ta and/or T1 papillary tumors of the bladder following TUR.59 Several meta-analyses have subsequently concluded that BCG prevents, or at least delays, progression to invasive disease in high-risk or intermediate-risk disease. The largest of these meta-analyses analyzed 24 trials (n=4,863 patients) and showed a 27% reduction in the odds of progression (9.8% vs 13.8%; odds ratio (OR) 0.73; p=0.001) in patients treated with maintenance BCG compared with either TUR alone or TUR with chemotherapies other than mitomycin C.60–62 The meta-analyses illustrated that BCG is only superior to mitomycin C in situations where BCG maintenance is provided.60 62

Historically, there have been efforts to administer BCG in conjunction with recombinant interferons. However, a meta-analysis demonstrated that BCG alone was associated with lower risk of recurrence in comparison to BCG with interferon-α-2a (relative risk (RR) 0.57; 95% CI 0.39 to 0.82) and to BCG with interferon-α-2b (RR 0.42; 95% CI 0.30 to 0.59).63

The standard BCG dosing regimen contains an induction and a maintenance phase, with induction consisting of BCG instillation once a week for 6 weeks and maintenance consisting of repeat instillations at set time intervals (at 3, 6, 12, 18, 24, 30, and 36 months postinduction). Maintenance instillations occur once a week for 3 weeks at each time interval. This dosing regimen (the ‘6+3 regimen’) is supported by the results of EORTC-30962 (NCT00002990), a phase III clinical trial that randomized patients to receive full-dose or one-third-dose BCG with 1 year or 3 years of maintenance instillations. In this trial, no differences in toxicity were identified between one-third-dose and full-dose BCG. Further, for intermediate-risk patients (defined as lower than pT1 and lower than G3 in this study), no significant difference in the 5-year DFS rate was observed between 1 year and 3 years of maintenance. For high-risk patients (at least pT1 or G3 disease), 3 years of maintenance was superior to 1 year of maintenance by percentage of disease-free patients at 5 years for patients receiving full-dose BCG (HR 1.61; 95% CI 1.13 to 2.30; p=0.0087).64

A complicating factor is that BCG is currently subject to ongoing global shortages, which has impacted the ability of healthcare providers to provide BCG therapy to patients.65 66 In light of this ongoing shortage, several professional and advocacy societies have modified their guidelines for management of NMIBC and advocated a risk-stratified approach toward BCG administration.67–70 For this reason, current guidelines recommend a risk-stratified schedule of maintenance instillations: while patients with high-risk disease should receive a full 3-year course of maintenance, patients with intermediate-risk disease may receive shortened courses of 1 year of maintenance therapy.68 71 72

BCG treatment is associated with a number of potential AEs, which are cumulative over the course of BCG therapy, including cystitis, dysuria, frequency of urination, and, more rarely, infections or systemic side effects. Vigilance is important, since these symptoms overlap with other common AEs in patients receiving treatment for bladder cancer, including urinary tract infections, sensitivity related to urinary catheterization, or overactive bladder. AEs stemming from BCG therapy may be addressed through temporary withholding of BCG, conventional treatments (including systemic steroids,73 74 non-steroidal anti-inflammatory drugs (NSAIDS),74 anti-tuberculosis antibiotics,75 and quinolone antibiotics76 77), or permanent withdrawal of BCG for severe toxicity.78 A proinflammatory response to BCG, leukocyturia, is associated with both an increase in self-reported AEs79 80 and response to BCG.81 It may be the case that the occurrence of systemic side effects is indicative of an immune response to BCG and thus could be used as a marker of response to therapy, although more data is required on this subject.

One well-studied method to reduce BCG-related toxicity is dose reduction. Clinical trials have demonstrated that reduced-dose BCG may be efficacious, with significantly less toxicity than full-dose BCG. Long-term follow-up of 499 patients in one prospective trial found that while patients with multifocal tumors derived more benefit from the standard dose than the reduced dose (p=0.0151), the cause-specific survival at 5 years did not differ between the two arms (p=0.76). Patients who received reduced doses of BCG were significantly less likely to experience grade ≥3 toxicity (p<0.001).82 However, in the EORTC-30962 trial, there were no differences in toxicity between the one-third-dose and full-dose arms.64 Reduced dosing frequency, however, leads to inferior outcomes. The NIMBUS study (NTR4011) evaluated reduced frequency instillation in a randomized trial of 824 BCG-naive patients. Patients who received reduced frequency instillation experienced an increased rate of recurrence, at 27% compared with 12% for patients receiving standard BCG treatment (HR 0.40; 95% CI 0.24 to 0.67).83 While research is ongoing regarding possible alternative dosing regimens, the 6+3 regimen remains the gold standard based on current data.

BCG-unresponsive NMIBC is a term that encompasses both BCG-refractory and BCG-relapsing (within 6 months of last BCG exposure) NMIBC, as defined by the IBCG.84 The FDA has also issued guidance on BCG-unresponsive NMIBC with more specific criteria, defined as at least one of the following: (1) persistent or recurrent CIS (with or without recurrent Ta/T1 disease) within 12 months of completion of adequate BCG therapy, (2) recurrent high-grade Ta/T1 disease within 6 months of completion of adequate BCG therapy, or (3) T1 high-grade disease at the first evaluation following BCG induction. Adequate BCG therapy is defined as at least 5 of 6 doses of an initial induction course with at least 2 additional doses (either of maintenance therapy or of a second course of induction).85 Multiple clinical trials have been conducted using this guidance to define this at-risk patient population, for which the only FDA-approved systemic immunotherapy is pembrolizumab.

Pembrolizumab is approved for the treatment of high-risk, BCG-unresponsive CIS at a dose of 200 mg every 3 weeks or 400 mg every 6 weeks (an alternative treatment schedule approved across multiple indications) for up to 24 months. The approval is based on favorable outcomes for patients in cohort A of the phase II KEYNOTE-057 trial (NCT02625961), who had high-risk, BCG-unresponsive CIS with or without papillary disease and were treated with pembrolizumab every 3 weeks for up to 24 months. Patients were required to undergo full resection of papillary disease prior to their first dose of pembrolizumab. The primary endpoint of CR was assessed at 3 months, and a key secondary endpoint was DOR.86 At first analysis, 41 of 102 patients achieved CR at 3 months (40.2%; 95% CI 30.6% to 50.4%) and the median DOR of those patients who achieved CR was 12.7 months.87 At the 12-month landmark analysis for DOR (which occurred approximately 15 months from the start of pembrolizumab therapy), the rate of patients with observed DOR ≥12 months was 46% (18 of 39 initial complete responders) and 19% of all treated patients.42 None of the patients developed muscle-invasive or metastatic disease while on protocol treatment. Based on these data, in January 2020, the FDA approved the use of pembrolizumab for the treatment of BCG-unresponsive, high-risk CIS (with or without papillary tumors).3 Cohort B of KEYNOTE-057, which focuses on patients with fully resected papillary disease only (without concomitant CIS), was ongoing at the time of publication with a primary endpoint of RFS.86

Investigational strategies to overcome BCG-unresponsiveness include concomitant immunomodulation with recombinant cytokines. As an example, N-803 is a mutant IL-15-based immunostimulatory fusion protein complex (IL-15RαFc) that selectively promotes proliferation and activation of natural killer (NK) cells and CD8⁺ T cells. In one cohort of a phase II/III trial that enrolled 80 patients with BCG-unresponsive CIS for intravesical administration of N-803 with BCG, the CR rate at any time was 72% (n=51/71) and the probability of maintaining CR for 12 months was 59%, with a median CR duration of 19.2 months (range 7.6–26.4) months.88

A recent phase III, open-label, multicenter US trial (NCT02773849) showed antitumor efficacy with intravesical nadofaragene firadenovec (rAD-IFNα2b), a replication-defective adenoviral gene transfer vector that delivers interferon-α-2b expression to the bladder epithelium. In the trial, patients with BCG-unresponsive CIS with or without Ta/T1 disease achieved a CR rate at 3 months of 53.4% (95% CI 43.3% to 63.3%) with 45.5% of these remaining free of high-grade recurrence at 12 months. A similar trend of durable response and RFS was observed in patients with papillary high-grade Ta/T1 BCG-unresponsive NMIBC, with 43.8% of patients remaining recurrence-free at 12 months. Progression to muscle invasion occurred in 5% of the CIS cohort and 6% of the high-grade Ta/T1 cohort. Among the patients with CIS, 29% underwent cystectomy by 12 months, as did 21% of those with high-grade Ta/T1 disease. The cystectomy-free survival at 24 months for the whole cohort was 64.5%. Of patients who underwent cystectomy, 3 of 32 (9.3%) in the CIS cohort were found to have pT2 or higher stage disease.89 While nadofaragene firadnovec is not FDA-approved for the treatment of BCG-unresponsive NMIBC, at the time of writing, it has been granted priority review by the agency and previously received Fast Track and Breakthrough Therapy Designations.

Neoadjuvant cisplatin-based chemotherapy is the current SOC in MIBC. While there are no immunotherapies currently approved as neoadjuvant therapy for localized MIBC, immunotherapy, alone and in combination with chemotherapy, has shown efficacy in MIBC in phase II trials.40 41 90

Based on the phase II data, there are several ongoing randomized phase III trials evaluating the role of perioperative immunotherapy in MIBC and results from these trials may establish the utility of this approach. For cisplatin-eligible MIBC, three randomized phase III trials are active at the time of manuscript preparation: KEYNOTE-866 (pembrolizumab, gemcitabine, and cisplatin vs gemcitabine and cisplatin), NIAGARA (durvalumab, gemcitabine, and cisplatin vs gemcitabine and cisplatin), and ENERGIZE (nivolumab, gemcitabine, and cisplatin with or without linrodostat vs gemcitabine and cisplatin). For cisplatin-ineligible MIBC, KEYNOTE-905 is evaluating pembrolizumab with or without EV followed by radical cystectomy versus radical cystectomy alone and the PIVOT IO 009 (NCT04209114) trial is evaluating nivolumab with or without bempegaldesleukin (NKTR-214) followed by radical cystectomy versus radical cystectomy alone. In these trials, immunotherapy agents are continued after radical cystectomy as well. Further, the KEYNOTE-B15 phase III trial, which will examine EV with pembrolizumab versus gemcitabine with cisplatin in patients with cisplatin-eligible MIBC, is anticipated to begin enrollment soon.

Atezolizumab was tested in a randomized phase III trial as adjuvant therapy after radical surgery in patients with muscle-invasive (including node-positive) urothelial carcinoma. This trial, IMvigor010, enrolled patients with ≥ypT2 disease and/or nodal involvement at radical surgery after neoadjuvant chemotherapy or ≥ypT3 disease and/or nodal involvement if they did not receive prior neoadjuvant chemotherapy. Patients were randomized after radical surgery to receive either atezolizumab for 1 year or observation. There were no differences in DFS between the arms, which was the primary endpoint. The trial continues to be followed for OS.44

Nivolumab has also been evaluated in the adjuvant setting in the randomized placebo-controlled phase III CheckMate 274 trial, which met its co-primary DFS endpoints in the ITT population and the group with tumors with elevated PD-L1 expression in patients who had undergone radical surgery for muscle-invasive (including node-positive) urothelial carcinoma. Initial results from 709 patients (353 randomized to nivolumab including 140 with tumors PD-L1 ≥1% and 356 randomized to placebo including 142 with tumors PD-L1 ≥1%) demonstrated a statistically significant and clinically meaningful improvement in median DFS with adjuvant nivolumab after radical surgery compared with placebo, at 21.0 months (range 17.1–33.4) vs 10.9 months (range 8.3–13.9) in the ITT population (HR 0.70; 98.31% CI 0.54 to 0.89; p=0.0006) and median DFS not yet reached (range 22.0 to not estimable) vs 10.8 months (range 5.7 to 21.2) in the tumor PD-L1 ≥1% group (HR 0.53; 98.87% CI 0.34 to 0.84; p=0.0004). Significant improvement was also seen for the secondary endpoint of non-urothelial tract RFS for adjuvant nivolumab in both the ITT group (HR 0.72; 95% CI 0.58 to 0.89) and in patients with tumors with PD-L ≥1% (HR 0.54; 95% CI 0.38 to 0.77).45 The FDA granted priority review status to the Biologics License Application for nivolumab for adjuvant treatment of patients with surgically resected, high-risk MIBC in April 2021.

Pembrolizumab is also being tested in a randomized phase III trial as adjuvant therapy after radical surgery in patients with muscle-invasive (node-positive) urothelial carcinoma. This trial, AMBASSADOR, is enrolling patients with ≥ypT2 disease and/or nodal involvement at radical surgery after neoadjuvant chemotherapy, or ≥ypT3 disease after no neoadjuvant chemotherapy and will randomize patients to 1 year of pembrolizumab or observation. Accrual is still ongoing.

Immunotherapy in the neoadjuvant setting has been tested in phase I and II trials, some of which have completed accrual and have reported results (see table 4). Approaches include ICI as monotherapy before surgery as well as ICI in combination with another ICI or platinum-based chemotherapy. These trials have reported pathologic complete responses (pCRs), also referred to as ypT0, in a subset of patients at the time of surgical intervention, a finding that has correlated with better long-term survival in prior studies of neoadjuvant chemotherapy. Neoadjuvant ICI therapy has been examined in the cisplatin-ineligible (NABUCCO, HCRN GU14-188 Cohort II, NCT02812420) and cisplatin-eligible (BLASST-1, HCRN GU14-188 Cohort I, DUTRENEO) patient populations.

PD-L1 positivity of the tumor by CPS (IHC 22C3 assay) correlated with pembrolizumab responses in the PURE-01 study and ipilimumab and nivolumab in the NABUCCO study. Additionally, in the DUTRENEO trial, for patients in the PD-L1-high group (as measured by the E1L3N XP antibody) the ypT0 rate was 57.1%, while it was 14.3% for patients in the PD-L1-low group.91 However, all other monotherapy and combination studies in table 4 did not identify a significant response correlation with tumor or IC PD-L1 status.

Overall, ICI monotherapy or combination therapy has been reported to have manageable toxicity and has not been associated with delay of cystectomy. These promising data support ongoing phase III trials, but are not yet practice-changing. Comparisons across these small-moderate sized single-arm phase II trials, however, should be approached cautiously. Significant differences in initial staging techniques, type and length of treatment, biomarker assays, and eligible patient populations exist across these studies. Additionally, the potential correlation between pathologic assessment after neoadjuvant immunotherapy regimens and OS (and/or RFS) remains unclear.

Radiation and chemotherapy cause immunogenic cell death,92 93 which may synergize with ICI therapy to potentiate antitumor responses. Two randomized phase III trials are investigating concurrent anti-PD-(L)1 therapy in combination with external beam radiation and radiosensitizing chemotherapy. The randomized, phase III trial NCT03775265 (SWOG/NRG 1806) is testing atezolizumab and the randomized, phase III KEYNOTE-992 (NCT04241185) is testing pembrolizumab in this setting. In addition, the phase II, randomized NCT03768570 is examining durvalumab following trimodal therapy (surgery, chemotherapy, and radiotherapy).

The phase II, single-group assignment clinical trial IMvigor210 examined the efficacy of atezolizumab for the treatment of mUC. In cohort I of the trial (NCT02951767), 119 cisplatin-based chemotherapy-ineligible patients received first-line atezolizumab.96 The ORR was 24% (95% CI 16% to 32%), and the median DOR was not reached at 2-year follow-up.31 97 On the basis of ORR and DOR data from both cohorts of the IMvigor210 trial, the FDA approved atezolizumab for the first-line treatment of PD-L1-positive mUC in patients who are ineligible for cisplatin-containing chemotherapy or those who are not eligible for any platinum-based chemotherapy, regardless of PD-L1 status, in April 2017.5

In the KEYNOTE-052 (NCT02335424) phase II trial, 374 patients with mUC who were ineligible for cisplatin-based chemotherapy received pembrolizumab as first-line treatment, irrespective of PD-L1 status.98 At the most recent updated analysis, the ORR was 29% and the median DOR was 30.1 months.30 Subgroup analysis of the PD-L1-negative (CPS<10) and PD-L1-positive (CPS≥10) groups showed higher DOR and OS in the PD-L1-positive group. A follow-up report of long-term outcomes found that 2-year OS was 31.2%.30 Based on OS and ORR data from KEYNOTE-045 (described in the Immunotherapies for R/R mUC section) and ORR and DOR data from KEYNOTE-052, in May 2017, the FDA approved pembrolizumab for use as a first-line treatment of mUC (in patients who are ineligible for cisplatin-based chemotherapy and PD-L1-positive, or any patient ineligible for platinum-based chemotherapy) and for the treatment of R/R mUC (in patients who have experienced disease progression following platinum-based chemotherapy) regardless of PD-L1 status.3

The activity of PD-(L)1 inhibitors for the first-line treatment of cisplatin-ineligible patients with mUC raised the hypothesis of whether PD-(L)1 blockade as single-agent therapy should be extended to cisplatin-eligible patients as well and whether regimens combining platinum-based chemotherapy and PD-(L)1 blockade might further improve outcomes. IMvigor130 was a placebo-controlled phase III trial randomized 1:1:1 to test whether atezolizumab monotherapy or atezolizumab with gemcitabine and platinum (cisplatin or carboplatin) improved survival compared with placebo plus gemcitabine and platinum. The study enrolled 1,213 patients and demonstrated longer PFS for atezolizumab plus chemotherapy compared with chemotherapy alone (8.2 months vs 6.3 months; stratified HR 0.82; 95% CI 0.70 to 0.96; one-sided p=0.007). The interim analysis for OS showed a trend toward longer OS with the atezolizumab plus chemotherapy combination, which was not statistically significant (HR 0.83; 95% CI 0.69 to 1.00; one-sided p=0.027). The median OS for atezolizumab monotherapy was 15.7 months compared with 13.1 months for chemotherapy. In patients with high levels of PD-L1 expression in tumor-infiltrating ICs (IC2/3 by SP142 assay), atezolizumab monotherapy median OS was not estimable at interim analysis (95% CI 17.7 to not estimable) vs 17.8 months (95% CI 10.0 to not estimable) in the chemotherapy-alone group (stratified HR 0.68; 95% CI 0.43 to 1.08). As atezolizumab versus chemotherapy could not be formally compared at this interim analysis due to the hierarchical statistical analysis plan, these arms will be further evaluated at future analyses. In patients with tumors harboring low PD-L1 expression levels (IC0/1), the median OS was 13.5 months (95% CI 11.1 to 16.4) in the atezolizumab monotherapy group vs 12.9 months (95% CI 11.3 to 15.0) for patients treated with chemotherapy (unstratified HR 1.07; 95% CI 0.86 to 1.33).32 As noted in the Diagnostic tests and biomarkers for urothelial cancer immunotherapy section, however, an unplanned analysis prompted by the Data and Safety Monitoring Committee revealed increased early deaths in patients with low PD-L1 expressing tumors treated with atezolizumab compared with platinum-based chemotherapy—a result that became apparent even before the trial completed accrual and results were publically available.22 32 This led the FDA and EMA to restrict the label for atezolizumab monotherapy for cisplatin-ineligible patients to only those having tumors with high levels of PD-L1 expression (IC2/3 by Ventana SP142 assay) or, in the US only, patients considered platinum-ineligible (unable to receive even carboplatin) regardless of PD-L1 expression status.

The KEYNOTE-361 trial (NCT02853305) was a phase III randomized trial that assigned 1,010 patients to receive pembrolizumab monotherapy, pembrolizumab with chemotherapy, or SOC chemotherapy. The final results of KEYNOTE-361 revealed no significant improvement in PFS or OS with pembrolizumab plus platinum-based chemotherapy versus platinum-based chemotherapy or with single-agent pembrolizumab versus platinum-based chemotherapy.33 99 In parallel with events unfolding in the IMvigor130 trial, as described above and in the Diagnostic tests and biomarkers for urothelial cancer immunotherapy section, the FDA restricted the prescribing label for pembrolizumab when used in the cisplatin-ineligible frontline setting to patients with PD-L1 expression ≥10 by CPS or to those who were considered platinum-ineligible (unable to receive carboplatin) regardless of PD-L1 expression. Importantly, neither IMvigor130 nor KEYNOTE-361 were designed to specifically compare single-agent PD-(L)1 blockade versus carboplatin-based chemotherapy in patients ineligible for cisplatin, the current labeled indication for front-line treatment of mUC.

DANUBE (NCT02516241) was a phase III trial that randomized 1,126 patients to receive first-line durvalumab with tremelimumab, durvalumab monotherapy, or SOC chemotherapy. DANUBE did not reach its co-primary endpoints of an improvement in OS for durvalumab versus chemotherapy in patients with PD-L1-high tumors or with tremelimumab plus duravlumab versus chemotherapy in the ITT population.100

First-line platinum-based chemotherapy results in disease control in mUC in 65%–75% of patients, but PFS and OS are relatively short, with median PFS less than 9 months.94 101 While significant progress has been made with five anti-PD-(L)1 inhibitors being approved by the FDA for mUC patients who progress after platinum-based chemotherapy, response rates with second-line anti-PD-(L)1 inhibitors are modest and only a minority of patients obtain durable clinical benefit. Moreover, there is a substantial patient drop off from first-line to second-line therapy in mUC, with only approximately 40% of patients with mUC receiving second-line therapy as seen from patterns of care and outcomes in real-world settings.102 103 The current ‘watch-and-wait’ approach for patients with mUC following response or stable disease after first-line platinum-based chemotherapy and prior to initiation of second-line treatment is suboptimal as almost all patients progress within 9 months of active therapy.94 Maintenance therapy with anti-PD-(L)1 inhibitors after cessation of first-line platinum-based chemotherapy is an attractive treatment strategy in mUC to improve patient outcomes. Historically, maintenance therapy approaches with targeted therapies have not been successful in mUC.104 105 However, the antitumor activity and relatively favorable safety profile of PD-(L)1 antagonists in mUC make them a potentially attractive option for maintenance therapy.

Two recent trials have evaluated the efficacy of single-agent checkpoint inhibition after completion of first-line systemic platinum-based chemotherapy (HCRN GU14-182 and JAVELIN Bladder 100).37 106 In the HCRN GU14-182 trial, a phase II study of maintenance pembrolizumab versus placebo after first-line chemotherapy, 108 patients with mUC achieving at least stable disease after up to 8 cycles of first-line platinum-based chemotherapy were enrolled and treated with pembrolizumab or placebo for up to 24 months. The primary endpoint was PFS. Significantly longer PFS was achieved with maintenance pembrolizumab (5.4 months; 95% CI 3.1 to 7.3) compared with placebo (3.0 months; 95% CI 2.7 to 5.5). Median OS was 22 months (95% CI 12.9 to not reached) with pembrolizumab and 18.7 months (95% CI 11.4 to not reached) with placebo, a secondary endpoint for which the trial was not adequately powered and did not reach statistical significance.106

The JAVELIN Bladder 100 trial was a phase III, multicenter, multinational, randomized, open-label, parallel-arm study investigating first-line maintenance treatment with avelumab plus best supportive care (BSC) versus BSC alone in patients with mUC who did not have disease progression after first-line platinum-containing chemotherapy. A total of 700 patients were randomly assigned to receive either avelumab plus BSC or BSC alone, and the primary endpoint was OS. Median OS with avelumab and BSC was significantly longer compared with BSC alone (21.4 vs 14.3 months, respectively; HR 0.69; 95% CI 0.56 to 0.86; p=0.001). Median PFS was 3.7 months with avelumab and BSC (95% CI 3.5 to 5.5) compared with 2 months with BSC alone (95% CI 1.9 to 2.7; HR 0.56; 95% CI 0.40 to 0.79; p<0.001).38 Based on the results of this trial, the FDA approved avelumab as maintenance therapy for patients with locally advanced or mUC that has not progressed with first-line platinum-containing chemotherapy. It is important to note that the upfront chemotherapy and maintenance avelumab was approved based a randomized phase III trial, representing a higher LE than the approvals for pembrolizumab and atezolizumab, which were based on phase II data.

Cohort II of the IMvigor210 trial (NCT02108652) enrolled 310 patients with mUC who had experienced disease progression following platinum-based chemotherapy and explored the activity of atezolizumab.107 The ORR was 16% (95% CI 13% to 21%), and the median DOR was 27.7 months (95% CI 2.1 to 33.4).17 97 On the basis of ORR and DOR data from both cohorts of the IMvigor210 trial, the FDA granted accelerated approval to atezolizumab for the treatment of R/R mUC in May 2016.5

A randomized open-label phase III trial, IMvigor211, randomized patients with disease progression following platinum-based chemotherapy to either atezolizumab or investigator’s choice of chemotherapy (single-agent paclitaxel, docetaxel, or vinflunine (European Union only)). The primary endpoint of this trial was OS in patients with PD-L1-high expression on tumor infiltrating ICs (IC 2/3 by SP142 assay). Of the 931 patients randomized, 234 were PD-L1-high. In that group, there was no significant difference in OS (stratified HR 0.87; 95% CI 0.63 to 1.21; p=0.41). Based on the study design, no additional formal analyses were performed, though an OS benefit was observed with atezolizumab versus chemotherapy (regardless of PD-L1 status) in the ITT population in an exploratory analysis. Atezolizumab therapy was associated with fewer grade 3 or 4 AEs and numerically longer DOR than chemotherapy. These results were comparable to those of IMvigor210 cohort II.28 Because IMvigor211 failed to meet its primary OS endpoint, however, the indication for atezolizumab in patients with mUC who have previously received platinum-based chemotherapy was voluntarily withdrawn in March 2021.

Avelumab received accelerated approval from the FDA in May 2017, for the treatment of R/R mUC based on the JAVELIN Solid Tumor (NCT01772004) phase I expansion trial, which enrolled 242 patients.4 108 The FDA based its decision on an ORR of 17% (95% CI 11% to 24%) and DOR data (median not reached at 6 months).26 In an updated safety and efficacy analysis with more than 2 years of follow-up, ORR was 16.5% (95% CI 12.1% to 21.8%), median DOR was 20.5 months (95% CI 9.7 to not reached), and the 24-month OS was 20.1% (95% CI 15.2% to 25.4%) with avelumab.109

Study 1108 (NCT01693562), a phase II trial, examined the efficacy of durvalumab in 191 patients with R/R mUC (182 of which had previously received platinum-based chemotherapy).25 110 The ORR was 20.4% (95% CI 13.1% to 29.5%) and DOR (median not reached at 1 year) data from this trial formed the basis of FDA-accelerated approval for the use of durvalumab to treat R/R mUC (that had progressed following platinum-based chemotherapy) in May 2017.111 112 In November 2020, however, the FDA indication for durvalumab for use in previously treated patients with locally advanced or metastatic bladder cancer was voluntarily withdrawn because the phase III DANUBE trial did not meet its primary end points.

Nivolumab was granted accelerated approved by the FDA for the treatment of R/R mUC for patients with disease progression following platinum-based chemotherapy in February 2017.6 This approval was based on ORR (20.7%; 95% CI 16.1% to 26.1%) and DOR (median 20.3 months; 95% CI 11.5 to 31.3) data from the phase II CheckMate 275 (NCT02387996) trial, which evaluated nivolumab monotherapy in 386 patients with R/R mUC (270 of these patients had experienced disease progression after platinum-based chemotherapy).113 114 Extended follow-up of this trial confirmed the safety and efficacy data previously reported.21

The KEYNOTE-045 (NCT02256436) phase III trial compared pembrolizumab to chemotherapy for 542 patients with mUC who had experienced disease progression after prior platinum-based chemotherapy.115 At updated long-term follow-up (median 28 months), patients treated with pembrolizumab exhibited a statistically significant advantage compared with chemotherapy in 2-year OS rates (26.9% vs 14.3%; HR 0.70; 95% CI 0.5 to 0.85; p=0.00015) and ORR (21.1% vs 11.0%; p=0.002). However, there was no significant difference in PFS at the 1-year or 2-year landmarks.3 116 Median DOR was not reached with pembrolizumab at this updated analysis, but was 4.4 months for chemotherapy.