Document resource
WHAT IS ALREADY KNOWN ON THIS TOPIC Radical cystectomy remains the standard of care for bacillus Calmette–Guérin (BCG)-unresponsive non-muscle-invasive bladder cancer, but this surgery is highly invasive and leads to a substantial decline in the patient’s quality of life. Therefore, the development of bladder-sparing therapies is a pressing global issue.WHAT THIS STUDY ADDS Although an increase in Food and Drug Administration-guided clinical trials has led to the recent approval of four new therapeutic agents, this study systematically reveals a critical gap in current treatment development: these approvals were primarily based on the results of single-arm trials targeting carcinoma in situ, indicating a lack of comparative data. Randomised controlled trials (RCTs) are needed to assess true efficacy.HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY To address the remaining unmet needs in existing and developing treatments, there will be an increased necessity for RCTs specifically in the setting of BCG-exposed patients.Introduction Bladder cancer is the ninth most common cancer worldwide, with an estimated 614 298 new cases diagnosed each year globally. 1 Approximately 75% of newly diagnosed bladder cancers are non-muscle-invasive bladder cancer (NMIBC),2 which is generally managed by transurethral resection of bladder tumours (TURBT), followed by intravesical instillation. Patients diagnosed with high-risk NMIBC and carcinoma in situ (CIS) are generally recommended to receive intravesical bacillus Calmette–Guérin (BCG) therapy according to major clinical guidelines.2–4 However, some patients develop resistance, and the efficacy of repeated intravesical BCG treatments in these resistant cases is limited;5 therefore, radical cystectomy is then considered. Major clinical guidelines recommend RC for patients with recurrent high-risk NMIBC after BCG,2–4 which trades off quality of life (QoL) mainly due to the accompanying urinary diversion against better oncological outcomes.6 7 In daily clinical practice, it is determined through shared decision-making whether individual patients are selected or not for radical cystectomy (RC). Previously, effective options of bladder-sparing therapy (BST) for patients with high-risk NMIBC who had experienced BCG failure have been scarce. Although clinical trials aimed at bladder preservation have been conducted, no agents have received approval, with the exception of valrubicin, largely attributed to factors such as the varied prior BCG exposures of enrolled patients and a lack of consistent evaluation endpoints across trials. Moreover, the efficacy of valrubicin itself proved to be limited.8 The difficulty in conducting clinical trials for BST for BCG-unresponsive high-risk NMIBC arises primarily from the lack of clearly defined trial populations and evaluation criteria, as well as the ethical challenges associated with establishing appropriate control groups, as effective treatment options are limited, except RC. Regarding the definition of trial populations, the most widely accepted criterion is the ‘BCG-unresponsive’ classification proposed by the Food and Drug Administration (FDA),9 which is based on the types of BCG failure categorised by the International Consensus Panel.10 ‘BCG-unresponsive’ specifically refers to a group comprising both BCG-refractory and BCG early-relapse cases, for whom RC is generally recommended. Because of the lack of effective control arms other than RC, the FDA currently permits single-arm trials for patients with BCG-unresponsive CIS and recommends using the complete response (CR) as the primary endpoint.9 In patients without CIS, who are presumed to have undergone complete tumour resection by TURBT, randomised controlled trials (RCTs) using survival-based endpoints such as recurrence-free survival (RFS) are recommended. Patients without CIS should not be included in clinical trials in which the primary endpoint is CR.9 For successful BST, several therapeutic options have been developed and clinical trials of novel agents are currently ongoing. We conducted a systematic review of such novel intravesical, local and systemic therapies for NMIBC after BCG failure. Effective BST strategies would offer more favourable oncological outcomes with better QoL for more patients at high risk of disease progression.Methods Eligibility criteria Studies were eligible for inclusion when they met all of the following criteria: (1) prospective interventional studies, including RCTs or single-arm studies; (2) studies evaluating treatments for patients experiencing recurrence of NMIBC after BCG therapy or BCG-unresponsive CIS; (3) studies reporting RFS or CR rates as outcome measures; and (4) studies available in English. Studies were excluded if they met any of the following criteria: (1) studies including patients with muscle-invasive bladder cancer (MIBC) (≥pT2); (2) studies targeting paediatric populations; (3) retrospective studies; (4) review articles; or (5) case reports. This systematic review was not registered in any database.Information sources A systematic search of PubMed, Cochrane Library and ClinicalTrials.gov from inception to 30 June 2025 was conducted based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist. 11Search strategy A literature search was performed using “NMIBC” OR “(non-muscle-invasive bladder cancer)” OR (“BCG” AND (“unresponsive” or “refractory” or “relapse” or “recurrence”)).Selection process Initially, the titles of articles or clinical trials were screened based on the eligibility criteria. Subsequently, each study was assessed for study design, duplicates were removed and studies were selected according to the predefined eligibility criteria. At this stage, phase 1 or phase 2 studies were excluded if a phase 3 study evaluating the same intervention was available, and similarly, phase 1 studies were excluded if corresponding phase 2 studies were identified. Clinical trials requiring further evaluation regarding eligibility were reviewed and resolved through discussion between two urologists.Data collection process Information regarding study design, clinical trial status, inclusion criteria related to previous treatments of enrolled patients, therapeutic agents and their administration methods, evaluation methods and outcomes, if available, were extracted.Data items The inclusion criteria, planned enrolment, actual enrolment, primary endpoints and current status of each clinical trial were collected.Study risk of bias assessment To assess the risk of bias, the Methodological Index for Non-Randomized Studies (MINORS) was used for single-arm studies, 12 while the Risk of Bias 2.0 (RoB 2) tool was applied to RCTs.13 No automation tools were used in the process.Effect measures The outcome measures collected included CR rates for CIS and RFS for NMIBC other than CIS.Results Study selection We reviewed 2493 studies (Cochrane Library n=1243, ClinicalTrials.gov n=445, PubMed n=805), of which 114 met the eligibility criteria. As shown in figure 1, initial screening involved the exclusion of duplicate National Clinical Trial (NCT) numbers (n=125) and non-English reports (n=62). Subsequently, duplicate clinical trials and publication titles were removed (n=478). Of the remaining 1828 reports, 114 relevant trials were selected according to the specified eligibility criteria.Figure 1PRISMA 2020 flow diagram for systematic review. BCG, bacillus Calmette–Guérin; NCT, National Clinical Trial; NMIBC, non-muscle-invasive bladder cancer.Study characteristics Forty-nine trials have been completed, of which 37 have been published. Forty-five trials are currently ongoing, and 20 trials were either of unknown status, discontinued or withdrawn ( online supplemental table 1). Online supplemental figure 1 shows trends of clinical trials according to types of therapeutic agents. Cytotoxic anticancer agents and immunotherapy and biological response modifiers constituted the majority of eligible studies, followed by immune checkpoint inhibitors (ICIs), tyrosine kinase inhibitors (TKIs), antibody–drug conjugates (ADCs), radiation and photodynamic therapy (online supplemental figure 1). Studies on ICIs and ADCs increased in these 10 and 5 years, respectively (online supplemental figure 2). All eligible studies are listed in online supplemental tables 1-3. 59 studies in which results have been reported are listed in tables 1–3. Of the 114 eligible studies, only 19 (17%) were RCTs; most of the remaining were early-phase or single-arm studies. The median MINORS score for the 34 trials was 10.5 (maximum 16 points) (online supplemental table 4). The quality assessment showed that 28 trials (82%) were of moderate quality (9–14 points), while six (18%) were of poor quality (below 8 points), and none were of good quality (15 or 16 points). The risk of bias assessment showed that, among the five reported RCTs, one was rated as low risk, three as having some concerns and one as high risk (online supplemental table 5).SP410.1136/bmjonc-2025-000753.supp4Supplementary dataSP110.1136/bmjonc-2025-000753.supp1Supplementary dataSP210.1136/bmjonc-2025-000753.supp2Supplementary dataSP510.1136/bmjonc-2025-000753.supp5Supplementary dataSP610.1136/bmjonc-2025-000753.supp6Supplementary dataTable 1Outline, study design and outcome of reported clinical trials on cytotoxic anticancer agentsNo.Start/published yearTrial name/PMID (Ref.))orNCT numberStudy design/phaseStudy statusTreatment agents/competitorIntervention methodCIS/papillaryOutcomeCytotoxic anticancer drug11993/20001068797218SAPublishedValrubicinIVeCIS is necessaryCR at any time 21%21998/2009ECOG E-3897/1863947019SAPublishedValrubicinIVeCIS+papillaryRFS at 12 m 20%32020/SunRIse-1/NCT04640623RCTInterim reportedTAR-200+cetrelimabIVe+DIVCIS is necessaryCR at 12 m 56.7%TAR-200IVeCIS is necessaryCR at 12 m 52.4%CetrelimabDIVCIS is necessaryCR at 12 m 22.8%TAR-200IVePapillary onlyRFS at 9 m 81.1%62006/2013SWOG S0353/2359745225SAPublishedGem 2 g×6IVeCIS+papillaryRFS at 12 m 28%8/20102016270626RCTPublishedGem 2 g×12 (6 w) + maintenanceIVeCIS+papillaryRFS at 24 m 19%BCG+maintenanceIVeCIS+papillaryRFS at 24 m 3%9/20101985871024SAPublishedGem 2 g×6 + maintenanceIVeCIS+papillaryRFS at 12 m about 38%10/20061678291322SAPublishedGem 2 g×6IVeCIS+papillaryCR at 12 m about 10%11/20071798703523SAPublishedGem 2 g×12IVePapillary onlyRFS at 12 m about 80%12/20071984133027RCTPublishedGemIVePapillary onlyRFS at 12 m about 85%MMCIVePapillary onlyRFS at 24 m about 65%142009/201628RCTPublishedGem+MMCIVeUnknownCR at any time 84%MMCIVeUnknownCR at any time 66%152010/20172851615629SAPublishedGem+everolimusIVe+oralCIS+papillaryRFS at 12 m 20%18/20172816308631SAPublishednab-PTXIVeCIS+papillaryCR at any time 36%19/20112116751778SAPublishedONCOFID-PBIVeCIS is necessaryCR at any time 60%202017/20223464977279SAPublishedONCOFID-PBIVeCIS is necessaryCR at 12 m 40%24/20101991389032SAPublishedDocIVeCIS+papillaryRFS at 12 m 61%252022/NCT06732531SACompletedBH011IVeCIS+papillaryCR at 12 m 50%28/20142449076280SAPublishedMMC hyperthermia (BSD-2000)IVeCIS+papillaryRFS at 12 m about 70%29/20152602681881SAPublishedMMC hyperthermia (UniThermia)IVePapillary onlyRFS at any time 59%302009/2019HYNM/3027469936RCTPublishedMMC hyperthermia (Synergo)IVeCIS+papillaryRFS at 24 m 35%BCG or chemotherapyIVeCIS+papillaryRFS at 24 m 41%332000/NCT04311580SAPublishedEMDA MMCIVeCIS+papillaryRFS at 12 m 54%34/20183052244537SAPublishedEMDA MMCIVeCIS+papillaryRFS at 12 m 50%352018/NCT03672240SACompletedNitroxolineOralCIS+papillary362014/NCT04498702SACompletedNitroxolineOralCIS+papillaryRFS at 12 m 54.3%BCG, bacillus Calmette–Guérin; CIS, carcinoma in situ; CR, complete response; CT, clinical trial; DIV, drip intravenous infusion; Doc, docetaxel; EMDA, electromotive drug administration; Gem, gemcitabine; IVe, intravesical therapy; MMC, mitomycin C; NCT, national clinical trial; PMID, PubMed ID; PTX, paclitaxel; RCT, randomised control trial; RFS, recurrence-free survival; SA, single arm trial.Table 2Outline, study design and outcome of reported clinical trials on immunotherapy and biological response modifiersNo.Start/published yearTrial name/PMID (Ref.))orNCT numberStudy designStudy statusTreatment agents/competitorIntervention methodCIS/papillaryOutcomeImmunotherapy and biological response modifiers372016/20213325364139SAPublishedNadofaragene FiradenovecIVeCIS+papillaryCR at 12 m 24.3%412017/2022QUILT-3.032/3832001141SAPublishedNAIIVeCIS+papillaryCR at 12 m 45%422020/BOND-003/NCT04452591SAInterim reportedCretostimogene GrenadenorepvecIVeCIS is necessaryCR at 12 m 28%432020/2024CORE-001/38 844 794 or43SAPublishedCretostimogene Grenadenorepvec+PemIVe+DIVCIS is necessaryCR at 12 m 57%452021/LEGEND/NCT04752722SAInterim reportedDetalimogene VoraplasmidIVeCIS is necessaryCR at 6 m 47%46/20001103920282SAPublishedIFNa-2IVeUnknownCR at any time 80%471999/NCT00004122SACompletedIFNα−2b+BCGIVeCIS+papillary48/20112178805045SAPublishedIFNα + BCGIVeCIS is necessaryRFS at 6 m 66%522008/2019NCT0069479846SAPublishedCADI-05SCCIS+papillaryRFS at 15 m 35%532012/NCT01625260SAInterim reportedALT-801+GemDIVCIS+papillaryCR at 18 m 33%542006/2017EN3348−301/2814993748SAPublishedMCNAIVeCIS+papillaryRFS at 12 m 34.8%562015/2022SAKK 06/14/3501288950SAPublishedVPM1002BCIVeCIS+papillaryRFS at 15 m 49.3%602024/NCT06351904SAInterim reportedRAG-01IVeCIS+papillaryRFS at any time 66.7%61/20233813936451SAPublishedOncoTheradIVeCIS+papillaryRFS at 12 m 86%62/1996869364683SAPublishedBropirimineoralCIS is necessaryCR at any time 50%632014/2021NCT02009332SACompletedABI-009IVeCIS+papillary652021/NCT05126472SAInterim reported2141-V11IVeCIS+papillaryCR at any time 45%661998/NCT00006034RCTCompletedKeyhole Limpet HemocyaninIVeCIS is necessaryno result data availableDoxorubicinIVeCIS is necessaryno result data available672013/2021NCT02015104RCTCompletedPANVAC+BCGIVe+SCCIS+papillaryRFS at 12 m 44.4%BCGIVeCIS+papillaryRFS at 12 m 42.9%BCG, bacillus Calmette–Guérin; CIS, carcinoma in situ; CR, complete response; CT, clinical trial; DIV, drip intravenous infusion; IFN, interferon; IVe, intravesical therapy; NAI, nogapendekin alfa inbakicept; NCT, national clinical trial; PMID, PubMed ID; RCT, randomised control trial; RFS, recurrence free survival; SA, single arm trial; SC, subcutaneous injection.Table 3Outline, study design and outcome of reported clinical trials on immune checkpoint inhibitors, tyrosine kinase inhibitors and othersNo.Start year/published yearTrial name/PMID (Ref.))or NCT numberStudy designStudy statusTreatment agents/competitorIntervention methodCIS/papillaryOutcomeImmune checkpoint inhibitor732016/KEYNOTE-057/3405117753SAPublishedPembrolizumabDIVCIS is necessaryCR at 12 m 18.8%742017/ADAPT-BLADDER cohort 4/NCT03317158SAInterim reportedDurvalumab+Gem/DocDIVCIS+papillaryCR at 12 m 68.6%762017/20223600819354SAPublishedPembrolizumab+BCGIVeCIS+papillaryRFS at 12 m 22%792017/NCT02901548SACompletedDurvalumabDIVCIS is necessaryCR at 6 m 23.1%802018/20253982702184SAPublishedDurvalumabIVeCIS+papillaryRFS at 6 m 39%832018/2024NCT03258593SACompletedDurvalumab+VicineumDIV+IVeCIS+papillaryRFS at 12 m 20%852016/20233680384085SAPublishedAtezolizumab+BCGDIVCIS is necessaryCR at 6 m 37.5%862017/2024SWOG S1605/3759619186SAPublishedAtezolizumabDIVCIS is necessaryCR at 6 m 27%872019/ABC Trial/NCT03892642SACompletedAvelumab+BCGDIVCIS+papillaryCR at 3 m 67%902021/NCT04706598SAInterim reportedCamrelizumabIVeCIS+papillarymedian RFS 12.68 mTyrosine kinase inhibitor952020/THOR-2 cohort 1/NCT04172675RCTInterim reportedErdafitinibOralPapillary only*RFS at 12 m 77%Gem or MMC or hyperthermic MMCOral+IVePapillary only*RFS at 12 m 41%2020/THOR-2 cohort 2/NCT04172675SAErdafitinibOralCIS is necessary*CR at 8 m 72.7%982016/2020NCT02657486SACompletedInfigratinibOralPapillary only*RFS at 7 w 75%1002013/2018HCRN 12–157/2793241658SAPublishedDovitinibOralCIS+papillary*CR at 6 m 8%1012008/20193123178560SAPublishedSunitinibOralCIS+papillaryRFS at 3 m 44%Cytotoxic agent other than anticancer drugs1082015/2020VISTA / NCT02449239SACompletedVicineumIVeCIS+papillaryCR at 12 m about 21%Radiation1102009/2021RTOG 0926/3922651466SAPublishedRT 61.2 Gy/34 fr+CDDP/MMC/ 5-FUDIV+RTPapillary is necessaryRFS at 24 m 55.9%Photo dynamic therapy1142019/NCT03945162SAInterim reportedTLD-1433IVeCIS is necessaryCR at 12 m 35%*FGFR mutation positive.BCG, bacillus Calmette– Guérin; CDDP, cisplatin; CIS, carcinoma in situ; CR, complete response; CT, clinical trial; DIV, drip intravenous infusion; Doc, docetaxel; FGFR, fibroblast growth factor receptor; 5-FU, 5-Fluorouracil; Gem, Gemcitabine; IVe, intravesical therapy; MMC, mitomycin C; NCT, national clinical trial; PMID, PubMed ID; RCT, randomised control trial; RFS, recurrence free survival; RT, radiation therapy; SA, single arm trial; SC, subcutaneous injection.Novel agents or modalities for NMIBC after BCG failure Figure 2 summarises status of clinical trials for individual patients with NMIBC after BCG failure.Figure 2Summary of clinical trial status for individual agents (excluding combination therapies). FDA, Food and Drug Administration.Currently, the FDA-approved therapeutic agents for NMIBC after BCG failure include valrubicin,8 pembrolizumab, nadofaragene firadenovec, nogapendekin alfa inbakicept (NAI) in combination with BCG14 and TAR-200.15 All of these have been approved for CIS, with or without concomitant papillary tumours; however, no established treatment option currently exists for patients with papillary tumours without CIS, highlighting the need for future drug development in this population. In this section, novel agents or therapeutic modalities are introduced by treatment or modality type.In recent years, the development of device-assisted chemotherapy for bladder cancer treatment has advanced. Within this review, we specifically mention TAR-200 (Gemcitabine) and TAR-210 (fibroblast growth factor receptor 3 (FGFR3) inhibitors). These systems are designed to be placed within the bladder via a urethral catheter and aim to maintain a consistent intravesical drug concentration by releasing the agent slowly over a 3 week period. These devices are anticipated to be inserted and removed in an outpatient setting.16Cytotoxic anticancer drugs Valrubicin Valrubicin is an anthracycline-based cytotoxic agent that was approved by the FDA in 1998 for the treatment of BCG-refractory CIS of the bladder. 17 Two prospective clinical trials were conducted,18 19 reporting CR rates of 21% and 24% in patients with CIS. Retrospective real-world data demonstrated RFS rates of 51.6% at 3 months, 30.4% at 6 months and 16.4% at 12 months in patients with BCG-unresponsive NMIBC receiving valrubicin.20 An international study analysing actual treatments for BCG-unresponsive high-risk NMIBC from 2011 to 2018 revealed that its use was limited to a small proportion of cases, specifically 10.9%.21Gemcitabine Intravesical gemcitabine (Gem) therapy was evaluated in multiple clinical trials conducted in the USA and Italy in the early 2000s, demonstrating RFS rates of approximately 50%. 22–25 In two RCTs reported in 2010, Gem was compared with BCG and mitomycin C (MMC) in the initial BCG failure26 and BCG-naïve setting,27 respectively, and both studies demonstrated significant improvement in RFS. An RCT demonstrated the superiority of a Gem and MMC regimen over MMC monotherapy in terms of CR rates; nevertheless, the extent of BCG refractoriness in the study population was not clearly defined,28 and this trial was assessed as high risk according to the ROB2 (online supplemental table 5). Currently, a phase 2 clinical trial of sequential MMC and Gem therapy, aligned with the FDA definition of BCG-unresponsive disease, is underway in South Korea (NCT06388720). A clinical trial conducted in the United States evaluating Gem in combination with everolimus was associated with a high rate of treatment discontinuation due to toxicity (NCT01259063), although this trial was assessed as poor quality according to the MINORS criteria (online supplemental table 4).29 In the SunRISe-1 clinical trial, TAR-200, a Gem-releasing intravesical device, has demonstrated promising outcomes including a 1-year CR rate of 56.7% in combination with cetrelimab, an anti-PD-1 antibody, and a 1-year CR rate of 57.4% as monotherapy for BCG-unresponsive CIS30 (NCT04640623). The ongoing SunRISe-5 trial is currently evaluating TAR-200 in patients with recurrent, papillary-only, high-risk NMIBC after BCG failure, in comparison with Gem or MMC (NCT06211764). The Alliance for Clinical Trials in Oncology has initiated recruitment for an RCT comparing BCG re-induction combined with Gem to BCG monotherapy for BCG-exposed NMIBC (GAIN-BCG, NCT07000084), with trial completion anticipated in 2028.Paclitaxel Multiple nanoparticle formulations of paclitaxel are undergoing evaluation for intravesical administration. Due to its inherent poor water solubility, paclitaxel requires specific modifications to achieve dissolution. Nab-paclitaxel (nab-PTX), a prominent example of such modification, improves solubility through its association with albumin. Nab-PTX has been investigated in patients with BCG-unresponsive NMIBC, reporting an initial CR rate of 36%. 31 Currently, ONCOFID-PB, a novel complex formed by paclitaxel and hyaluronic acid, is being assessed in an international phase 3 single-arm study for BCG-unresponsive CIS (NCT05024773). Furthermore, a phase 1 study is in progress for a formulation that incorporates paclitaxel within PLZ4-coated micelles.Docetaxel Intravesical docetaxel (Doc) therapy for BCG-refractory NMIBC was first evaluated in a phase 1 clinical trial in 2010, demonstrating promising outcomes with a 1 year RFS rate of 61%. 32 Although retrospective studies have reported the efficacy of sequential combination therapy with Gem and Doc for high-risk NMIBC after BCG failure,33 34 no prospective studies are currently being conducted on this combination. In this context, NDV01, a sustained-release formulation containing both Gem and Doc, is undergoing a phase 2 clinical trial (NCT06663137). In parallel, efforts to improve Doc formulations have led to the development of Nanoxel, a polymeric micelle-based formulation designed to eliminate the use of polysorbate 80 surfactants that can induce fatal anaphylaxis;35 however, its planned phase 3 clinical trial was terminated (NCT02982395). Similarly, BH011, a modified Doc formulation that also omits polysorbate 80 to reduce adverse effects, is currently being evaluated in phase 2 clinical trials (NCT06732531).Mitomycin C Standard intravesical MMC therapy is considered insufficient for treating BCG-unresponsive NMIBC, 27 prompting investigations into enhanced MMC administration methods such as hyperthermic MMC and electromotive drug administration (EMDA)-MMC. An RCT comparing hyperthermic chemotherapy using the recirculating intravesical thermotherapy (RITE) system against second-course BCG therapy or conventional chemotherapy showed no significant differences in the CR rate or RFS rate.36 A single-arm phase 2 trial evaluating EMDA-MMC for BCG-refractory NMIBC reported a 1-year RFS rate of 50%.37Nitroxoline Nitroxoline (APL-1202) is an oral methionine aminopeptidase 2 (MetAP2) inhibitor and the only orally administered anticancer agent currently under clinical investigation for BCG-unresponsive NMIBC. A clinical phase 2 trial demonstrated safety profiles of nitroxoline and an RFS rate of 39% in patients with high-risk NMIBC after BCG or intravesical chemotherapy failure ( NCT04498702).38 In a clinical phase 3 trial of nitroxoline in combination with doxorubicin versus doxorubicin alone, which is ongoing in China, the study subjects did not include patients with high-risk NMIBC after BCG failure but those with NMIBC after chemotherapy failure (NCT04490993).Immunotherapy and biological response modifiers Nadofaragene firadenovec Nadofaragene firadenovec (Adstiladrin, NCT02773849) is a non-replicating adenoviral vector encoding the interferon alfa-2b (IFNα−2b) gene, administered intravesically once every 3 months, approved by the FDA in 2022 for the treatment of BCG-unresponsive CIS. In a phase 3 single-arm trial (ADSTILADRIN), the CR rate was 53.4% and CR had been maintained in 24% of patients at 12 months.39 In a long-term follow-up analysis, the 5 year RC-free survival rate was 49% overall and 43% in the CIS subgroup, indicating that approximately half of the patients were able to avoid cystectomy.40 An RCT is currently underway comparing nadofaragene firadenovec with or without Gem/Doc or pembrolizumab for BCG-unresponsive NMIBC (ABLE-22, NCT06545955). Additionally, recruitment has begun for an RCT comparing nadofaragene with Gem/Doc for BCG-failure NMIBC, using less stringent criteria than BCG-unresponsive disease (COMPARE IT Trial, NCT06929286).NAI Nogapendekin alfa inbakicept (NAI) is an IL-15 superagonist complex that has demonstrated promising results in combination with BCG in the QUILT 3.032 trial. 41 In patients with BCG-unresponsive CIS, a CR rate of 71% was achieved, with a median duration of CR of 26.6 months among responders. Based on these results, NAI was approved by the FDA in 2024.Cretostimogene grenadenorepvec Cretostimogene grenadenorepvec (CG0070) is an oncolytic virus vaccine engineered from an adenovirus vector carrying the granulocyte–macrophage colony-stimulating factor ( GM-CSF) gene, designed to selectively replicate within tumour cells. A phase 3 trial of CG0070 monotherapy for BCG-unresponsive high-risk NMIBC (BOND-003, NCT04452591) is currently underway. According to its interim analysis, a CR rate was 74.5% with a median duration of response of 27 months or longer (not reached).42 Final results are awaited. A phase 2 trial of CG0070 in combination with pembrolizumab (CORE-001, NCT04387461) demonstrated a CR rate of 57.1% at 12 months in patients with BCG unresponsive CIS.43 A phase 2 multicohort study for BCG-exposed NMIBC has also been initiated, investigating combination therapies including gemcitabine (NCT06567743).Detalimogene voraplasmid Detalimogene voraplasmid (EG-70) is a non-viral genetic medicine-based immunotherapy. On cellular uptake, it delivers a plasmid designed to simultaneously express interleukin 12 (IL-12) and regulators of retinoic acid-inducible gene I. This dual action stimulates both innate and adaptive immunity, ultimately leading to its antitumour effects. Currently, a phase 2 trial is underway for patients with BCG-unresponsive CIS (LEGEND trial, NCT04752722). Part of the preliminary outcomes of this trial was reported at the ASCO-GU 2025 annual meeting, demonstrating an impressive overall CR rate of 71% in 21 efficacy-evaluable patients.44Interferon α Interferon α has long been studied as an agent for intravesical therapy in bladder cancer. In a subgroup analysis of a clinical trial originally designed for a different purpose, the combination of BCG plus interferon α in patients with CIS who had received two or more prior courses of BCG demonstrated a 24-month RFS rate of 23%. 45IAP0971 IAP0971 is a novel heterodimeric fusion protein combining an anti-PD-1 antibody with the IL-15/IL-15Rα complex. It is designed to enhance immune activation while limiting nonspecific responses by targeting PD-1. A phase 1/2 trial ( NCT06255964) launched in China in 2024 will evaluate intravesical administration, either alone or in combination with BCG.Catumaxomab A phase 1/2 clinical trial of catumaxomab ( NCT04799847), a bispecific antibody targeting both epithelial cell adhesion molecule (EpCAM) and CD3, was initiated in China in 2021. One arm of the antibody binds to EpCAM on the surface of tumour cells, while the other engages CD3 on T cells, aiming to mediate tumour cell lysis and promote subsequent immune activation. Results have not yet been reported.CADI-05 CADI-05 is a Toll-like receptor 2 agonist given via an intradermal injection. In a study involving 20 patients with NMIBC after BCG failure, CADI-05 achieved an RFS rate of 35% at both 15 and 30 months. 46Mycobacterium phlei cell wall-nucleic acid complexMycobacterium phlei cell wall-nucleic acid complex (MCNA) is a cell wall extract of Mycobacterium phlei that exerts antitumour activity through two mechanisms: cytokine induction and direct apoptosis of cancer cells. A single-arm trial demonstrated RFS rates of 25% at 1 year and 19% at 2 years.47 Following the formal definition of types of BCG failure, a post hoc analysis showed RFS rates of 48.9% at 6 months, 34.8% at 1 year and 28.3% at 2 years in patients with BCG-unresponsive NMIBC.48 An RCT comparing MCNA with MMC was initiated but terminated due to insufficient patient enrolment (NCT01200992).VPM1002BC VPM1002BC is a genetically modified BCG initially designed to augment the protective effects of the traditional BCG tuberculosis vaccine. This recombinant BCG has been endowed with the ability to secrete listeriolysin from Listeria monocytogenes, a modification anticipated to enhance CD8+ T-cell stimulation.49 Clinical evaluation in a phase 1/2 trial for patients with BCG-unresponsive NMIBC has demonstrated favourable outcomes, achieving a 60-week RFS rate of 49.3%. Notably, adverse events of grade 3 or higher were infrequent, with urinary tract infection being the sole reported event in 5% of participants.50OncoTherad OncoTherad, a nanometre-scale immunomodulator originating from Brazil, is expected to elicit an immune response via mechanisms including the Toll-like receptor pathway. Clinical investigation in BCG-unresponsive NMIBC demonstrated promising results: no patient experienced recurrence within 6 months, and 62.5% of patients maintained RFS within a 2-year period. 51Herpes simplex virus Several oncolytic herpesviruses are being investigated for intravesical instillation. These include OH2, Talimogene laherparepvec (T-VEC) and T3011, all of which carry the GM-CSF gene, with T3011 additionally engineered with an IL-12 gene.52 All of these oncolytic herpesvirus formulations are administered via intravesical instillation, where the virus locally invades tumour cells, inserts its genetic material and induces an immune response that exerts antitumour effects. While T-VEC was approved for other cancers in 2015, its development for NMIBC was withdrawn. Clinical trials for OH2 for high-risk NMIBC (NCT05232136) and T3011 for BCG-unresponsive or BCG-exposed NMIBC (NCT06971614) are currently ongoing.ICIs Pembrolizumab Pembrolizumab, an anti-PD-1 antibody, is the only ICI approved by the FDA for the treatment of NMIBC. In the phase 2 single-arm trial (KEYNOTE-057, NCT02625961), a CR rate of 41% at 3 months was reported in patients with BCG unresponsive CIS.53 A phase 3 RCT (KEYNOTE-676, NCT03711032) comparing pembrolizumab plus BCG versus BCG alone is currently ongoing for both BCG-unresponsive and BCG-exposed NMIBC. Additionally, a phase 1 trial evaluating ‘intravesical’ pembrolizumab in combination with BCG reported RFS rates of 67% and only 22% at 6 and 12 months, respectively (NCT02808143).54 Furthermore, a phase 2 single-arm study combining pembrolizumab with Gem/Doc is scheduled in Argentina (NCT06972615).Durvalumab Durvalumab, an anti-PD-L1 antibody, is the most frequently investigated ICI in clinical trials for the BCG-unresponsive NMIBC. Although the phase 3 trial had not started at this time, several phase 2 trials including durvalumab monotherapy administered intravenously or intravesically and combination therapies of durvalumab with either tremelimumab (an anti-CTLA-4 antibody), monalizumab (an anti-NKG2A antibody) or montanide (a water-in-oil emulsion adjuvant to enhance immune responses) have been conducted for NMIBC after BCG failure. Part of the results of the ADAPT-BLADDER trial ( NCT03317158) demonstrated that durvalumab in combination with Gem/Doc showed a CR rate of 89% at any time point in 37 patients with BCG-unresponsive NMIBC and that the CR rate remained 69% at 12 months.55Nivolumab CheckMate 7G8 ( NCT04149574) was a phase 3 RCT comparing nivolumab in combination with BCG versus BCG monotherapy. This trial was terminated early due to insufficient patient enrolment.TKIs FGFR3 inhibitors Multiple clinical trials of FGFR3 inhibitors have been conducted in the field of NMIBC with FGFR3 alterations. Erdafitinib has been evaluated in BCG-unresponsive CIS in the THOR-2 study cohort 2, demonstrating a CR rate of 100% at week 9 and 75% at week 25 in the interim analysis.56 TAR-210, an intravesical sustained-release formulation of erdafitinib, has shown promising results in the interim analysis; a 12-month RFS rate was 90% in 21 patients with BCG-unresponsive or BCG-exposed high-risk NMIBC patients.57 An international phase 3 RCT is planned to begin, comparing TAR-210 with Gem or MMC for BCG-exposed or BCG-unresponsive papillary NMIBC (MoonRISe-3, NCT06919965). Other clinical trials involving infigratinib, rogaratinib and dovitinib have also been planned; however, the development of rogaratinib was discontinued, and dovitinib showed poor response rates in BCG-unresponsive NMIBC with FGFR3 mutations or overexpression.58Sunitinib Sunitinib, a multitarget TKI, had been the standard of care as the first-line systemic therapy for metastatic clear cell renal cancer before the advent of ICI combination treatments. 59 A phase 2 trial failed to demonstrate a clinical benefit in patients with BCG-refractory high-risk NMIBC.60ADCs Enfortumab vedotin and 9MW2821 Intravesical administration of enfortumab vedotin, an ADC comprising an anti-Nectin-4 antibody and monomethyl auristatin E (MMAE), has been investigated (EV-104, NCT05014139). An interim analysis of a dose-escalation study reported a CR rate of 75% at the 125 mg dose level among four patients with BCG-unresponsive CIS.61 A next-generation ADC, 9MW2821, is composed of an anti-Nectin-4 antibody and MMAE, using a site-specific conjugation method and a novel linker to enhance stability and reduce off-target toxicity compared with enfortumab vedotin.62 A phase 1 study of intravesical administration of 9MW2821 has been initiated for high-risk NMIBC after failure of intravesical therapies including BCG (NCT06551233).Disitamab vedotin Clinical trials of disitamab vedotin, an ADC composed of an anti-HER2 antibody and MMAE, were initiated for HER2-positive NMIBC in 2023. These included intravenous monotherapy ( NCT05996952) and the combination with intravesical Gem (NCT05943379) for high-risk NMIBC, including that after BCG failure, and combination with tislelizumab (an anti-PD-1 antibody) for high-risk NMIBC after BCG failure (NCT05957757).Cytotoxic agent other than anticancer drugs Vicineum Vicineum (VB4-845) is a fusion protein composed of an antibody targeting EpCAM linked to Pseudomonas exotoxin A (ETA). EpCAM is a membrane protein highly expressed in many cancers, and binding of Vicineum facilitates intracellular delivery of ETA, which is expected to induce tumour cell death. A phase 3 trial (VISTA trial, NCT02449239) demonstrated a CR rate of approximately 40% at 3 months and 21% at 12 months in patients with BCG-unresponsive CIS.63 This trial was rated as poor quality according to the MINORS criteria due to the short observation period and the high rate of participant attrition (online supplemental table 4). This agent failed to gain FDA approval64 and its development has been paused.65Radiation Several clinical trials of radiotherapy combined with radiosensitisers have been conducted for high-risk NMIBC after BCG failure. In the RTOG 0926 study, radiotherapy combined with cisplatin or 5-FU plus MMC was given to 34 patients with T1 NMIBC that had failed BCG as an alternative to RC ( NCT00981656), demonstrating a 3-year bladder preservation rate of 88.2%.66 Other ongoing trials include the PREVERT trial evaluating radiotherapy with avelumab in BCG-unresponsive high-risk NMIBC (NCT03950362) and the PARRC trial comparing radiotherapy with pembrolizumab versus chemotherapy (cisplatin, Gem or MMC) in high-grade T1 diseases including those after BCG failure (NCT06770582).Photodynamic therapy A phase 2 trial of photodynamic therapy using the photosensitiser TLD-1433 is currently underway for BCG-unresponsive CIS with or without papillary disease ( NCT03945162). Interim analysis of 45 evaluable patients revealed a 3-month CR rate of 50% and a 1-year CR rate of 35% without severe adverse events directly related to photodynamic therapy.67 General anaesthesia is required during the light irradiation procedure.Discussion RC and urinary diversion are currently the standard of care for patients with non-metastatic MIBC in combination with neoadjuvant and/or adjuvant systemic therapy and for those with BCG-unresponsive high-risk NMIBC. Despite improved surgical outcomes by robotic surgical platforms, RC is still associated with a complication rate of~20% and a perioperative mortality rate of 1%–3%. 68 69 In addition, RC can compromise post-surgical QoL in terms of urinary, sexual and gastrointestinal functions as well as changes in body image in cases of non-continent urinary diversion.70 Given the drawbacks of RC, many patients appear to prefer BST options, if present, to RC. In fact, an RCT comparing RC and BST in patients with MIBC was launched but was closed 3 years later because of poor accrual; many patients preferred the BST arm and declined randomisation.71 Although the development of BST options had long been an unmet need for patients with BCG-unresponsive high-risk NMIBC, no effective therapeutics were available until recently.Looking back at its history, the development of novel BST therapies for NMIBC after BCG failure started with intravesical immunotherapy, biological response modifiers and conventional anticancer agents. Since the late 2010s, the mainstay has shifted to innovative drug delivery of anticancer agents, gene therapy-based biological response modifiers, TKIs and ICIs. More recently, clinical trials of ADCs have been launched and their outcomes are awaited. As of June 2025, the FDA has approved valrubicin, pembrolizumab, nadofaragene firadenovec and NAI for BCG-unresponsive NMIBC. As approval of these agents has been primarily based on single-arm trials targeting patients with BCG-unresponsive CIS with or without concurrent papillary tumours, the FDA has not granted specific approval for their use in those with papillary NMIBC without CIS after BCG failure. However, the National Comprehensive Cancer Network (NCCN) lists pembrolizumab and nadofaragene firadenovec as Category 2B recommendations for this subgroup.4 As such, recommendations by clinical guidelines do not always agree with regulatory approvals. The NCCN probably incorporates the results of subgroup analysis and expert consensus to support practical off-label use,72 which is relatively common in clinical practice in the United States.Several novel agents are expected to be sent for approval by the FDA in the near future, including cretostimogene grenadenorepvec (either alone or in combination with pembrolizumab), detalimogene voraplasmid and durvalumab (in combination with Gem/Doc) because interim or final analyses have demonstrated promising early CR rates comparable with or higher than those of FDA-approved pembrolizumab, nadofaragene firadenovec and NAI (online supplemental figure 3). Expanding therapeutic options would further increase the chance of bladder preservation in patients with high-risk NMIBC after BCG failure and may eventually improve their oncological outcomes.SP310.1136/bmjonc-2025-000753.supp3Supplementary dataReal-world data have indicated that approximately 25% of patients do not complete BCG induction therapy mainly because of toxicities.73 74 Reportedly, up to 97% of patients experienced voiding pain, urinary frequency, haematuria, fever, general malaise and/or rare but severe complications such as contracted bladder and systemic BCG infection.75 76 A condition in which patients cannot complete the BCG regimen due to adverse events is referred to as BCG intolerance. Importantly, BCG-intolerant patients with NMIBC are underserved as current treatment options are essentially limited to RC or conventional intravesical chemotherapy worldwide. There is a clear unmet need for therapeutic development in this subgroup.Recently, the broader concept of BCG-exposed disease has been proposed, referring to patients who have received prior BCG therapy but do not meet the strict criteria for BCG-unresponsive disease.77 This category encompasses patients with BCG-resistant disease or late relapse after adequate or inadequate BCG therapy. The International Bladder Cancer Group recommends that clinical trials targeting BCG-exposed NMIBC should be conducted as RCTs, as BCG rechallenge remains a reasonable therapeutic option in this population and should serve as the control arm, unlike BCG-unresponsive NMIBC, for which BCG is no longer considered effective. Clinical trials such as KEYNOTE-676 (pembrolizumab and BCG vs BCG alone), THOR-2 (erdafitinib vs intravesical chemotherapy), MoonRISe-3 (TAR-210 vs Gem or MMC) and GAIN-BCG (BCG vs BCG and Gem) have adopted this approach, targeting a broader BCG-exposed population.Overall, the acceptance of single-arm trial designs by the FDA appears to facilitate drug development in the field of NMIBC after BCG failure. In fact, after a long period of time since the approval of valrubicin in 1998, three additional agents (pembrolizumab, nadofaragene firadenovec, NAI and TAR-200) have been approved based on single-arm data to expand treatment options for BCG-unresponsive NMIBC. In the risk-of-bias assessment for each study, it was observed that most recent studies maintained a quality of moderate or higher. This can be considered a product of the standardisation of research methodologies. Nevertheless, in its 2024 draft guidance, the FDA emphasised that RCTs are preferable especially for patients with resected BCG-unresponsive papillary NMIBC.9 Currently, single-arm trials still remain dominant in ongoing clinical research. Future trends in trial design and regulatory policy are crucial to drug development as the therapeutic landscape continues to evolve in this field.Limitations of this systematic review include the predominance of early-phase trials, heterogeneity in outcome measurements and a limited number of comparative studies. It is worth noting, however, that the quality of the included single-arm trials was generally reliable, as evidenced by 82% of them receiving a MINORS score of moderate or higher. Despite this, the current evidence provides only limited guidance for decision-making regarding the optimal treatment in clinical practice.Conclusions To date, a wide range of clinical trials on therapeutic approaches have been investigated, including cytotoxic agents, immunotherapy, ICIs, ADCs, TKIs and radiotherapy for NMIBC after BCG failure. Notably, ICIs and some immunotherapies have already received regulatory approval, making meaningful progress after decades of limited therapeutic options. Despite these advances, significant unmet needs persist. Therapeutic development for BCG-intolerant and broader BCG-exposed populations is still in its early stages. Furthermore, the predominance of single-arm studies, along with heterogeneous endpoints and populations, limits the ability to draw definitive conclusions regarding optimal management. Therefore, future research should prioritise rigorously designed comparative studies and outcome measures. Continued progress in trial standardisation, in parallel with emerging therapeutic platforms, is expected to further expand bladder-preserving options and ultimately improve oncologic and quality-of-life outcomes for patients with high-risk NMIBC after BCG failure.