High-risk, bacillus Calmette–Guérin-unresponsive non-muscle-invasive bladder cancer remains a major therapeutic challenge, particularly for patients seeking an alternative to radical cystectomy. Results from Cohort C of the phase 3 BOND-003 trial showed that intravesical cretostimogene grenadenorepvec produced a complete response at any time in 75% of evaluable patients with BCG-unresponsive non-muscle-invasive bladder cancer and carcinoma in situ.
At 24 months, 42% of the full efficacy population were in complete response. Among patients who achieved a complete response at 3 or 6 months, the estimated probability of remaining disease-free at 24 months was 60.1%. No grade 3 or 4 treatment-related adverse events were reported.
The article, titled “Intravesical cretostimogene grenadenorepvec oncolytic immunotherapy in high-risk, BCG-unresponsive, non-muscle-invasive bladder cancer with carcinoma in situ (BOND-003 Cohort C): a single-arm, phase 3 trial,” was published in the August 2026 issue of The Lancet Oncology.
Authors: Mark D Tyson II, Jong-Kil Nam, Shreyas S Joshi, Edward M Uchio, Seung Il Jung, Trinity J Bivalacqua, Gary D Steinberg, Neal D Shore, James M Burke, Hiroshi Kitamura, Ben Tran, and Roger Li.
The Need for Bladder-Sparing Treatment
Patients with high-risk non-muscle-invasive bladder cancer who develop persistent or recurrent disease after adequate BCG therapy have limited bladder-preserving treatment options.
Radical cystectomy remains the guideline-recommended treatment in this setting. Although surgery can provide effective disease control, it is a life-altering intervention associated with substantial morbidity and complications. The need therefore remains for treatments capable of producing durable tumor control while allowing patients to preserve their bladder.
Cretostimogene, previously known as CG0070, is an intravesically administered oncolytic immunotherapy with two complementary mechanisms of action. The therapy is designed to selectively replicate in and lyse cancer cells with alterations in the retinoblastoma–E2F pathway. It also expresses granulocyte–macrophage colony-stimulating factor, which is intended to recruit and activate antigen-presenting cells and enhance antitumor immunity.
BOND-003 Cohort C Study Design
BOND-003 Cohort C was an international, single-arm, phase 3 study conducted at 41 academic and community centers across North America, Asia, and Australia.
Eligible patients were aged 18 years or older, had an Eastern Cooperative Oncology Group performance status of 0–2, and had pathologically confirmed high-risk, BCG-unresponsive non-muscle-invasive bladder cancer with carcinoma in situ. Patients could also have concomitant resected high-grade Ta or T1 disease.
BCG-unresponsive disease was defined as persistent or recurrent carcinoma in situ, with or without concomitant high-grade Ta or T1 disease, within 12 months after the last dose of adequate BCG treatment. Patients received intravesical cretostimogene at a dose of 1 × 10¹² viral particles in 0.8 mL once weekly for 6 weeks. Patients with persistent carcinoma in situ, high-grade Ta disease, or carcinoma in situ with high-grade Ta disease at the 3-month assessment could receive a second 6-week induction course.
Patients who achieved a complete response subsequently received maintenance treatment. This consisted of once-weekly administration for 3 weeks every 3 months during the first year, followed by once-weekly administration for 3 weeks every 6 months during the second and third years.
The primary endpoint was centrally confirmed complete response at any time, which included complete response at 3 months or, among patients who underwent re-induction, at 6 months.
Secondary endpoints included complete response at 12 months, duration of response, progression-free survival, progression to a higher disease stage, cystectomy-free survival, overall survival, and safety.
Cystectomy-free survival was defined as the time from treatment initiation until radical cystectomy or death. The primary statistical hypothesis used a historical complete response benchmark of 20%, based on results with intravesical valrubicin and discussions with the US FDA. The trial is registered with ClinicalTrials.gov under the identifier NCT04452591 and remains ongoing.
Patient Population
Between October 9, 2020, and August 3, 2023, 165 patients were assessed for eligibility and 115 were enrolled.
A total of 112 patients received at least one dose of cretostimogene and were included in the safety analysis. Two treated patients withdrew because of unrelated adverse events before the 3-month efficacy assessment, leaving 110 patients in the efficacy analysis.
The median age was 74 years, 74% of patients were male, and 85% had an ECOG performance status of 0. At study entry, 80% had carcinoma in situ alone, 15% had carcinoma in situ with high-grade Ta disease, and 4% had carcinoma in situ with T1 disease.
Patients had received a median of 12 previous BCG instillations. Nearly half, 47%, had also received at least one non-BCG treatment directed at non-muscle-invasive bladder cancer.
Complete Response at Any Time Reached 75%
At the June 23, 2025, data cutoff, the median follow-up was 25.8 months. Complete response at any time was observed in 83 of 110 evaluable patients, corresponding to a complete response rate of 75%. The 95% confidence interval was 66.3%–83.2%.
The result was significantly higher than the prespecified historical benchmark of 20%, with a p value of less than 0.0001. Fourteen patients achieved a complete response following re-induction.
Among the 27 patients who did not achieve a complete response, 21 had persistent disease, 4 experienced progression within the non-muscle-invasive setting, 1 progressed to muscle-invasive disease, and 1 had recurrent disease. Complete response results were generally consistent across the prespecified patient subgroups.
Responses at 12 and 24 Months
At 12 months, a complete response was observed in 51 patients, or 46% of the efficacy population. The corresponding Kaplan–Meier estimate was 50.7%.
At 24 months, a complete response was observed in 46 patients, or 42% of the efficacy population. The corresponding Kaplan–Meier estimate was 42.4%. The assessment of complete response at 24 months was conducted as a post-hoc analysis.
These findings should be distinguished from the duration-of-response analysis, which included only patients who achieved a complete response at 3 or 6 months.
Among these responders, the estimated probability of remaining disease-free was 64.2% at 12 months and 60.1% at 24 months. The estimated median duration of response was 27.9 months. At the data cutoff, one patient had remained disease-free for more than 51 months after completing maintenance treatment. Among the 14 patients who achieved a complete response following re-induction, 9 maintained a response lasting at least 15 months.
Cystectomy-Free Survival and Disease Progression
The estimated cystectomy-free survival rate was 89% at 12 months and 81% at 24 months. Median cystectomy-free survival had not been reached.
Because cystectomy-free survival was defined as the time to radical cystectomy or death, these findings represent a composite time-to-event endpoint rather than the proportion of patients who avoided cystectomy alone.
Eighteen patients underwent radical cystectomy after disease recurrence or progression. Of these, 15, or 83%, had either non-muscle-invasive bladder cancer or no residual tumour, classified as pT0, on final pathological examination.
The median time between recurrence or progression and cystectomy was 20 weeks.
Progression to a higher disease stage occurred in 13 of 110 patients. Nine experienced progression within the non-muscle-invasive setting, from carcinoma in situ or carcinoma in situ with Ta disease to T1 disease, while 4 developed muscle-invasive T2 disease during study treatment.
Five of 112 treated patients died during the trial. All deaths occurred at least 7 months after treatment initiation and were considered unrelated to cretostimogene. Median overall survival had not been reached.
Safety
Treatment-emergent adverse events were reported in 106 of 112 patients, or 95%. A total of 71 patients, or 63%, experienced at least one treatment-related adverse event. Treatment-related events were predominantly grade 1 or 2 local urinary symptoms. The most frequently reported events were bladder spasm in 25% of patients, pollakiuria in 22%, micturition urgency in 21%, dysuria in 19%, and hematuria in 13%.
The median time to resolution of treatment-related adverse events was 1 day. No grade 3 or 4 treatment-related adverse events were reported. No treatment-related adverse events resulted in treatment discontinuation or death.
Four patients discontinued treatment because of treatment-emergent adverse events, but none of these events were considered related to cretostimogene. Serious adverse events were reported in 15 patients. Two patients experienced serious treatment-related adverse events: one case of non-infective cystitis and one case of urinary bladder hemorrhage. Both events were grade 2.
Interpretation and Study Limitations
The BOND-003 findings indicate that intravesical cretostimogene can produce clinically meaningful and potentially durable responses in patients with high-risk, BCG-unresponsive non-muscle-invasive bladder cancer with carcinoma in situ. The adverse-event profile was predominantly characterized by low-grade, transient urinary symptoms.
However, the study did not include a control group. The findings can therefore be interpreted only in relation to historical benchmarks and cannot establish the relative efficacy or safety of cretostimogene compared with other available bladder-sparing therapies.
The investigators cautioned that relative clinical benefits among available treatments cannot be established without head-to-head trial comparisons. Additional limitations included potential variation in cystoscopy and transurethral resection techniques across participating centers. The amount of carcinoma in situ present before treatment was also not measured as a potential confounding factor.
Longer-term follow-up is required to better define the durability of response. Biomarkers capable of predicting which patients are most likely to benefit have not yet been established. Health-related quality-of-life and cost-effectiveness outcomes were not included in the current report, although dedicated analyses are planned.
Conclusion
Cohort C of the phase 3 BOND-003 trial met its primary endpoint, with a complete response at any time observed in 75% of evaluable patients with high-risk, BCG-unresponsive non-muscle-invasive bladder cancer and carcinoma in situ.
At 24 months, 42% of the full efficacy population were in complete response. Among patients who responded at 3 or 6 months, the estimated probability of remaining disease-free at 24 months was 60.1%. The estimated cystectomy-free survival rate was 81% at 24 months. Treatment-related adverse events were primarily low-grade urinary symptoms, with no grade 3 or 4 treatment-related events and no treatment-related discontinuations or deaths.
Although the single-arm design prevents direct comparison with other bladder-sparing treatments, the results support the continued evaluation of cretostimogene as an intravesical treatment for patients with high-risk, BCG-unresponsive non-muscle-invasive bladder cancer with carcinoma in situ.
The study was funded by CG Oncology. The funder contributed to the study design, data analysis, data interpretation, and manuscript development but had no role in data collection.
The full article is available in The Lancet Oncology.
Find more videos on OncoDaily YouTube TV.