Impact of American Urological Association risk category on outcomes of intravesical BCG vs gemcitabine/docetaxel in high-grade Ta non-muscle-invasive bladder cancer.

OBJECTIVES

To compare oncological outcomes between intravesical bacillus Calmette-Guérin (BCG) and gemcitabine/docetaxel (Gem/Doce) in patients with primary high-grade (HG) Ta non-muscle-invasive bladder cancer (NMIBC) and assess the prognostic utility of the American Urological Association (AUA) risk stratification system, which classifies these tumours as intermediate risk (IR) or high risk (HR).

PATIENTS AND

METHODS

Retrospective cohort of 128 treatment-naïve patients with HG Ta NMIBC treated from January 2011 to December 2024 was analysed. After resection, patients received intravesical therapy with either BCG (n = 65) or Gem/Doce (n = 63), with Gem/Doce on a 24-month maintenance schedule and BCG at 3, 9, and 15 months. Cox regression models evaluated the impact of patient, disease, and treatment characteristics on oncological outcomes.

RESULTS

The follow-up was longer in the BCG group (median [interquartile range, IQR] 81 [44-113] months) than in the Gem/Doce group (median [IQR] 37 [26-55] months). Recurrence-free survival (RFS) at 12 and 24 months was 80% and 76% for Gem/Doce, and 76% and 70% for BCG, respectively. HG-RFS at 12 and 24 months was 82% and 78% for Gem/Doce, and 79% and 75% for BCG, respectively. The 24-month progression-free survival was 100% for Gem/Doce and 91% for BCG. There were no significant associations between AUA risk classification (IR vs HR; hazard ratio 1.05, P = 0.89 for RFS; hazard ratio 1.12, P = 0.75 for HG-RFS) or treatment type (BCG vs Gem/Doce; hazard ratio 1.19, P = 0.60 for RFS; hazard ratio 1.02, P = 0.96 for HG-RFS) and recurrence outcomes. Adverse events were less frequent with Gem/Doce than BCG (40% vs 46%), with no Grade III events in the Gem/Doce group.

CONCLUSION

Treatment with Gem/Doce is a well-tolerated, effective alternative to BCG for HG Ta NMIBC. In this subset, AUA risk stratification (IR vs HR) was not significantly associated with recurrence outcomes, warranting prospective validation.