Mark Farha, MD, on a genomic score to predict early BCG failure in NMIBC

In the following interview, Mark Farha, MD, BSc, MBA, discusses the development of a normalized genomic score (NGS) designed to predict early BCG failure in patients with non–muscle invasive bladder cancer (NMIBC). Initial findings on the NGS were presented at the 2026 American Society of Clinical Oncology Genitourinary Cancers Symposium in San Francisco, California.¹

Farha is a PGY-4 urology resident at Weill Cornell Medicine in New York, New York.

Farha explained that although BCG remains the standard of care for intermediate- and high-risk disease, outcomes remain variable, with approximately half of patients experiencing recurrence and a substantial proportion progressing to muscle-invasive disease. Despite decades of use, there are no widely implemented biomarkers to guide treatment selection. Against a backdrop of expanding intravesical treatment options, the study sought to leverage genomic profiling to better identify which patients are most likely to benefit from BCG vs those who may require alternative approaches.

The investigators developed a next-generation sequencing–based genomic score using tumor samples from patients treated with transurethral resection and BCG. Patients were categorized into phenotypes—those with early recurrence within 6 months (n = 24) and those with durable responses exceeding 5 years (n = 40)—to identify genomic features associated with outcomes. These features were incorporated into machine learning models, with the highest-performing model used to generate a normalized genomic score ranging from 0 to 100. Patients were subsequently stratified into low-, intermediate-, and high-risk groups based on this score.

Clinically, the NGS demonstrated meaningful prognostic utility. Patients classified as NGS-low experienced significantly improved high-grade recurrence-free survival compared with higher-score groups, and this association remained independently predictive after adjusting for established clinical risk factors (HR, 0.32; 95% CI, 0.11 to 0.96; P = .04). Notably, the genomic score added substantial discriminatory power to traditional risk stratification, improving the area under the curve for relapse prediction by 22% (AUA Risk AUC: 0.545, Combined: 0.666).

Farha concluded, “We clearly demonstrate that the addition of genomic features to clinical variables improves relapse risk prediction.”

REFERENCE

1. Farha M, Lang J, Freydenlund N, et al. Development of a non-muscle invasive bladder cancer (NMIBC) genomic score (NGS) to predict early BCG failure. Presented at: 2026 American Society of Clinical Oncology Genitourinary Cancers Symposium. February 26-28, 2026. San Francisco, California. Abstract 850. https://www.asco.org/abstracts-presentations/257115