Introduction:

Over seventy percent of bladder cancer presents as non-muscle invasive bladder cancer (NMIBC). The standard of care for high-grade NMIBC is intravesical Bacillus Calmette-Guerin (BCG) immunotherapy. Unfortunately, forty percent of patients do not respond to BCG therapy and experience recurrence or progression of disease. Identifying genomic alterations in high-grade bladder cancer associated with clinical response to BCG immunotherapy can provide clinical utility for both prognostic purposes and for identifying novel targeted therapy. Our group has previously identified an association between ARID1A mutations and an increased risk of recurrence after BCG therapy. By utilizing next-generation sequencing to evaluate pretreatment and posttreatment specimens from BCG therapy within the same patient, we can identify resistance patterns within tumors.

Methods:

We identified high-grade NMIBC patients who underwent BCG immunotherapy who were enrolled on a prospective IRB-approved protocol for which targeted exon capture sequencing (MSK-IMPACT) was performed on pretreatment and posttreatment tumor DNA and matched germline DNA in a CLIA-certified laboratory. The primary endpoint of analysis was high-grade recurrence following intravesical BCG therapy. Patients were classified as having BCG refractory, BCG relapsing, and/or BCG unresponsive disease. BCG refractory defined as failure to be disease free within 6 months after initial BCG with maintenance or re-treatment at 3 months due to rapid recurrence, BCG relapsing is tumor recurrence after 6 months of BCG with early (<12 months) or late (>12 months) and BCG unresponsive as recurrent CIS (within 12 months) or Ta/T1 (within 6 months) after completion of adequate BCG. Genomic alterations and tumor mutational burden (TMB) were correlated with BCG refractory versus BCG relapsing disease.

Results:

Thirteen patients with high-grade recurrence following BCG treatment were identified who had their pretreatment and posttreatment specimens undergo genomic analysis. The median age at diagnosis was 57 years (IQR 48-76) with ten patients (77%) male. Nine patients (69%) were cTaHG, 3 patients (23%) cT1 and 1 patient (8%) cTis with median follow-up of 57 months (IQR 42-67). As demonstrated in Figure 1, the concordance between pretreatment BCG and posttreatment BCG specimens was very high with only select cases of unique mutations between matched specimens. One select case in the late BCG relapsing cohort (Patient 13) with a known FGFR3-TACC3 fusion had a pretreatment CDKN1A mutation and after BCG treatment had two unique secondary CDKN1A mutations. Another case in the late BCG relapsing cohort (Patient 10) was found to have both ARID1A, PIK3CA and FGFR3 hotspot mutations. As a tumor harboring an ARID1A mutation there is an increased risk of recurrence and with two actionable mutations in PIK3CA and FGFR3 this may have implications for possible targeted therapy in the future.

When comparing genomic alterations between BCG refractory/relapsing cohorts only TP53 (HR 4.5 (1.1-18.3), p=0.03) and KMT2D (HR 4.4 (1.4-13.8), p=0.01) were significantly associated with BCG refractory disease. The numbers are very small but comparing genomic alterations between BCG unresponsive to remainder of cohort identified CDKN2A and BRCA2 to be significantly associated. There was no significant difference in the TMB between BCG refractory and BCG relapsing (13.5 vs 12.1, p=0.8) nor between pre-treatment versus post-treatment specimens (12.5 vs 13.0, p=0.9).

Conclusion:

BCG recurrence can be prevalent in NMIBC and with recurrent BCG shortages further knowledge of resistance patterns can ultimately help guide clinical practice. On the whole we found the concordance to be exceptionally high between pretreatment and posttreatment specimens. The cohort identified here is small but there are some select differences between pretreatment and posttreatment specimens that may warrant more evaluation. While we are looking at only patients with BCG recurrences, there are a number of actionable mutations that serve as a potential for novel targeted therapies. These recurrent tumors after BCG therapy are being explored further with whole-exome sequencing and other next-generation sequencing to analyze expression across the transcriptome. The ability to sequence tumors within the same patient at different temporal points after treatments allows for further elucidation of clonal evolution of tumors.

Funding: This work was supported by the Sidney Kimmel Center for Prostate and Urologic Cancers, the Michael and Zena Wiener for Therapeutics Program in Bladder Cancer, Pin Down Bladder Cancer, Cycle for Survival, the Marie-Josee and Henry R. Kravis Center for Molecular Oncology, NIH/NCATS Grant Number UL1-TR002384, the National Cancer Institute Cancer Center Core Grant Number P30-CA008748 and by SPORE in Bladder Cancer P50-CA221745.