Inferring optimal MTAP homozygous loss (homozygous deletion) detection in urothelial cancer: Prevalence and co-alterations based upon treatment exposure and anatomical location.

641 Background: Validated homozygous loss detection is becoming increasingly important in clinical practice, with prostate NCCN guideline inclusion of BRCA1/2 loss (portending durable benefit from PARP inhibitor) and multi-tumor trials of PRMT5 and MAT2A inhibitors ongoing (biomarker: MTAP deletion). Homozygous losses are challenging to detect and require intentional NGS assay design and validations. Using a well-validated test, we sought to evaluate the most common losses in metastatic urothelial cancer tissue samples, evaluate their prognostic associations on standard of care regimens, prevalence before and after standard of care regimens, and overlap with other biomarkers. Methods: This study used the nationwide (US-based) de-identified Flatiron Health-Foundation Medicine Urothelial Cancer clinico-genomic database (FH-FMI CGDB), originating from approximately 280 US cancer clinics (~800 sites of care). Patients with metastatic urothelial cancer and tissue tumor genomic testing by FoundationOne CDx were eligible for analysis. Prevalence of alterations was compared between non-sequential groups of specimens obtained from patients who were naïve or exposed to immune checkpoint inhibitors (ICI) or chemo. Outcome analyses were adjusted for: ECOG performance score, age, and socioeconomic status. Results: 1446 specimens from 1446 unique patients met criteria for assessment. Only three genes featured homozygous losses with a prevalence greater than 3%: CDKN2A (36.5%), CDKN2B (34.6%), and MTAP (28.8%). Prevalence of each of these alterations was similar in specimens from patients who were naïve to ICI or chemo compared to those exposed to ICI or chemo. Multivariable models did not suggest associations with outcome on ICI, chemo, or enfortumab vedotin. In 401 cases (27.7%) MTAP , CDKN2A and CDKN2B genes were co-lost, with 97 (6.7%) of cases where CDKN2A and CDKN2B were lost but not MTAP . MTAP loss without CDKN2A/B loss was very rare (n = 15, 1.04%). In multivariable models adjusting for ICI and chemo exposure, upper tract urothelial carcinoma had significantly higher prevalence of MTAP loss (OR: 2.53, p < 0.0001) with a prevalence of 48.3%. RB1 mutations were strongly inversely associated with MTAP loss (OR: 0.037, p < 0.0001) and FGFR3 mutations were strongly associated with MTAP loss (OR: 2.46, p < 0.0001). Conclusions: MTAP loss prevalence in urothelial cancer is about 29% via an assay and algorithm designed to detect homozygous losses, with higher prevalence in upper tract urothelial carcinoma. Our results are consistent with a model of MTAP loss in bladder cancer as a bystander effect of CDKN2A/B loss, and that it is a truncal event, inferring that both archival or contemporaneous tissue specimens are valid for assessment of MTAP loss for patient enrollment into clinical trials for PRMT5 and MAT2A inhibitors.