Real-world analysis of folate receptor alpha (FRα; FOLR1) expression in pan-tumor samples from over 6000 patients in the US.

Abstract

5568 Background: FRα is overexpressed in several cancers, including ovarian and endometrial. The FRα-targeted antibody-drug conjugate mirvetuximab soravtasine-gynx (MIRV) showed survival benefit vs chemotherapy in patients with platinum-resistant ovarian cancer (PROC) with high FRα expression (Moore K, et al. N Engl J Med . 2023;389(23):2162-2174). Greater understanding of FRα expression and distribution in real-world (RW) settings may help enable FRα testing implementation for biomarker-guided treatment strategies. Here, we present analyses of FRα expression in RW patient tumor samples referred for FRα testing by immunohistochemistry (IHC). Methods: Uniquepatient tumor samples across different malignancies (N=6695) were acquired from RW healthcare settings in the US over a 12-month period. FRα expression was assessed by IHC using the VENTANA FOLR1 (FOLR1-2.1) RxDx Assay. Trained pathologists scored samples with ≥75% tumor cells with ≥2+ membrane staining as FRα positive. The clinicopathological features of each case, FRα prevalence, and impact of sample age were evaluated. Results: From the entire dataset (N=6695), 41% of samples were FRα positive, 54% were negative, and 5% were not evaluable. Of 2000 samples with known primary tumor origin: 1632 (82%) were primary or metastatic sites of disease from patients with ovarian cancer (39% FRα positive); 266 (13%) originated from distinct non-ovarian tumor types, including primary peritoneal, Müllerian, pelvic, and gynecological sites of unspecified origin (combined 33.1% FRα positive); the remaining 102 (5%) had unconfirmed tumor origin. Of 1678 samples with known histological subtype, FRα positive expression was observed in samples from serous (43.5% [646/1486]), mixed (37% [14/38]), and endometrioid (10% [5/48]) histology, whereas FRα positive expression was absent across all clear cell (0/91) and mucinous subtypes (0/15). FRα positive expression was detectable in unique patient samples across all sample ages, which ranged from <1 to >5 years. Conclusions: RWanalysis of 6695 patient tumor samples referred for FRα testing demonstrated FRα prevalence of 41%. Ovarian cancers represented most cases evaluated, and the 39% FRα positive rate was consistent with the FRα positive prevalence observed in MIRV clinical trials (35%). This RW analysis demonstrates that FRα is measurable in samples ranging from <1 to >5 years, suggesting prior samples may be utilized for FRα testing if a recent sample is unattainable. This is consistent with previous analyses from MIRV clinical trials where most patients with PROC were selected for enrollment based on high FRα expression from archival tumor specimens. Collectively, this large RW data set helps to characterize the RW prevalence of the clinically actionable biomarker FRα and supports using FRα testing for personalized ovarian cancer treatment plans.