Immunologic and clinical outcomes of perfluoroalkyl substances (PFAS) exposure in solid tumors treated with pembrolizumab.

2539 Background: Perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS) are two major PFAS with persistent and widespread human exposure. Epidemiologic studies in non-cancer individuals suggest PFAS-associated immune dysregulation and anti-inflammatory effects. We assessed immunologic and clinical effects of PFAS on patients (pts) with advanced solid tumors treated with pembrolizumab. Methods: In the investigator-initiated INSPIRE study (NCT02644369) (n = 106) of advanced solid cancer pts treated with pembrolizumab (head and neck squamous cell cancer, triple-negative breast cancer, high-grade serous carcinoma, melanoma and other mixed solid tumors), we quantified baseline PFOS and PFOA via HPLC-tandem mass spectrometry. PFAS were evaluated as continuous variables and dichotomized at the median ( < median vs > median; PFOS 3.6 ng/ml; PFOA 1.0 ng/ml). Associations with clinical and demographic factors were assessed using correlation/regression analyses. Association with toxicity and objective response (ORR) used Wilcoxon rank-sum tests; progression-free survival (PFS) and overall survival (OS) used Cox proportional hazards models. Immune correlates (immune infiltration measured by multiplex IHC, bulk-RNA sequencing of tumor tissue and plasma cytokine levels measured by Luminex multiplex cytokine assay) were evaluated using median-split Wilcoxon comparisons with Benjamini-Hochberg false discovery rate (FDR) control (q < 0.2). Results: Baseline PFOS and PFOA levels were moderately correlated (Pearson r = 0.49, p < 0.001). PFOS levels were significantly higher with increasing age (median age 59.4 years; range 21.1-81.8; p = 0.016). PFOA levels were significantly higher with increasing age (p = 0.001), in males (p = 0.042), and in the melanoma cohort (p = 0.015), and significantly lower in pts with prior systemic therapy (p = 0.001). For immune parameters, higher PFOA exposure was associated with reduced CD4 T-cell infiltration in tumor and stromal areas (q = 0.07). RNA sequencing of genes and immune signatures showed no associations reaching FDR significance. 10 cytokines were significantly lower in pts with high PFOA levels: IL-16, IL-2R, IL-3, HGF, IFN-γ, MCP-2/CCL8, MDC/CCL22, TNF-RII, TSLP, and BLC/CXCL13 (all q = 0.15) in ovarian cancer and melanoma pts. Neither PFOS nor PFOA levels showed significant associations with ORR, toxicity, PFS or OS. Conclusions: PFOS and PFOA levels vary by demographic and clinical factors but show no direct association with toxicity and treatment outcomes in pts treated with pembrolizumab. Elevated PFOA exposure correlates with reduced T-cell infiltration and cytokine suppression, potentially inducing systemic and local immune dysregulation. These findings highlight immunologic effects of PFAS and warrant mechanistic and longitudinal studies.