Abstract 7036: Integrative biomarker profiling for predicting immunotherapy outcomes in gynecologic cancers

Abstract Immunotherapy with immune checkpoint inhibitors (ICI) has proven effective for various solid tumors, such as those in the lungs, urinary tract, and skin. However, its efficacy in gynecologic (GYN) cancers has been limited, with response rates of 11-17% in recurrent cases. Only a minority of patients with DNA mismatch repair deficiencies exhibit favorable responses to PD-1 inhibitors. The predictability of PD-L1 expression in GYN cancers remains inconsistent. Our research aimed to identify and prioritize biomarkers and mechanisms linked to ICI resistance in GYN oncology, targeting those who could benefit from these treatments. We gathered clinical and pathological data from 49 patients with ovarian, endometrial, and cervical cancers at various stages, all treated with ICI. We performed comprehensive analyses, including transcriptome-based immune cell abundance estimation, signaling pathway enrichment, and correlation with ICI treatment responses and survival outcomes. Our previous work highlighted TGF-β and CD47's roles in immunotherapy resistance, facilitating an immunosuppressive environment within our GYN cancer cohort. Moreover, we conducted a stoichiometry constraint-based genome-scale metabolic model and computed correlation-grouped metabolic activity scores (cgScore) using metabolic enzyme gene expression as constraints. Among 1533 metabolic reactions, pyruvate transport and folate metabolism stood out as the most upregulated in responders versus those who progressed. Upon identifying several promising biomarkers, we used logistic regression classifiers to decipher the predictive value of each. Our model included clinical-demographic (age, BMI, cancer type, stage) and molecular-genomic features (immune marker gene expression, TopTGF-β score, immune cell abundance from transcriptomic deconvolution, and the top 2 cgScores). Analysis divided into training and testing sets highlighted features such as stage, BMI, eosinophils, CD274, CD47, and the pyruvate transport reaction as significant predictors of ICI response. The model achieved an 80% accuracy rate on the test set, confirmed by permutation importance from scikit-learn. We also used a Regularized Cox model to link predictors with progression-free survival, emphasizing the roles of BMI, plasma cells, CD276, pyruvate transport reaction, and folate metabolism. In conclusion, there are immunological and metabolic markers have the potential to serve as predictive indicators for ICI treatment response and survival in GYN cancers. Understanding these predictors can profoundly influence clinical decision-making in immunotherapy, ultimately leading to enhanced patient outcomes and personalized treatment approaches. Citation Format: Gideon Idumah, Lin Li, Haider Mahdi, Ying Ni. Integrative biomarker profiling for predicting immunotherapy outcomes in gynecologic cancers abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts) ; 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84 (6Suppl): Abstract nr 7036.