Abstract 743: A prognostic model based on plasma proteomics for predicting immuno-chemotherapy outcomes in advanced non-small cell lung cancer

Abstract Background: Predicting benefit from immune checkpoint inhibitors in advanced non-small cell lung cancer (aNSCLC) remains challenging. We developed a plasma proteomic model to evaluate immuno-chemotherapy (IC) outcomes. Methods: Blood samples were collected prior to 1st-line IC for analysis of plasma proteins and peripheral blood mononuclear cells (PBMC) from 194 aNSCLC patients without EGFR/ALK mutations. The Olink Immuno-Oncology Panel was used to detect the soluble proteins in plasma, and transcriptome sequencing was conducted on PBMC. Results: We optimized a prior prognostic model by incorporating a validation cohort and employing the Survival SVM method. This improved model, trained on plasma levels of eight proteins (CD27, TWEAK, CD8A, GZMH, IL6, PGF, TIE2, and TNFRSF12A), independently predicted overall survival (OS). Kaplan-Meier analysis validated significant differences in PFS and OS (p < 0.001, respectively). Repeated 10-fold cross-validation in the discovery cohort demonstrated model robustness, with stable predictive power for 12-month OS and PFS (AUCs: 0.86 and 0.68, respectively), which was further supported in the validation cohort (AUCs: 0.78 and 0.63, respectively). The model's independent prognostic value was further substantiated through Cox regression analyses (Table). Our model exceeded the performance of conventional PD-L1 TPS cutoff values of 1% or 50% in stratifying patient risk. Low-risk patients exhibited enriched memory B cells (p < 0.05), while high-risk patients showed an upregulation of angiogenesis-related proteins (p < 0.05), consistently observed across both cohorts. Validation of the eight proteins is ongoing using ELISA assays to further confirm their clinical utility. Conclusion: We established and validated an optimized plasma protein-based prognostic model that identifies distinct immune phenotypes and provides critical insights into aNSCLC progression. Citation Format: Zhihuang Hu, Jialei Wang, Changbin Zhu, Xin Zhang, Shenglin Huang, Si Sun, Zhenhua Wu, Yao Zhang, Ying Lin, Jin Wang, Tao Hu, Zhan Huang, Hui Yu, Xinmin Zhao, Huijie Wang, Xianghua Wu. A prognostic model based on plasma proteomics for predicting immuno-chemotherapy outcomes in advanced non-small cell lung cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 743.