Abstract 1854: Plasma proteomic biomarkers identify non-responders and reveal biological insights about the tumor microenvironment in melanoma patients after PD1 blockade

Samuel J. Wright,Deepika Yeramosu,Milan Parikh,Marijana Rucevic,Ngan Nguyen,Russell W. Jenkins,Keith T. Flaherty,Nir Hacohen,Genevieve M. Boland,Arnav Mehta

DOI: https://doi.org/10.1158/1538-7445.am2024-1854

IF: 11.2

2024-03-26

Cancer Research

Abstract:Most patients treated with immune checkpoint blockade (ICB) do not have durable treatment responses, stressing a critical need to identify early non-invasive biomarkers of response. Circulating biomarkers provide easy access for serial monitoring and can provide insight on the mechanisms of response to ICB. We performed plasma proteomics (~3000 proteins) on 250 metastatic melanoma patients before and on ICB treatment and performed analysis to identify response- and time point-associated peripheral protein biomarkers. We next generated patient-matched peripheral blood lymphocyte (PBL) surface proteomic data and tumor sample bulk transcriptomic data to build a multimodal dataset. We leveraged these data to train machine learning models for unsupervised analysis and to classify treatment responses and learned features predictive of survival and patient endotypes. Using univariate linear models to predict protein abundance based on time point, response and the interaction term for time point and response, we identified 343 time point-associated proteins, 141 response-associated proteins, and one interaction-associated protein. We further investigated response-related proteins by fitting multivariate logistic regression and cox proportional hazards regression models to further investigate the ability of these proteins to predict response, progression-free survival (PFS), and overall survival (OS) and identify those proteins that are most important in predicting response. We calculated covarying modules of proteins that were found to be either response-associated or time point-associated to discover underlying biological networks at different treatment time points and disease response states. Additionally, we leveraged these modules to endotype patients and identified trends in protein modules across patient subsets, including modules related to biological processes such as immune infiltration, apoptosis, metaplasia and cell adhesion. To understand the contribution of different compartments on plasma proteomic abundance, we studied correlations of plasma protein abundance with PBL surface protein abundance and tumor bulk mRNA-sequencing expression. Finally, we integrated our data modalities and fit various statistical models to evaluate the relative importance of plasma protein biomarkers in predicting response compared to PBL surface proteins and tumor bulk mRNA-sequencing expression. Together, these data represent one of the deepest peripheral biomarker studies using paired samples in melanoma patients treated with anti-PD1 therapy. Citation Format: Samuel J. Wright, Deepika Yeramosu, Milan Parikh, Marijana Rucevic, Ngan Nguyen, Russell W. Jenkins, Keith T. Flaherty, Nir Hacohen, Genevieve M. Boland, Arnav Mehta. Plasma proteomic biomarkers identify non-responders and reveal biological insights about the tumor microenvironment in melanoma patients after PD1 blockade [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular s); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl) nr 1854.

oncology