Abstract 2860: Impact of hemolysis on multi-omic pancreatic cancer biomarker discovery: De-risking precision medicine biomarker development

Abstract

Abstract Biomarker analysis is critically dependent on the quality of biofluid or tissue samples obtained from human research studies. Although proteomic, lipidomic, and metabolomic analyses can be dramatically impacted by the time of sample collection, fasting status, and participant demographics, the hemolysis status of plasma, serum or buffy coat samples is a poorly understood confounder of sample quality. Hemoglobin levels can range between 0 - 10 g/L in samples (referred to as 1-4 in plasma/serum and 0-4 level in buffy coat) as a marker of hemolysis severity and sample contamination by reticulocyte-derived analytes. In oncology clinical trials, patients can be more susceptible to hemolysis due to chemotherapy treatment, which can impact sample assessment and study results. Herein, we analyzed 941 plasma and 950 serum samples using comprehensive proteomics, structural lipidomics, signaling lipidomics, and metabolomics in a pancreatic cancer biomarker discovery program referred to as Project Survival as well as 951 buffy coat samples using only proteomic analysis. Project survival is a 7-year longitudinal pancreatic cancer biomarker discovery trial analyzing 400+ pancreatic cancer and at-risk patients using multi-omic and multiple biofluid assessment. To date this study yielded samples in plasma with 92.3% - #1, 6.5% - #2, 1.2% - #3, and 0% - #4 hemolysis, serum with 94.8% - #1, 3.8% - #2, 1.4% - #3, and 0% - #4 hemolysis and buffy coat 42.7% - #0, 25.6% - #1, 20.8% -#2, 10.4% - #3, and 0.5% - #4 hemolysis. Multi-omic and regression analysis of sample data for hemolysis status revealed a distinct pattern of OMIC variables correlated with the degree of hemolysis. Proteomics analysis was the greatest impacted in terms of the protein identification and quantitation. Additionally, pathway analysis revealed expected pathways associated with hemolysis and coagulation, but also unknown pathways and corresponding proteins that were differentially correlated with hemolysis state. Additionally, metabolomics and lipidomics analysis also revealed distinct differentials associated with hemolysis state. Herein, our analysis is the first to analyze thousands of samples using multi-omics revealing critically informative molecular differentials across OMIC technologies demonstrating that caution should be given to avoid these identified biomarkers for translational development. Citation Format: Michael A. Kiebish, Punit Shah, Valerie Bussberg, Vladimir Tolstikov, Rick Searfoss, Kennedy Ofori-Mensa, Eric M. Grund, Abena Darkwah, Emily Y. Chen, Bennett Greenwood, Ellaine Adu Ntoso, Leonardo Rodrigues, Mia Liu, Elder Granger, Chas Bountra, Rangaprasad Sarangarajan, A J. Moser, Niven R. Narain. Impact of hemolysis on multi-omic pancreatic cancer biomarker discovery: De-risking precision medicine biomarker development [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2860.