Abstract 5937: Development of a protein and clinical multiparametric biomarker panel for assessing pancreatic adenocarcinoma risk

Detection of pancreatic adenocarcinoma represents a major burden on the healthcare system due to its latent detection, generally in the metastatic stage. To this end, we developed a prospective biomarker clinical trial to clinically and molecularly characterize at risk and confirmed pancreatic adenocarcinoma patients. This multicenter (n=6) clinical trial (NCT 02781012) of PDAC, referred to as Project Survival, combined high-fidelity longitudinal phenotypic characterization and multi-omic profiling to identify biomarkers with diagnostic and therapeutic utility. The study included collecting plasma, serum, buffy coat, urine, saliva, and tumor tissue samples and followed 140 non-cancer patients and 280 PDAC subjects for up to 7 years. Plasma samples were analyzed to identify circulating biomarkers using comprehensive proteomics profiling by LC MS/MS analysis. We first identify proteins that demonstrated a significant change between non-cancer (e. g., relatives with no signs of cancer, pancreatitis patients) and cancer patients. Then, top analytes were examined for impact from confounders (age, gender, BMI, and diabetes) using multivariate regression analysis. Four proteins were selected that demonstrated a robust signal with minimal impact from confounders. Next, multivariate logistic models were fit to distinguish cancer patients from non-cancer patients along with stratification across cancer staging to determine the predictive performance of the four-protein panel. Finally, an additional panel was developed that incorporated clinical features along with the protein panel to further improve predictive performance. We identified a distinct panel of four proteins that produced an optimal predictive diagnostic performance. The protein panel demonstrated ROC AUC of 0.77 for distinguishing cancer (n=273) from non-cancer (n=140) patients (PPV 0.76, NPV 0.70, OR 7.2), followed by ROC AUC of 0.70, 0.75, and 0.85 for distinguishing early stage (n=65), locally advanced (n=117), or metastatic (n=91) patients from non-cancer patients, respectively. Further, adding the four clinical features (Age, History of Pancreatitis, Family History of Cancer, and Weight Loss at Study Entry) further improved the ROC AUCs to 0.85, 0.79, 0.86, and 0.90 respectively. This prospective biomarker-driven clinical trial identified a distinct four-panel biomarker for Pancreatic Ductal Adenocarcinoma. The four proteins individually demonstrated a robust ability to distinguish cancer patients while controlling for confounding factors. The four-protein diagnostic panel showed superior ability to identify cancer patients and showed high predictive performance based on staging predictions. The panel is currently being explored for commercial development. Citation Format: Nischal Mahaveer Chand, A James Moser, Eric M. Grund, Poornima K. Tekumalla, Madappa N. Kundranda, Gregory M. Miller, Niven R. Narain, Vijay Modur, Michael A. Kiebish. Development of a protein and clinical multiparametric biomarker panel for assessing pancreatic adenocarcinoma risk [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 5937.