Abstract B005: Integrated profiling of pancreatic tumor interstitial fluid to develop blood-based early detection signatures

Abstract

Pancreatic ductal adenocarcinoma (PDAC) is the most common form of pancreatic cancer and lacks effective therapies, driving dismal survival. Because most patients present with metastatic disease, when curative intervention is no longer possible, there is an urgent unmet need for biomarker panels that enable early detection. Exploiting the high interstitial pressure and fluid buildup characteristic of pancreatic tumors, we developed a unique, IRB approved protocol to collect tumor interstitial fluid (TIF) from tumor punch biopsies before blood vessel clamping. Collecting fluid before clamping limits hypoxia and nutrient deprivation induced cell death, thereby maintaining cell viability and preserving the native TIF contents. The resulting sample provides a faithful snapshot of the factors secreted into the tumor microenvironment. Using TIF, we performed targeted (∼120 polar metabolites) and untargeted metabolomics alongside SomaScan aptamer-based detection of ∼11,000 proteins. On matched biopsy tissue, we performed Xenium and Visium HD spatial transcriptomics and spatial proteomics (∼60 markers spanning immune activation and cell identity), enabling powerful correlative, integrated -omics analyses of tumor biology alongside paired TIF and plasma comparisons. We aim to define whether a TIF informed multifactor panel can provide the basis for a simple blood test for earlier PDAC detection, how transcriptional programs in specific cell types shape TIF and blood contents, how spatial interactions among cell populations relate to secreted factors, whether clinical parameters (e.g., overall survival, treatments, stage, and patient background) correlate with the multi-parameter -omics signatures, and which secreted factors and spatial transcriptome signatures track with immune cell activation. In preliminary analyses of 10 TIF samples plus 10 healthy donor, 10 pancreatitis, and 10 PDAC plasma samples, 23 metabolites showed strong TIF-plasma Spearman correlations. Of these, 10 metabolites distinguished healthy donor, pancreatitis, and PDAC plasma. Pairwise plasma metabolite comparisons showed progressively more positive correlations across healthy donor, pancreatitis, and PDAC groups, consistent with common metabolic programs emerging in patients with PDAC. Weighted gene co-expression network analysis of SomaScan data identified five candidate functional modules reflecting the tumor microenvironment, offering preliminary insight into pathways and cellular contributions in PDAC. Future work will integrate the multiomics datasets to address these questions and validate findings in additional cohorts, building a comprehensive ATLAS as a resource for the research community. The overarching goal is to develop improved prognostic and diagnostic panels and, ultimately, blood-based signatures for early detection of this devastating disease.