Abstract 1950: Enhanced detection of pancreatic cancer through methylation-based liquid biopsy: insights into tumor heterogeneity, early alterations, and stromal contributions

Abstract Pancreatic cancer is one of the deadliest malignancies, with a high mortality rate attributed to late-stage diagnosis and limited treatment options. Recent advancements in liquid biopsy, particularly through cell-free DNA (cfDNA) analyses, offer hope for early detection and prognosis. Specifically, cfDNA methylation profiling shows promise, outperforming traditional mutation-based methods such as single nucleotide variants (SNVs) and copy number variations (CNVs) in terms of sensitivity and specificity. In this study, we conducted paired CAPP-seq and EM-seq analyses on 29 pancreatic cfDNA samples. Our results indicated that methylation-based detection achieved a detection rate of 62.10% with 95% specificity, significantly higher than the 44.82% detection rate observed with CAPP-seq. Following the application of JEEPERS (Jagged-End Error Polishing of Enzymatically misRepaired Sequences) to mitigate methylation errors from jagged ends, the detection rate further improved to 72.41%. This aligns with existing literature highlighting the superior sensitivity of methylation profiling. We propose three hypotheses to explain this enhanced sensitivity: (1) Some pancreatic tumors exhibit CNVs prior to SNVs or contain subclonal KRAS mutations that are not detectable via SNV-based methods; (2) Methylation alterations may occur early during the PanIN stage, allowing the release of both normal and tumor-derived epithelial cells into circulation, identifiable by tissue-specific methylation markers; (3) PDAC tumors characterized by increased fibroblast content and epithelial-to-mesenchymal transition engage unique methylation signatures. To validate these hypotheses, we analyzed both pancreatic tumor tissues and cfDNA samples using CAPP-seq, LPWGS, emWGS, and single-cell nuclear sequencing data. Our findings confirm that CNVs frequently precede SNVs and highlight that pancreatic cfDNA comprises a notable proportion of epithelial cells. Importantly, methylation-based methodologies are capable of detecting signals from both normal and tumor-derived epithelial cells, leveraging cell-type-specific patterns. In contrast, SNV-based approaches are restricted to tumor-only signals. Furthermore, fibroblast methylation markers elucidate stromal contributions to the overall methylation landscape. These findings underscore the potential of methylation-based cfDNA analysis as a robust method for pancreatic cancer detection. We advocate for further validation through the development of a targeted panel that integrates SNVs, CNVs, and cell-type-specific markers to enhance diagnostic accuracy in clinical settings. Citation Format: Rui Wang, Diego Almanza, Mari Olsen, Xiaoman Kang, Daniel King, Maximilian Diehn, Ash Alizadeh. Enhanced detection of pancreatic cancer through methylation-based liquid biopsy: insights into tumor heterogeneity, early alterations, and stromal contributions [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 1950.