Abstract A027: Validation of an early-stage pancreatic cancer classification model using 5-Hydroxymethylation profiles in plasma-derived cell-free DNA

Abstract

Abstract Background: Pancreatic cancer outcomes are poor, mostly due to the detection of cancer at late stages. Early-detection enables prompt initiation of treatment and has been reported to significantly improve outcomes. We have developed a novel sequencing-based epigenomics technology for cancer early-detection that enriches 5-hydroxymethylcytosine (5hmC) loci in cell-free DNA (cfDNA), and via machine learning algorithms, enables the detection of pancreatic cancer. Method: Whole-blood was obtained from a training cohort of 660 individuals (consisting of 132 pancreatic cancers (PaCa) and 528 non-cancers) and a validation cohort of 593 individuals (consisting of 86 PaCa and 507 non-cancers). cfDNA was isolated from plasma and 5hmC and whole-genome libraries were sequenced. Logistic regression algorithms were employed using 5hmC feature sets combined with physical characteristics to optimally partition cancers from non-cancers. Results: We trained and cross-validated a binomial logistic-regression model on PaCa and non-cancer samples. Cross validation of the training model yielded an overall sensitivity of 65.9%, early-stage (stage I-II) sensitivity of 57.1% and a specificity of 98%. The model was further validated on an independent set and yielded an overall sensitivity of 64%, early-stage sensitivity of 64.7% and a specificity of 98%. Conclusion: Our results demonstrate that plasma-derived cfDNA 5hmC profiles enable the detection of PaCa in early-stage disease providing a valuable tool especially for those individuals at high risk. A larger clinical study (NODMED: NCT05188586) is underway to validate the test in a high-risk population that would most benefit from early-detection. Citation Format: Anna Bergamaschi, David Haan, Micah Collins, Gulfem Guler, Melissa Peters, Lauren Gigliotti, Shimul Chowdhury, Wayne Volkmuth, Samuel Levy. Validation of an early-stage pancreatic cancer classification model using 5-Hydroxymethylation profiles in plasma-derived cell-free DNA [abstract]. In: Proceedings of the AACR Special Conference on Pancreatic Cancer; 2022 Sep 13-16; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2022;82(22 Suppl):Abstract nr A027.