Development of a cfDNA-based protein-informed epigenetic signature (PEp-sig) to enable biomarker-based risk stratification for pancreatic ductal adenocarcinoma (PDA).

4149 Background: Novel blood-based biomarkers are needed for early risk stratification and personalized therapies for patients with PDA. An evidence-based 99-gene panel was developed focusing on genes whose protein products are implicated in PDA drug response. We hypothesized that methylation levels in these genes could serve as surrogates for their corresponding protein activity, enabling the prediction of treatment-related outcomes in PDA patients. Using this panel, we developed a cell-free DNA (cfDNA)-based PEp-sig for risk stratification of patients with PDA. Methods: Targeted enzymatic methylation sequencing was performed on plasma samples from PDA patients (01/2010–05/2022) receiving chemotherapy from the Ohio State University biorepository. Gene methylation levels were analyzed for associations with overall survival (OS) using univariate and multivariate Cox regression models. Significant genes and covariates such as first chemotherapy (FLC)- FOLFIRINOX (FFX) vs. gemcitabine (G)/nab-paclitaxel (NP) vs. other (Ot), and stage at diagnosis (StD) – early-stage (ES) that includes resectable and borderline resectable PDA vs. locally advanced (LA) and metastatic (Met)) were identified through backward selection. Risk scores from multivariate models were dichotomized at the median, stratifying patients into High (Hg) and low-risk groups (Lg), with survival differences assessed by Kaplan-Meier and log-rank tests. Results: The study cohort (SC) included 51 PDA patients (StD: 22 ES, 15 LA, and 14 Met). Among the ES cases, 3 progressed to Met after neoadjuvant therapy (NAT); in the LA sub-group, 2 proceeded to surgery post-chemotherapy. Ultimately, 21 patients had resection (Rs), while 30 underwent palliative therapy (PT), with FLC distribution as follows: PT group – 12 FFX, 16 G/NP, and 2 Ot; Rs group, NAT—9 FFX, 1 G/NP, and 1 Ot; adjuvant therapy after upfront surgery (UpS) —5 FFX, 2 G/NP, 3 Ot. Two 15-gene PEp-sig were developed: one for the SC and another for the PT group. A significant overlap (11/15) of genes was observed between the two PEp-sigs. These genes are linked to the response to G, NP, irinotecan, and platinums. The performance of PEp-sig models, with and without FLC and StD adjustments, is summarized below. Conclusions: In this proof-of-concept study, we present a cfDNA-based PEp-sig that effectively stratifies PDA patients by survival risk. Notably, it operates independently of FLC and StD within the PT group. Ongoing efforts aim to develop treatment selection algorithms based on these findings. Models tested SC PT-group Hg vs. LgOS (in months) Hazard Ratio (HR) p-value Hg vs. LgOS (in months) HR p-value PEp-sig alone 10.75 vs. 33 8.7 <0.001 5.3 vs. 16.83 9.2 <0.001 PEp-sig + FLC* 10.62 vs. 33 8.1 PEp-sig + StD* 8.4 vs. 33 16.9 8 *FLC and StD significantly impacted OS in SC but not in PT.