Initial Report: Personalized Circulating Tumor DNA and Survival in Patients with Resectable Pancreatic Cancer

ABSTRACT: BACKGROUND: Pancreatic adenocarcinoma (PDAC) is highly lethal with up to 80% of resected patients experiencing disease recurrence within 2 years (Watanabe, Nakamura, Kimura et al in Int J Mol Sci 23(19):11521, 2022). Cross-sectional imaging and serum tumor markers are used for monitoring post-operative recurrence; however, both have significant limitations (Edland, Tjensvoll, Oltedal et al in Mol Oncol 17:1857-1870, 2023). Circulating tumor DNA (ctDNA) has emerged as a valuable prognostic tool to measure molecular residual disease (MRD) and predict recurrence in solid tumors (Watanabe, Nakamura, Kimura et al in Int J Mol Sci 23(19):11521, 2022). In this study, we evaluated the feasibility of a personalized, tumor-informed ctDNA assay to detect recurrence prior to standard surveillance tools in patients with PDAC. PATIENTS AND METHODS: After Institutional Review Board (IRB) approval (Pro00106870), we assessed serial ctDNA measurements (n = 177) from 35 patients with resectable PDAC treated by either upfront resection or neoadjuvant chemotherapy. Plasma samples (median 4 ml, interquartile range 0.6-5.9 ml) were isolated from blood collected in EDTA tubes and banked at diagnosis, during neoadjuvant therapy if applicable, on the day of surgery, and every 2-3 months postoperatively. A tumor-informed assay (Signatera™, Natera, Inc.) that tracks up to 16 individual-specific, somatic single nucleotide variants in the corresponding patient's plasma samples were used for ctDNA detection. Survival was calculated using Kaplan-Meier curves, and significance was determined with the log-rank test. RESULTS: Personalized ctDNA assays were successfully designed for all patients (with 32/35 patients having 16-plex assays). Median follow-up from initial treatment was 13 months (range 1-26 months; Table 1). ctDNA-positivity at any time point was observed in 40% (14/35) of patients. During the follow-up period, 18 patients (51%) developed radiographic evidence of recurrence after a median of 9 months of follow-up (range 1-26 months). At the time of radiographic recurrence, 50% (9/18) of patients were ctDNA-positive. During the immediate postoperative period (up to 9 weeks post-surgery), RFS and OS were significantly inferior in patients who were ctDNA-positive versus ctDNA-negative (RFS 97 versus 297 days, p < 0.001; OS 110 versus 381 days, p < 0.001; Fig. 1). Table 1 Cohort demographics (N = 35); patient demographics, tumor characteristics, and survival Gender (%) Female 17 (49%) Male 18 (51%) Median age (IQR) 70 years (65-75 years) Neoadjuvant treatment (%) 11 (31%) Median sample plasma volume (IQR) 4.0 mL (0.6-5.9 mL) Median follow-up (range) 13 months (1-26 months) Median initial CA 19-9 in U/mL (IQR) 56 (18-160) Median tumor size in cm (IQR) 2.5 (1.8-3.3) Median number of positive lymph nodes (IQR) 1 (0-3) Median recurrence-free survival 9.4 months Median overall survival N/A (not reached) Fig. 1 a Overview plot showing longitudinal ctDNA status, treatment regimen, and clinical outcomes for each patient (N = 35); median follow-up from the start of the neoadjuvant therapy/surgery was 13 months (range 1-26 months); ctDNA at any time point was 40% (14/35); out of the 35 patients, 18 (51%) developed radiographic evidence of recurrence (median RFS: 9 months), and of these 18 patients with clinical recurrence, 9 (50%) were ctDNA-positive and the remaining ctDNA-negative; notably, all ctDNA-negative patients with recurrence had suboptimal plasma volume available for ctDNA analysis; b, c Kaplan-Meier estimates representing the association of ctDNA status with (b) RFS and (c) OS, at MRD time point (9 weeks post-surgery) DISCUSSION: Our study demonstrates the feasibility of tumor-informed ctDNA-based MRD testing in resectable PDAC and shows that MRD detected by ctDNA within the immediate postoperative period portends a dismal prognosis. This information is valuable for both patients and clinicians in setting prognostic expectations.