Investigating survival in patients undergoing surgery for pancreatic ductal adenocarcinoma using liquid biopsy: Preliminary results of PAN-CGE-BLISS pilot study.

e16454 Background: Analysis of cell-free DNA (cfDNA) is an alternative approach to conventional tumor markers in monitoring solid tumors. Although promising, data about cfDNA utility are lacking in pancreatic cancer, particularly for the follow-up of patients undergoing surgery. The present study aimed to evaluate the role of cfDNA levels and genomic variants detected by next-generation sequencing (NGS) as predictive biomarkers of best outcomes in patients with resected pancreatic ductal adenocarcinoma (PDAC). Methods: This multicentric retrospective study included 15 patients with resectable PDAC. All patients were previously enrolled in the PANCREAS CGE study (NTC02818907). Blood samples were collected at baseline; post-neoadjuvant treatment or prior surgery; post-surgery; and post-adjuvant chemotherapy. cfDNA from 59 plasma samples was quantified and qualified using fluorescence-based assays and sequenced using the Illumina instrument and a bespoke gene-panel. FFPE samples were collected at surgery and analyzed using a 519-gene panel (Nonacus). The detection and tracking of somatic variants in cfDNA was performed., The relationship of variant allele frequency (VAF) and cfDNA concentration with clinical outcomes were investigated. Results: At baseline, patients with low cfDNA concentrations ( < 0.4 ng/µL) demonstrated a significantly higher overall survival (OS) than those with high cfDNA concentrations (44.5 versus 24.3 months; hazard ratio (HR) = 0.154; 95% confidence interval (CI) = 0.02-0.88; p = 0.036, respectively). Following surgery, the median OS was 23.7 and 45.9 months in patients with high and low cfDNA concentrations ( > 1.05 and < 1.05 ng/µL), respectively (HR = 0.170; 95% CI = 0.03-1.44; p = 0.0140). Patients with cfDNA concentration at least twice as high postoperatively as preoperatively exhibited a lower OS than those with stable or decreased concentration during the perioperative period (27.2 versus 45.9 months; HR = 0.2; 95% CI = 0.04-0.97; p = 0.0281). NGS-based detection of variants in cfDNA revealed that increasing VAFs were associated with poorer outcomes. cfDNA concentrations after adjuvant treatment and corresponding variant burden were predictive of disease-free survival (DFS), with a cut-off of 0.25 ng/µL (HR = 0.136; 95% CI = 0.04-0.48; p = 0.0247). Conclusions: This study demonstrates that baseline level of cfDNA concentration and somatic VAFs detected by NGS are independent prognostic factors in PDAC patients. The combination of cfDNA concentrations, VAF dynamics, and variant detection during the perioperative period provides deeper insights into tumor biology and patient prognosis. Additionally, cfDNA concentrations and variant burden following adjuvant treatment are effective in predicting recurrence. These results highlight the application of long-term cfDNA and NGS-based monitoring in clinical practice for PDAC patients as readily available, minimally invasive, and cost-effective biomarkers.