1555P Cell-free DNA Analysis in Patients with Metastatic Gastroesophageal Adenocarcinoma: Preliminary Results of the REGIRI - PRODIGE 58 Ancillary Study

Cell-free DNA (cfDNA) analysis has shown promising results for the management of patients with cancer, but few data exist in patients with metastatic gastro-esophageal adenocarcinomas (mGEA). In this study, we analyzed cfDNA extracted from blood of patients included in the REGIRI - PRODIGE 58 trial (Regorafenib + irinotecan as 2nd line in patients with mGEA). 239 samples from 34 patients with mGEA included in the Regiri arm were available for analysis. Plasma samples were taken before treatment’s initiation, then at days 1, 2, 8 and 15 of Cycle 1 and days 1, 8 and 15 of Cycle 2. cfDNA was extracted from plasma, concentrations and fragments sizes determined by fragment analysis. First and last time points were selected to estimate concentration’s evolution for each patient. A cfDNA threshold concentration (tc) of 0.55 ng/μL was settled as the best threshold at baseline to separate patients into 2 groups “low” and “high” corresponding to concentrations lower or greater than tc respectively. Among the 34 patients analyzed, 1 patient was excluded because of genomic DNA contamination in plasma at baseline. 16 patients (48.48%) were in the “low” group and 17 (51.52%) in the “high” group. For cfDNA concentration at baseline and PFS, a trend non-statistically significant difference (HR=0.53; CI95[0.2595;1.085]; p=0.0564) was found: Low cfDNA concentration at baseline was associated with a better PFS (2.6 vs 1.7 months). 4 distinct subgroups emerged: Patients with low concentration at baseline that remains low (low-low), low at baseline with an increase during follow-up (low-high), high at baseline with a decrease during follow-up (high-low) and high at baseline that remains high (high-high). A statistically significant difference was found for PFS (4.55 vs 2.11 months) between the low-low and high-high subgroups respectively (HR=0.4139; CI95[0.1652;1.037]; p=0.022). Our findings suggest that cfDNA concentration at baseline and cfDNA clearance might have an interest to predict PFS in patients with mGEA. Further investigations are ongoing to decipher if genomic signatures can be predictive of PFS in patients with mGEA.