Serial ctDNA dynamics predict clinical outcomes in metastatic and locally advanced PDAC: a systematic review.

Background: Circulating tumor DNA (ctDNA) is being studied for prognosis and treatment monitoring in pancreatic ductal adenocarcinoma (PDAC). Given the limitations of CA19-9, including non-secretion and confounding by cholestasis, we assessed whether serial ctDNA dynamics predict progression-free survival (PFS) and overall survival (OS) in localized and advanced PDAC, and whether ctDNA adds information to CA19-9. Methods: We conducted a PRISMA-guided systematic review of PubMed, Translational Research, MDPI, and JAMA. Eligible studies included patients with PDAC who had ≥2 ctDNA measurements during perioperative care, treatment, or surveillance and reported associations with OS, PFS, DFS/RFS, recurrence, progression, EMR, molecular progression, or lead-time versus imaging/CA19-9. Two reviewers screened 119 records, of which 17 studies met inclusion criteria. Because of heterogeneity in assays, sampling schedules, and reporting, results were synthesized narratively. Results: Across studies, ctDNA detectability and clinical utility varied according to disease stage, assay type, and sampling window. In resected PDAC, a CLIA-validated ddPCR KRAS assay study reported preoperative ctDNA detection in 49% of patients (29/59). In the same cohort, ctDNA detected during follow-up predicted recurrence with 90% sensitivity (95% CI, 74%-98%) and 88% specificity (95% CI, 62%-98%) and preceded clinical or radiographic recurrence by a median of 84 days (IQR, 25-146). In advanced PDAC treated with palliative chemotherapy, longitudinal sampling every 4 weeks showed that ctDNA monitoring identified progression before radiology in 20 of 30 patients (67%) with baseline-detectable ctDNA, with a median lead time of 23 days (P = 0.01). In the same cohort, CA19-9 increases, when present, showed a shorter median lead time of 9.5 days (P = 0.03), and ctDNA detected progression in some patients without a corresponding CA19-9 increase during therapy. Across the included studies, baseline ctDNA positivity was generally associated with higher-risk disease, whereas early ctDNA decline or clearance during treatment and postoperative or longitudinal ctDNA negativity were associated with more favorable outcomes. Conclusion: Serial ctDNA dynamics in PDAC were associated with recurrence, progression, treatment response, and survival across perioperative and systemic therapy settings. ctDNA appears most consistent as a postoperative MRD marker and may complement CA19-9 and imaging during treatment monitoring. Prospective studies are needed to standardize sampling schedules, response thresholds, and ctDNA-guided clinical use.