Use of serial ctDNA dynamics to predict clinical outcomes in metastatic and locally advanced PDAC: A systematic review.

Abstract

785 Background: Circulating tumor DNA (ctDNA) is a potential prognostic marker detectable in a substantial proportion of pancreatic ductal adenocarcinoma (PDAC). Because CA19-9 is limited by non-secretion and cholestasis, we assessed whether serial ctDNA dynamics predict progression free survival (PFS) and overall survival (OS) across localized and metastatic PDAC and how ctDNA performs relative to and in combination with CA19-9. Methods: We conducted a PRISMA guided literature search across PubMed, Embase, Web of Science, and Google Scholar. After removing duplicates, two reviewers screened 318 publications and 20 met prespecified criteria. Designs included metastatic and/or locally advanced cohorts with repeated on-treatment sampling (including post-hoc analysis), localized/resectable cohorts ± perioperative sampling, and mixed all stage observational series. Primary outcomes were OS and PFS; secondary outcomes were ctDNA detectability, early molecular response (EMR), and lead-time to progression/recurrence versus imaging and/or CA19-9. Due to assay/timing heterogeneity, results were synthesized narratively and where ≥3 comparable studies existed, we summarized pooled effect ranges. Results: ctDNA detectability varied by setting and platform (8.33-68.57% in localized; 38.8-86.1% in metastatic cohorts). Baseline ctDNA positivity was consistently associated with worse survival: localized (resectable) meta-analysis OS HR 2.27 (95% CI 1.13-4.56); advanced (post-hoc analysis) PFS HR 1.50 (1.03-2.18, p=0.034) and OS HR 1.62 (1.05-2.50, p=0.029). The wide CI in localized disease indicates uncertainty, likely due to small cohorts and variable follow-up, and warrants cautious interpretation. In locally advanced cohorts on systemic therapy, EMR aligned with longer PFS and OS, whereas molecular progression (rise or conversion to positive) frequently preceded radiographic progression by 7-12 weeks. CA19-9 kinetics were prognostic when available; ctDNA remained independently prognostic after adjusting for CA19-9 and added discrimination when combined. Reports suggesting ctDNA utility in CA19-9 non-secretors are promising, but non-secretor prevalence was inconsistently reported, limiting generalizability. Conclusions: Across localized and metastatic PDAC, baseline ctDNA positivity identifies higher risk patients, while early declines on serial ctDNA identify more favorable outcomes and often anticipate imaging by 1-3 months. ctDNA complements CA19-9 and warrants further clinical evaluation; routine ctDNA guided treatment changes are not yet recommended and require future prospective trials.