Circulating tumor DNA for predicting recurrence and mortality in patients with resected Melanoma: A systematic review and meta-analysis

BACKGROUND: Resectable melanoma remains associated with substantial relapse risk despite adjuvant therapy, and prognostic biomarkers for better risk stratification are urgently needed. We evaluated the prognostic and diagnostic performance of circulating tumor DNA (ctDNA) in patients with resectable melanoma. METHODS: PubMed, Embase, and Cochrane databases were systematically searched through July 2025 for studies evaluating ctDNA detection in patients with resected melanoma and reporting outcomes of interest. All ctDNA data were assessed at three timepoints: baseline (pre-surgery), landmark (≤12 weeks post-surgery), and longitudinal surveillance. Random-effects models estimated hazard ratios (HR) with 95% confidence intervals (CI) for relapse-free survival (RFS), distant metastasis-free survival (DMFS), and overall survival (OS). Diagnostic performance was evaluated using both univariate and bivariate models. RESULTS: Twenty-four cohorts comprising 3,032 patients were included, with a median follow-up of 38.5 months. Baseline ctDNA positivity was associated with worse RFS (HR 3.14; 95% CI 1.82-5.41) and DMFS (HR 2.25; 95% CI 1.50-3.36), but not OS. At the landmark timepoint, ctDNA positivity predicted worse RFS (HR 3.42; 95% CI 2.51-4.65), DMFS (HR 5.38; 95% CI 3.09-9.37), and OS (HR 3.55; 95% CI 2.63-4.79). Longitudinal ctDNA detection showed the strongest prognostic association with RFS (HR 3.76; 95% CI 2.83-4.99). Diagnostic performance demonstrated modest sensitivity (31-53%) but high specificity (>90%) at landmark and longitudinal timepoints, with no significant differences in accuracy between these timepoints. CONCLUSION: Detectable ctDNA in resectable cutaneous melanoma is a negative prognostic marker associated with significantly higher rates of relapse, distant metastasis, and mortality. Diagnostic metrics revealed that plasma ctDNA has high specificity with suboptimal sensitivity.