Skip to content
    Reviews

    Circulating tumor DNA clearance as a trial-level early efficacy signal in solid tumors: a systematic review and meta-analysis of randomized trials

    Tang et alTranslational cancer research

    Circulating tumor DNA clearance as a trial-level early efficacy signal in solid tumors: a systematic review and meta-analysis of randomized trials
    In an exploratory, hypothesis-generating systematic review and trial-level meta-analysis of 18 randomized controlled trials across solid tumors, circulating tumor DNA (ctDNA) clearance failed to establish validation as a trial-level surrogate endpoint for clinical outcomes. While ctDNA clearance serves as a sensitive marker of treatment response and minimal residual disease, the trial demonstrated that its association with hard clinical endpoints is highly setting- and endpoint-dependent.
    Statistical analyses utilized weighted regression, leave-one-out cross-validation, copula correlation, Bayesian hierarchical modeling, and sensitivity evaluations, prioritizing cross-validated coefficient of determination (R²cv) over apparent R² to assess out-of-sample predictive capability. The strongest trial-level correlation was observed in progression-free survival (PFS) among 8 palliative or metastatic trials, yielding an apparent R² of 0.77 (95% CI: 0.03–0.96) and an R²cv of 0.41. In contrast, overall survival (OS) showed weaker correlation, with an apparent R² of 0.56 (95% CI: 0.12–0.84) and an R²cv of 0.26.
    Surrogacy signals broke down entirely in curative settings evaluating recurrence-related outcomes, demonstrating an apparent R² of 0.04 (95% CI: 0.00–0.92) and an out-of-sample R²cv of -1.66. Sensitivity analysis excluding the ctDNA-guided DYNAMIC-III trial improved the apparent R² to 0.67, but cross-validated performance remained unviable with an R²cv of -0.18.
    Although post-treatment ctDNA status remained broadly prognostic across evaluated endpoints, wide confidence intervals, a limited pool of contributing randomized studies, and suboptimal cross-validation demonstrate that ctDNA clearance cannot yet be adopted as a surrogate efficacy endpoint in clinical trial qualification or patient-level treatment modifications without dedicated prospective validation.

    Key Takeaways

    • Trial-level out-of-sample predictive performance (R²cv) for ctDNA clearance was 0.41 for progression-free survival (apparent R² = 0.77; 95% CI: 0.03–0.96) across 8 palliative/metastatic trials, and 0.26 for overall survival (apparent R² = 0.56; 95% CI: 0.12–0.84).
    • For recurrence-related outcomes, ctDNA clearance demonstrated negative cross-validated performance with an R²cv of -1.66 (apparent R² = 0.04; 95% CI: 0.00–0.92), which remained negative at -0.18 (apparent R² = 0.67) even after excluding the ctDNA-guided DYNAMIC-III trial.
    • Post-treatment ctDNA status retained prognostic value across endpoints but did not demonstrate sufficient trial-level validity to support ctDNA clearance as a surrogate endpoint for regulatory trial qualification or clinical treatment decisions.
    Read the original sourceTranslational cancer research