ctDNA Shows Prognostic Value in Resectable Melanoma, but Sensitivity Gaps Limit Clinical Utility

Detectable circulating tumor DNA (ctDNA) after curative-intent surgery was consistently associated with significantly worse recurrence-free, distant metastasis–free, and overall survival (OS) across multiple time points in patients with resectable melanoma. However, high false-negative rates may limit its use in surveillance or treatment de-escalation, according to results of a study recently published in Cancer Treatment Reviews.¹

Lack of Prognostic Biomarkers Remains a Challenge

Melanoma is the leading cause of skin cancer–related deaths globally, with rising incidence reported worldwide. Although patients with early-stage disease are frequently cured through surgical resection, those with stage IIB through resected stage IV melanoma remain at high risk of recurrence and are candidates for adjuvant systemic therapy.² However, approximately 50% of these patients will relapse despite treatment, and therapies carry the risk of long-term toxicity.¹

The researchers highlighted that no adjuvant systemic therapy trial has demonstrated a major OS benefit in this population, and clinicians lack reliable biomarkers to guide treatment decisions. Although widely used, the tumor-node-metastasis–based staging system is limited by tumor heterogeneity and offers incomplete prognostic resolution at the individual patient level.

Meanwhile, ctDNA, a subset of cell-free DNA shed by tumors into the bloodstream, has emerged as a promising noninvasive risk stratification tool. Detection of ctDNA following curative-intent surgery, defined as minimal residual disease (MRD), has been associated with recurrence and worse outcomes across multiple solid tumor types.

However, the researchers noted that evidence in melanoma has been fragmented, hindered by low sensitivity, assay variability, and a lack of robust, prospective data. To fill this knowledge gap, the researchers conducted a systematic review and meta-analysis, searching PubMed, Embase, and Cochrane databases through July 2025.

ctDNA was assessed at 3 predefined time points: baseline (presurgery), landmark (within 12 weeks post surgery), and longitudinal (beyond 12 weeks post surgery). Prognostic outcomes were pooled using random-effects models, and diagnostic performance was assessed via univariate and bivariate models.

ctDNA Detection Associated With Poor Survival Outcomes

From 3131 initial records, 21 studies comprising 24 cohorts and 3032 patients with available ctDNA data were included: 6 phase 2 and 3 trials, 11 prospective studies, and 4 retrospective studies.¹

The cohort had a median age of 60 years (range, 18-93), 58.4% were male, and approximately 97% had high-risk melanoma (stage IIB to IV resected). Among the patients, approximately 65% received adjuvant therapy, 30% received placebo or observation, and 2% received neoadjuvant therapy. Additionally, the median follow-up was 38.5 months.

ctDNA detection rates varied by time point: 38% at baseline (95% CI, 34%-43%); 15% at landmark (95% CI, 12%-19%); and 22% at longitudinal assessment (95% CI, 17%-28%). Across all 3 time points, detectable ctDNA was consistently associated with worse relapse-free survival. Baseline ctDNA positivity conferred an HR of 3.14 (95% CI, 1.82-5.41), rising to 3.42 (95% CI, 2.51-4.65) at the landmark time point and 3.76 (95% CI, 2.83-4.99) during longitudinal surveillance, the strongest prognostic association observed.

For distant metastasis-free survival, baseline ctDNA detection was associated with an HR of 2.25 (95% CI, 1.50-3.36), which strengthened substantially at the landmark time point to 5.38 (95% CI, 3.09-9.37). However, longitudinal distant metastasis-free survival data were available from only 2 small studies and did not reach statistical significance.

Regarding OS, landmark ctDNA positivity, evaluated in 6 cohorts comprising 907 patients, was associated with an HR of 3.55 (95% CI, 2.63-4.79). The baseline OS analysis did not reach significance, likely owing to limited data from 2 studies comprising 77 patients.

Lastly, on the diagnostic side, ctDNA demonstrated high specificity, particularly in the MRD setting, but modest sensitivity. At the landmark time point, pooled sensitivity was 31.0% (95% CI, 22.5%-41.1%) and specificity was 94.3% (95% CI, 87.1%-99.0%). Longitudinal assessment yielded sensitivity of 45.2% (95% CI, 34.3%-56.7%) and specificity of 96.3% (95% CI, 92.5%-99.0%).

Although longitudinal testing produced numerically higher diagnostic ORs and area under the curve values than landmark testing, a direct comparison among the 15 cohorts reporting both time points did not show a statistically significant difference in overall diagnostic accuracy (P = .106).

High Specificity, Limited Sensitivity Constrain ctDNA’s Clinical Utility

The researchers acknowledged several study limitations, including substantial heterogeneity among ctDNA assay platforms and cutoff thresholds that may limit the generalizability of their findings. Also, despite efforts to harmonize collection time points, considerable variability persisted. Nevertheless, the researchers expressed confidence in their findings.

“…Our meta-analysis revealed that [patients with] resectable melanoma with detectable plasma ctDNA exhibited higher recurrence rates and worse survival outcomes,” the authors concluded. “Diagnostic metrics revealed that plasma ctDNA has high specificity and thus can risk-stratify patients with high-risk melanoma and inform trials investigating treatment intensification strategies. However, limited analytical sensitivity precludes its use for surveillance or de-escalation strategies.”

References

1. Abrahão Reis PC, Noronha MM, Oliveira JP, et al. Circulating tumor DNA for predicting recurrence and mortality in patients with resected melanoma: a systematic review and meta-analysis. Cancer Treat Rev. Published online May 5, 2026. doi:10.1016/j.ctrv.2026.103146
2. Gershenwald JE, Scolyer RA, Hess KR, et al; for members of the American Joint Committee on Cancer Melanoma Expert Panel and the International Melanoma Database and Discovery Platform. Melanoma staging: evidence-based changes in the American Joint Committee on Cancer eighth edition cancer staging manual. CA Cancer J Clin. 2017;67(6):472-492. doi:10.3322/caac.21409