Malignant melanoma (MM) is an aggressive and fatal tumor caused by the abnormal proliferation of melanocytes (1). In its early stages, MM is typically confined to the skin and presents with nonspecific symptoms, often delaying diagnosis. Advanced MM, however, progresses rapidly, metastasizes early, and has a poor prognosis (2–4). The global incidence of MM has been steadily increasing in recent years (5). Surgery remains the primary treatment for early-stage localized MM, but traditional chemotherapy and radiotherapy have shown limited efficacy in advanced or metastatic cases (6, 7). For instance, dacarbazine, once a standard chemotherapy for advanced MM, has demonstrated low response rates and minimal survival benefits (8). In the 21st century, immunotherapy has revolutionized MM treatment due to its unique susceptibility to immune modulation. ICI have significantly improved OS in patients with advanced MM (9–11). However, 40–60% of patients fail to respond to immunotherapy (12).

Unlike other solid tumors, MM lacks highly specific serum biomarkers, complicating effective monitoring. Traditional tissue biopsies are unsuitable for continuous surveillance. Liquid biopsy, a minimally invasive and highly sensitive approach, has recently gained traction in MM (13, 14). Identifying serum biomarkers with high sensitivity and specificity is critical for improving early diagnosis, real-time disease monitoring, and personalized treatment strategies. PD-L1 expression on tumor cells is currently the most accepted predictive biomarker and is associated with a higher likelihood of treatment response (15). However, PD-L1 expression exhibits significant heterogeneity, and even PD-L1-negative patients can achieve response rates of up to 41.3% (16). This underscores the need for more reliable biomarkers to predict immunotherapy outcomes and minimize unnecessary immune-related adverse effects.

Circulating tumor DNA (ctDNA), consisting of tumor-derived genetic material such as mutations, gene rearrangements, and copy number variations, has emerged as a promising noninvasive biomarker through liquid biopsy (17, 18). Studies have demonstrated the prognostic value of ctDNA in other cancers. For instance, ctDNA detection correlates with poor OS and PFS in metastatic breast cancer (19), poor recurrence-free survival and OS in early-stage breast cancer (20), and promising efficacy and prognosis assessments in locally advanced rectal cancer treated with neoadjuvant chemoradiotherapy (21–23). However, the role of ctDNA in predicting immunotherapy response and prognosis in MM patients treated with ICI remains unclear. To address this, we conducted a systematic meta-analysis of published literature to evaluate the prognostic value of ctDNA in MM patients undergoing ICI therapy, aiming to inform future clinical applications.

Malignant melanoma is a highly invasive malignant skin tumor, commonly occurring in the skin, mucous membranes, and extremities. It is a major cause of skin cancer-related deaths, accounting for approximately 80% of such cases (37–39). In recent years, MM incidence has increased significantly, making it one of the fastest-growing malignancies worldwide (40). In the early stages, surgical excision is the primary treatment. However, once melanoma enters a rapid growth phase, prognosis deteriorates, with high mortality rates, and traditional radiotherapy and chemotherapy have limited effectiveness (41, 42). The advent of ICI has significantly improved treatment and prognosis for metastatic melanoma. Research shows that cytotoxic T-lymphocyte antigen-4 (CTLA-4) inhibitors, the first ICI approved by the U.S. Food and Drug Administration (FDA), can improve survival rates in patients with advanced melanoma (43, 44). However, the durable clinical responses induced by these therapies are observed in only a subset of patients. Therefore, to minimize adverse effects and further improve patient survival, biomarkers are urgently needed to guide clinical treatment decisions.

Circulating tumor DNA, DNA fragments shed into the bloodstream by apoptotic or necrotic tumor cells, has shown promise as a tumor-specific biomarker (45). It is non-invasive, easily accessible, suitable for repeated sampling, and can overcome the challenges of tumor heterogeneity, thus providing a valuable complement to tissue biopsies for clinical diagnosis and disease monitoring. Studies by the Pinzani team (46) revealed that ctDNA testing in MM patients has a sensitivity of 72% and a specificity of 89% compared to tumor tissue detection, with an 80% consistency with pathological results. Additionally, Haselmann et al. (47) found in a study of 187 MM patients that ctDNA assessment could be standardized before tumor recurrence, exhibiting higher specificity than serum S100 and lactate dehydrogenase, suggesting its diagnostic potential. Additionally, during immune checkpoint inhibitor (ICI) therapy, spatial CITE-seq can reveal the activation status of immune cells within the tumor tissue and their interactions with tumor cells (48). Meanwhile, ctDNA can reflect the overall tumor burden and treatment efficacy systemically. The combination of these two approaches offers a more accurate prediction of clinical outcomes in immune therapy. However, factors such as detection technology, sample origin, and disease staging can influence the accuracy of ctDNA testing, and its prognostic value in ICI-treated MM patients remains a subject of ongoing debate.

This study systematically reviewed the relationship between ctDNA and survival rates in ICIs-treated MM patients, analyzing 11 articles covering over 1,000 melanoma cases. Meta-analysis results showed a significant association between ctDNA fluctuations before and after ICIs treatment and the prognosis of MM patients. Compared with patients without detectable ctDNA, those with detectable or higher levels of ctDNA exhibited poorer OS and PFS. Despite the limited number of studies available, ctDNA testing shows potential in clinically assessing OS and PFS risk ratios, suggesting its value in monitoring immune therapy response (49, 50). Furthermore, no significant heterogeneity was observed in the included studies, as indicated by an acceptable I² statistic. Key factors affecting the results included study design, tumor staging, ctDNA measurement timing, and types of prognostic events considered. Although variability exists, the forest plot structure demonstrates stability in the pooled results, supporting ctDNA as an independent prognostic marker for advanced MM patients undergoing ICIs therapy. Another study suggests that ctDNA may help differentiate between true progression and pseudo-progression in MM patients receiving PD-1 antibody therapy. Early monitoring of ctDNA fluctuations could assist clinicians in identifying patients who are not responding to therapy, facilitating timely treatment adjustments and reducing unnecessary costs and ineffective treatment (51, 52). Atsuko et al. indicated that ctDNA could independently reflect the impact of adverse events on tumor burden during ICIs treatment (53). Although this study provides valuable insights into the relationship between ctDNA and prognosis in MM patients receiving ICIs therapy, several limitations must be considered. Firstly, the limited number of studies and relatively small sample sizes may restrict the generalizability of these findings. Furthermore, variability in study design, detection methods, and patient characteristics across studies could introduce bias, affecting the robustness of the results. Additionally, inconsistencies in the timing of ctDNA measurement and threshold definitions may impact the interpretation and comparability of results. Lastly, the short follow-up durations in some studies hinder a comprehensive assessment of the long-term prognostic value of ctDNA. Therefore, future research should focus on developing standardized ctDNA measurement protocols, including unified extraction and detection techniques, threshold definitions, and multicenter validation, to enhance predictive accuracy and reduce methodological variability. Moreover, integrating other omics data (such as genomics, proteomics, and immunomics) with long-term monitoring of ctDNA fluctuations may further improve the accuracy of predicting immune therapy responses. For example, combining spatial multi-omics sequencing technologies, such as DBiT ARP-seq and DBiT CTRP-seq, with multiplex immunofluorescence imaging techniques like CODEX to map immune responses could aid in monitoring disease progression (54). With these improvements, ctDNA has the potential to become a more reliable prognostic biomarker in clinical practice, supporting personalized immune therapy strategies.

This meta-analysis demonstrates a significant association between the presence of ctDNA and the prognosis of MM patients receiving ICI therapy, establishing ctDNA as a specific prognostic biomarker. Additionally, ctDNA shows potential as a tool for stratified diagnosis before treatment and as an effective indicator for monitoring treatment response and disease progression during therapy. Looking forward, we anticipate further high-quality research to provide strong evidence-based support for the clinical application of ctDNA in MM genetic testing.