Melanoma is the deadliest form of skin cancer, with an incidence that has continued to increase over the past several decades (1). The introduction of novel treatments, however, has contributed to the substantial improvement in patient outcomes, with an approximately 5% decline in the mortality rate since 2013 (2).

With the integration of paradigm-shifting therapies, including immune checkpoint inhibitors (ICIs) and BRAF-MEK inhibitors, clinical outcomes have improved significantly for patients with advanced melanoma. However, several challenges have also emerged. Pseudoprogression is a well-recognized radiological phenomenon that can occur during treatment with ICI, due to immune cell infiltration in the tumor microenvironment. This phenomenon has been observed to occur in up to one-third of melanoma patients treated with ICI (3). It is difficult to distinguish between true progression and pseudoprogression using standard-of-care imaging, and misinterpretation of radiological findings can adversely affect optimal clinical decision-making. New assessment tools have been developed to evaluate response to ICI (iRECIST), but their use remains limited in the context of clinical trials. Moreover, utilizing imaging as a single modality for surveillance has inherent disadvantages, such as the lack of ability to capture early response to therapy, the early stages of progression, or to monitor disease status over short intervals (4). As such, the identification and validation of novel biomarkers that can accurately monitor disease status and assess treatment efficacy in real time is critically needed.

Circulating tumor DNA (ctDNA) has emerged as a biomarker that has been shown to accurately reflect disease burden in real time. Testing relies on the detection of tumor-related DNA fragments (140–170 base pairs long) in plasma collected from peripheral blood (5). It has been used successfully in many cancer indications for disease screening, disease monitoring, prognostication, and treatment evaluation (6). Multiple platforms have been developed to analyze ctDNA in different cancers including single mutation detection by droplet-digital PCR (ddPCR), next-generation sequencing (NGS), or whole-exome sequencing (WES), which have been reviewed previously (7, 8). Existing evidence on the clinical utility of ctDNA testing in melanoma has emerged recently in retrospective and prospective studies (9, 10). The use of personalized and tumor-informed ctDNA testing has been shown in several studies to detect molecular residual disease (MRD), and serve as an important prognostic tool to predict relapse in colorectal, breast, lung and urothelial cancers (11–14).

Here, we present a case series of four patients with metastatic melanoma treated with different modalities, demonstrating our experience using personalized and tumor-informed testing for longitudinal monitoring of ctDNA, and how ctDNA results could inform clinical decision-making. We also discuss the challenges of using such an assay in the management of patients with metastatic melanoma.

The emergence of data establishing ctDNA as a predictive and prognostic biomarker to monitor disease status in multiple indications represents a milestone for cancer management. Several platforms have been developed to monitor ctDNA and have been validated in melanoma patients treated with targeted therapy and immunotherapy (15, 16). However, several challenges exist that may hinder the reliability of these tests in clinical practice (17). Technical factors in each specific ctDNA assay are subject to false negativity due to a low amount of tumor DNA allele fraction in isolated plasma as well as low ctDNA shedding from the tumor (17). False positivity represents another challenge that hinders the accurate interpretation of ctDNA, which could arise due to the detection of clonal hematopoiesis of indeterminate significance (CHIP) that is not tumor-specific (18). To this end, personalized, tumor-informed, patient-specific assays of both tumor tissue and matched normal blood samples allow for filtering of CHIP variants, thereby providing an accurate assessment of patient’s tumor burden. The initial analysis of such ctDNA approach is promising and has been validated in several cancer studies, showing high sensitivity ranging from 88% to 100% (11–14). Longitudinal analysis of ctDNA, as measured by ddPCR, was found to be predictive of disease relapse and response to adjuvant treatment in the absence of radiological findings in melanoma (9, 19). Similarly, ctDNA monitoring of melanoma-associated somatic alterations was shown in a cohort of patients to predict disease progression ahead of radiological findings (10). Of note, the majority of the published studies in melanoma have used single mutation or limited targeted panel sequencing for ctDNA analysis (BRAF, NRAS, KIT, and TERT) (16, 20, 21). This is of importance, as targeted panel sequencing detects a single, or a limited number of mutations, and its predictive and prognostic value is subject to tumor heterogeneity, as these tests are not able to capture subclonal mutations and new acquired mutations arising during treatment (22, 23). In addition, some mutations common to metastatic melanoma, such as mutations in the BRAF gene, can also be detected in patients with benign nevi syndromes (24). Another limitation is the subclonal architecture that could arise during treatment, which cannot be detected with targeted ctDNA panels. These limitations should be considered when interpreting results from panel-based ctDNA assays. Therefore, guidelines do not recommend using ctDNA outside of clinical trials or research given the lack of evidence regarding their utility and validity (25). Of interest, utilizing ctDNA as a supplementary method in the context of clinical, laboratory, and radiological findings might prove beneficial and is an active area of research.

Our experience in this patient cohort demonstrates the clinical utility of ctDNA testing in monitoring response of melanoma to therapy. The ability for ctDNA to detect tumor response to ICI early can aid in distinguishing true progression vs. pseudoprogression. In case 1, ctDNA was used as an adjunctive tool to imaging, to monitor disease response to the therapy. The patient had undetectable levels of ctDNA during treatment with ICI, despite having an abnormal PET-CT scan. The combined approach of short-term imaging with longitudinal ctDNA monitoring helped postpone unnecessary therapeutic intervention, as the FDG avid lesion improved on serial imaging. Lee et al. similarly found that in a cohort of 125 melanoma patients, ctDNA monitoring was able to help differentiate true progression from pseudoprogression during treatment with programmed death-1 (PD-1) inhibitors (26). Therefore, ctDNA testing can potentially reduce the need for unnecessary and invasive procedures normally employed when pseudoprogression is suspected, including repeat biopsy, radiation therapy, and change of treatment (26). The reliability of such an approach for clinical practice will depend on further validation of ctDNA assays in metastatic melanoma, as different detection methods have different sensitivities.

In case 2, we observed a significant decrease in ctDNA levels after starting imatinib, in relapsed metastatic melanoma. The decline in ctDNA occurred prior to the radiological evidence of tumor response to imatinib. In line with this finding, several studies in melanoma demonstrated that ctDNA clearance can precede radiological response during treatment with ICI and targeted therapy (15, 27–32). Case 2 is unique in several ways. Currently, only 30% of relapsed-refractory melanomas respond to KIT inhibitors (1). This case is in concordance with a prior case report demonstrating a possible role for ctDNA in monitoring KIT-mutant melanoma response to treatment (31).

Longitudinal monitoring of ctDNA can help clarify radiographic concerns for disease progression that cannot be confirmed by tissue biopsy. This is exemplified in case 3, where elevated ctDNA levels were concordant with a questionable FDG-avid aortocaval lymph node. Interestingly, ctDNA levels increased significantly after starting SBRT for the target lesion prior to ctDNA clearance, consistent with reports that trauma and necrosis due to radiation of tumor cells can lead to increased shedding of ctDNA into the peripheral blood (33, 34). The rise in ctDNA levels prior to clearance confirmed our suspicion that the lesion was secondary to disease progression.

Finally, case 4 highlights the limited ability of ctDNA in metastatic melanoma to monitor response to treatment and disease progression. Immune-privileged sites such as the brain, eye, or testes have highly developed blood–organ barriers that may limit the shedding of ctDNA into the peripheral bloodstream. Several studies have previously shown a poor correlation between ctDNA and intracranial disease status in melanoma (35, 36). In case 4, the patient continued to have undetectable ctDNA levels, despite continued progression of brain metastases. This observation is consistent with the theory that the blood–brain barrier prevents ctDNA shedding. Therefore, ctDNA testing may not be appropriate for monitoring intracranial metastases. Some reports have suggested the potential benefit of using CSF analysis for ctDNA monitoring in this setting. However, this procedure is highly invasive, and needs further validation (37–40).

In summary, our case series highlights the potential clinical utility and limitations of using patient-informed, tumor-specific ctDNA assay for clinical decision-making in the treatment of metastatic melanoma patients in the real-world setting. Analysis of ctDNA might be beneficial in disease monitoring as a complementary tool added to standard-of-care surveillance methods such as physical examination and imaging. Further validation of the utility of ctDNA in larger melanoma patient cohorts remains essential prior to complete integration of ctDNA testing in the clinical setting.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

Because this work is a case series reporting clinically conducted activities, this activity is not considered to meet federal definitions under the jurisdiction of an IRB and therefore falls outside the purview of the HRPO. IRB ID: 202210030.

KK and GA conceptualized the idea of the manuscript. KK and AZ performed literature search and wrote the manuscript. KK, GB, and AK conceptualized figures. GB and AK created figures. KK, AZ, OB, GB, AK, and GA reviewed the manuscript and agreed on the final version. All authors contributed to the article and approved the submitted version.

The authors would like to thank Meenakshi Malhotra Ph.D. for reviewing this work.

GB is an employee of Natera and AK is a previous employee of Natera, Inc. Both own stocks in the company.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.