Quantitative assessment of BRAF V600 mutant circulating cell-free tumor DNA as a tool for therapeutic monitoring in metastatic melanoma patients treated with BRAF/MEK inhibitors

The detection of mutations in circulating cell-free tumor DNA (ctDNA) is under investigation as a specific biomarker for the diagnosis and monitoring of patients with different cancer types [1‐4]. Mutations in the BRAF gene at position V600 are detected in 40–50 % of cutaneous melanomas, and represent the most common oncogenic driver mutation in this disease [5]. Therefore, quantitative measurement of BRAF V600 mutant ctDNA in cell-free DNA (cfDNA) extracted from plasma could serve as a specific biomarker in this patient population [6, 7].

Treatment with a combination of a BRAF- and MEK inhibitor results in a high tumor response rate (64–69 %) and improves the survival of patients with BRAF V600 mutant melanoma [8‐10]. Immune-checkpoint inhibition of either the CTLA-4 or PD-1 receptor can also improve the survival of patients with advanced melanoma, irrespective of the BRAF V600 mutation status [11‐13]. Optimal sequencing of available treatment options for patients with BRAF V600 mutant melanoma has not been defined. Retrospective series have raised concern that ipilimumab may have lesser activity when applied after the emergence of resistance to BRAF-inhibitors [14, 15]. Alternatively, ipilimumab fails to improve the survival of patients with a life-expectancy of less than 3–4 months from initiating therapy, and durable complete responses have been reported on ipilimumab in patients who developed prior resistance to treatment with a BRAF-inhibitor [16]. Of concern is the high incidence of progression within the central nervous system (CNS) at first progression on BRAF-inhibitors, as metastases to the CNS often imply a poor prognosis and necessitate the use of corticotherapy, implying an unfavorable condition for initiating immunotherapy [17, 18]. As conventional clinical tools for assessment of early tumor progression lack sensitivity, many patients will be symptomatic at the time of progression or will experience rapid progression and deterioration in the few weeks that follow the diagnosis of progression [17, 19‐21]. Therefore, more sensitive tools for monitoring of response and resistance to BRAF/MEK targeted therapy is of interest in order to optimize treatment of advanced BRAF V600 mutant melanoma. Furthermore, changes in the BRAF V600mut ctDNA concentration might be helpful for the interpretation of imaging results during immunotherapy where atypical tumor responses are more frequent [22, 23].

In this translational research study we analyze the value of longitudinal quantitative measurement of BRAF V600mut ctDNA from plasma as a therapeutic monitoring tool for patients with advanced BRAF V600 mutant melanoma treated with the BRAF/MEK inhibitors dabrafenib and trametinib.

Patients were eligible for plasma BRAF V600mut ctDNA monitoring when a BRAF V600 mutation had been detected in tumor tissue and were either treated or initiated treatment with dabrafenib and trametinib. Blood samples were prospectively collected after obtaining informed consent with an Ethical Committee approved document. Response evaluation to targeted therapy was performed every 2 months with standard imaging techniques [including 18-fluorodeoxyglucose positron emission tomography/computed tomography (FDG PET–CT), computed tomography (CT) of thorax and abdomen, magnetic resonance imaging (MRI) of the brain]. Plasma samples were collected together with routine blood collections (every 2 weeks during the first month of therapy; every month after the first month of therapy), until progressive disease (PD) was detected according to RECIST v1.1 [24].

Blood samples were collected in 10 mL EDTA tubes and immediately centrifuged at a relative centrifugal force of 1410.63×g, during 15 min at room temperature. Plasma was separated and stored in 1-mL aliquots at −80 °C. Silica-based extraction of DNA and subsequent allele-specific quantitative PCR (qPCR) to detect the BRAF wild-type gene and the G1798>A and T1799>A changes in the BRAF gene were performed with Idylla™ (Biocartis) on 1 mL of the stored plasma. The G1798>A change is present in patients with V600 K, V600R, and V600 M mutations, whereas the T1799>A change is present in patients with V600E, V600 K, V600E2, and V600D mutations. A linear correlation between Cq values reported by prototype Idylla assays and digital droplet PCR was previously established, allowing precise and sensitive quantification of mutant ctDNA fragments in plasma down to 3 mutant copies per PCR reaction with an analytical sensitivity of 0.01 %. The investigators that performed the DNA extraction and the subsequent qPCR were blinded for the clinical patient information. Quantitative PCR results were expressed as median and 25th/75th percentiles of the BRAF V600mut ctDNA copy number per milliliter, and the BRAF V600mut ctDNA fraction to the total amount of cfDNA. The evolution of the BRAF V600mut ctDNA copy number and fraction, were analyzed in relation to the clinical diagnosis of PD. Progression-free survival (PFS) was measured from the time of first administration of targeted therapy, to the time of PD or death, and was analyzed using the Kaplan–Meier method (SPSS software, IBM). For comparison of PFS between different groups, the log-rank test was used. We applied the Wilcoxon signed-rank test to compare changes in the concentrations during treatment. Fisher’s exact test was used to compare categorical variables. A 2-tailed p value ≤0.05 was considered to be statistically significant. Statistical analysis was performed with SPSS statistics 22.

To assess the early evolution of the BRAF V600mut ctDNA concentration during treatment with dabrafenib and trametinib, five samples were collected during the first 24 h of treatment in two additional patients treated “in hospital”. In these two patients, all plasma samples were analyzed in duplicate. The first sample of 1 mL was analyzed with the Idylla™ system using a fully automated cartridge that contains all reagents for the above mentioned extraction of DNA followed by qPCR detection and reporting of the results. In the second sample of 1 mL, DNA was extracted using the QIAamp Circulating Nucleic Acid Kit (Qiagen), followed by droplet digital PCR (ddPCR, Bio-rad) for detection of BRAF V600Emut ctDNA. Each ddPCR analysis was performed in triplicate. Five μL of eluate was added to each ddPCR reaction containing 11.5 μL ddPCR Supermix for probes (Bio-Rad) and 3 μL of BRAF V600E and wild-type assay (dHsaCP2000027, dHsaCP2000028, Bio-Rad), adjusted to a total volume of 23 μL. The reaction was then compartmentalized into ~15,000 droplets per sample in a QX-100 droplet generator (Bio-Rad) according to the manufacturer’s instructions. Emulsified PCR reactions were run in 96-well format on a Bio-rad CFX96 Touch Real-Time PCR Detection System according to the manufacturer’s instructions. Plates were subsequently read on a Bio-Rad QX-100 droplet reader using QuantaSoft software (Bio-Rad). No-template controls were included in each run and the BRAF V600E copies per sample were quantified using the Quantasoft software.

In this study, BRAF V600mut ctDNA was detected at baseline in 75 % of stage IV melanoma patients with a known BRAF V600 mutation. This is in line with two other studies using PCR based methods, where BRAF V600mut ctDNA was detected in plasma of 73–84 % of BRAF V600 mutant melanoma patients [6, 25]. These results are also in line with an overall detection rate of 77 % in 4 clinical studies with 746 stage IV BRAF V600 mutant melanoma patients using BEAMing (beads, emulsions, amplification, and magnetics analysis) after cfDNA extraction from plasma [26‐30]. Moreover, one study showed a 100 % agreement between digital PCR, qPCR and BEAMing, suggesting that the suboptimal detection rate of BRAF V600 mutant DNA in plasma in advanced melanoma is the result of a variable amount of BRAF V600mut ctDNA in advanced melanoma patients, rather than an variability in the sensitivity of different test-platforms [25]. We observed that after initiating BRAF/MEK inhibitor treatment, BRAF V600mut ctDNA can drop below the detection limit, notwithstanding the absence of radiological complete remission. This phenomenon of rapid decrease in the concentration of BRAF V600mut ctDNA has been reported before and has been attributed to the destruction of tumor cells and a subsequent rapid clearance of ctDNA [6, 31‐33]. Our combined observations of (1) decrease of the BRAF V600mut ctDNA concentration within days after treatment initiation, (2) of early increase during disease progression and (3) of early increase after discontinuation of targeted therapy, are suggestive of a correlation between ctDNA levels and the proliferation of melanoma cells. Therefore BRAF V600mut ctDNA seems to reflect the BRAF V600mut-dependent proliferative tumor burden, and not tumor mass as evidenced by CT-imaging. From this perspective, ctDNA could be attributed to increased cell death during melanoma cell proliferation or might be actively secreted or passively released by living cells, rather than by apoptotic or necrotic cells [34]. This latter hypothesis is supported by several preclinical arguments. Spontaneous active DNA release or apoptosis have been put forward as the main source of ctDNA, because plasma DNA often presents a ladder pattern, typical for active cleaving, when subjected to electrophoresis [32, 35]. Several arguments support the hypothesis that DNA is primarily released by living cancer cells rather than by apoptotic cancer cells: (1) DNA concentration increases in normal lymphocyte cultures following stimulation with phytohemagglutinin, lipopolysaccharide or antigen; (2) in a leukemic cell-line malignant cells released newly synthesized DNA; (3) cancer cell DNA concentration in cell culture supernatant increases with cell proliferation when few apoptotic or necrotic cells are present; and (4) circulating exosomal BRAF V600E mutant DNA was isolated from SK-MEL-28 melanoma-bearing mice [32, 33, 36, 37].

In our study population, PFS was significantly better for the patients in whom BRAF V600mut ctDNA became undetectable after the initiation of targeted therapy. When the BRAF V600mut ctDNA fraction increased during treatment, clinical PD always followed within 2 months. Moreover, in 44 % of cases an increase in the mutant fraction preceded PD. We recently reported that reappearance of BRAF V600mut ctDNA in plasma cfDNA can precede changes on PET–CT in patients under BRAF/MEK inhibitor treatment [34]. In this study an increase in the BRAF V600mut ctDNA fraction preceded PD with a median interval of 25 days. However, plasma samples were collected on a monthly basis, whereas imaging was performed every 2 months. In a study on ctDNA monitoring in metastatic breast cancer, increasing levels of ctDNA appeared on average 5 months before the establishment of PD by means of imaging [3].

An increase in the BRAF V600mut ctDNA fraction during BRAF/MEK targeted therapy could potentially serve as a trigger for early evaluation with imaging techniques, and allow for a timely switch to immunotherapy or other appropriate therapeutic interventions (e.g. stereotactic radiotherapy to brain metastases). Early detection of progression is important in metastatic melanoma, because of the aggressive course of the disease as exemplified by the observation that the central nervous system is a frequent site (30 %) of first progression under BRAF inhibitor therapy [17]. Moreover, potential second line therapeutic options such as ipilimumab and, to a lesser extent anti-PD-1 therapies, are associated with a latency of their anti-tumor effect. Therefore, the possibility of offering immunotherapy may be compromised if PD is only detected on imaging or when clinical symptoms are present [19, 20].

Our results indicate that quantitative analysis of BRAF V600mut ctDNA in plasma holds promise as a monitoring tool in BRAF V600 mutant melanoma during treatment with BRAF/MEK targeted therapy. Prospective trials are needed to assess if an add-on strategy or switch to immunotherapy upon early detection of therapy resistance through ctDNA evolution, adds value to the current treatment strategies. Finally we provide clinical arguments suggesting that an important fraction of BRAF V600mut ctDNA correlates with the BRAF V600mut-dependent proliferative tumor burden.