Molecular profiling with next-generation sequencing (NGS) has been applied in multiple solid cancers to discover potential therapeutic targets. Here, we describe the results of a clinical NGS panel in patients with advanced melanoma. Thirty-six tumor tissues from patients with BRAF wild-type melanoma at Seoul National University Hospital (SNUH; Seoul, Republic of Korea) were collected and deep-sequenced using the SNUH FIRST-Cancer NGS panel to assess single-nucleotide variants, small insertions/deletions, copy number variations, and structural variations to estimate tumor mutation burden (TMB). We discovered 106 oncogenic alterations and most of the patients (n = 33, 92%) harbored at least one oncogenic alteration, including 2 patients who were initially diagnosed as BRAF V600E–negative but were later confirmed to be positive. Altogether, 36 samples were classified into RAS/BRAF/NF1–mutant (n = 14, 39%) or triple wild-type (n = 22, 61%) melanoma subtypes. The estimated median TMB was 8.2 mutations per Mb, ranging from 0 to 146.67 mutations per Mb. Of the 36 patients, 25 (70%) had actionable alterations with currently developed drugs, and 7 (19.4%) were enrolled in clinical trials with an RAF inhibitor, multiple receptor tyrosine kinase inhibitor, and anti-programmed cell death-1 (PD-1) antibody. TMB tended to associate with progression-free survival (PFS) of treatment with anti-PD-1/PDL-1 antibody (HR, 0.96; 95% confidence interval, 0.92–1.00; P = 0.07). High-TMB (≥13) group was associated with longer PFS than the low-TMB group (median 34.0 vs. 11.0 weeks, P = 0.04). Overall, the clinical use of a NGS panel in patients with advanced melanoma shows association with clinical outcomes and several therapeutic strategies.

Advanced malignant melanoma is an aggressive and incurable disease with limited expected survival (1). However, the introduction of immune checkpoint inhibitor ipilimumab and anti-programmed cell death-1 (PD-1) antibody therapeutics expanded the median survival by as much as 3 years, and those who responded to the immunotherapy are expected to benefit even more than the median survival (2, 3). Moreover, the discovery of several molecular pathways and the development of drugs targeting those pathways have meaningful clinical benefits for patients with melanoma (4–6). Combination treatment with BRAF kinase inhibitor and MEK inhibitor for previously untreated advanced melanoma with BRAF V600E mutation showed a 1-year overall survival rate of 72%∼75.5% and objective response rate (ORR) of 64%∼68% (7–9). Imatinib, the KIT inhibitor, produced ORR of 54% for patients with KIT-mutant melanoma in a phase II trial (10). In the meantime, genomic analyses of melanoma had revealed not only genetic evolutions of key alterations in melanoma but also mutation landscapes defining disparate genomic subtypes, providing clinicians with various insights on pathogenesis and clinical features (11, 12). These encouraging results led to multiple therapeutic options for advanced melanoma and increased initiation of clinical trials with the new agents. As a result, there are increasing demands to evaluate multiple genomic profiles of melanoma efficiently.

Next-generation sequencing (NGS) is a prevalent technique that has been widely used in various cancer genomics researches (13). With the development and validation of various pipelines, NGS can identify multiple target alterations effectively with reduced clinical reporting time (13). Now the NGS is becoming to take a key role in current precision oncology and enables clinicians to identify even minor actionable targets in patients with cancer and to select those patients for appropriate clinical trials in ways such as umbrella trials and basket trials (14). Furthermore, recent reports show tumor mutation burden (TMB) calculated by an comprehensive targeted genomic profiling correlates with the TMB calculated by whole-exome sequencing, which typically takes far more resources and processes than the NGS-based panel (15). This implicates that NGS-based panel may be used to predict tumor response to anti-PD-1 antibodies with TMB (15–18). Altogether, an NGS-based panel harbors significant advantage as a tool for deciding the treatment strategies for advanced melanoma. Consequently, various NGS-based platforms have been developed for melanoma and described in earlier literature (19, 20).

There exist unmet clinical needs for such comprehensive genomic profiling of advanced melanoma especially in patients without BRAF mutation. For example, the demographics of patients with melanoma in the Korean population are quite different from those in the Western population. The incidence of melanoma is lower in the Korean population than in the Western population (21). The most common histologic subtype is not superficial spreading melanoma but acral lentiginous melanoma in the Korean population (22). Furthermore, BRAF-activating mutation frequency observed is 15% in the Korean patients with melanoma, which is lower compared with a frequency of 50% in the Western population (23, 24). As seen in Korean population, clinical application of NGS panel is essential to provide additional treatment options for these patients with BRAF wild-type melanoma.

In Seoul National University Hospital (Seoul, Republic of Korea), we used an NGS panel named SNUH FIRST-Cancer panel, which assesses 225 target genes for genomic profiling of melanoma. The panel includes not only most of the genes known to be associated with melanoma (BRAF, NRAS, NF1, and KIT; ref. 12) but also genes that are potentially actionable with currently developed drugs or new drugs under clinical trials. The panel is also capable of estimating TMB, which may be used to predict response to anti-PD-1/PD-L1 antibody. This study aims to describe the results of NGS panel in patients with BRAF wild-type melanoma along with clinical outcomes to evaluate their potential clinical applicability.

Sample collection

Patients with recurred or metastatic melanoma who visited Seoul National University Hospital (Seoul, Republic of Korea) and who were initially BRAF V600E–negative by either IHC or Sanger sequencing were enrolled in the study. DNA from the archival formalin-fixed paraffin-embedded (FFPE) tumor tissues were collected and profiled using NGS. Demographic features at the time of tissue acquisition were reviewed. Treatment history, response, and survival data were also collected. This study was approved by institutional review board (H-1803-138-933) and in accordance with the declaration of Helsinki.

DNA sequencing

Genomic DNA was extracted from FFPE tissues using the ReliaPrepTM FFPE gDNA Miniprep System (Promega). Approximately 250 to 500 ng of DNA was fragmented by Covaris Sonicator (Covaris) and the amount of fragmented DNA was measured by qPCR. The libraries were constructed using Agilent's SureSelectXT Target Enrichment protocol for Illumina paired-end sequencing. The quality and quantity of libraries were assessed by Bioanalyzer 2100 and DNA 1000 Chips (Agilent Technology). Subsequently, paired-end sequencing (2 × 101bp) was performed on the Illumina Hiseq 2500 Platform (Illumina Inc.).

Sequencing data analysis

Technical sequences (adaptors) and low-quality reads were trimmed from the raw FASTQ data by Trimmomatic software (25). The cleaned reads were aligned to the human reference genome (GRCh 37) by BWA v0.7.12 aligner (26) and preprocessed following GATK best practices.

We detected genetic alterations and estimated TMB (mutation/Mb) in patients with melanoma from deep sequencing data of custom-targeted panel and three versions of SNUH FIRST-Cancer Panel (v2, v3, and v3.1) including 225 cancer-associated genes (Supplementary Table S1). Single-nucleotide variants (SNV) were called by an in-house script to combine MuTect v1.1.7 (27) and OxoG filter algorithm (28). Small insertions and deletions (INDEL) were called by IndelGenotyper v36.3336. For functional analysis, confident mutations were annotated by ANNOVAR (29) and further filtered out to select rare nonsynonymous mutations only. Copy number variants (CNV) were detected by Z-score–based method and CNVkit software (30). Structural variants (SV) were called by DELLY (31) and manually reviewed with the Integrative Genomic Viewer (32, 33). We estimated TMB based on predicted somatic mutations by the modification of a method by Chalmers and colleagues (15). Details of the pipeline used to identify genomic alterations and predict TMB are described in Supplementary Methods.

Oncogenic/actionable alterations

We focused on genomic alterations with impact on the treatment plan and tumor biology using clinicians' knowledge and public database OncoKB (34) along with SNVs, CNVs, and SVs reported by SNUH FIRST-Cancer panel assay. The selected genomic alterations were separated into two classes: (i) oncogenic alterations evaluated as “oncogenic” or “likely oncogenic” in OncoKB and (ii) actionable alterations defined as oncogenic alterations to predict response or resistance against treatment. To investigate the feasibility of targeted panel sequencing for melanoma, properties of oncogenic alterations in this cohort were assessed on genomic subtypes of The Cancer Genome Atlas (TCGA; ref. 35); we also observed the incidence of oncogenic alterations in key genes and pathways that were reported recently from a large-scale landscape study (12).

Treatment response

We evaluated treatment responses using RECIST criteria v1.1 (36). ORR was defined as the percentage of patients with complete response (CR) and partial response (PR). Clinical benefit rate (CBR) was defined as the percentage of patients with CR, PR, and stable disease (SD) for 6 months. Progression-free survival (PFS) was defined as the time elapsed from treatment initiation and disease progression or death. Analysis of PFS was by the Kaplan–Meier method and the HR was calculated using the Cox-proportional hazard model. Comparison of continuous values between groups was with the Mann–Whitney test. To analyze the response to an anti-PD-1/PD-L1 antibody according to estimated TMB, patients were divided into two groups; high-TMB if estimated TMB is ≥13, and low-TMB if estimated TMB is <13. A P value below 0.05 was considered statistically significant. All statistical analyses were performed with R 3.4.3 software (https://cran.r-project.org/bin/windows/base/old/3.4.3/).

Patients

Tumor tissues were collected from 36 patients with recurred or metastatic melanoma who were candidates for systemic treatment. The primary sites of melanoma were mainly cutaneous lesions (n = 22, 61%), 17 of which were acral (77.3%). A majority of tissues were obtained from noncutaneous sites (n = 26, 72.2%) including the liver, lymph node, lung, bone, mucosa of the conjunctiva, nasal cavity, gastrointestinal, and genital tract. Thirty-one patients underwent surgical measures to obtain biopsy samples while the other 5 patients received needle biopsy. Demographic features of 36 patients with melanoma are summarized in Table 1.

Table 1.

Demographic features of patients and tissues.

DemographicsN (%)
Age 41–83 (median, 63) 
Gender 
 Male 18 (50) 
 Female 18 (50) 
Primary site 
 Acral 17 (47.2) 
 Mucosal 10 (27.8) 
 Cutaneous 5 (13.9) 
 Othersa 3 (8.3) 
 Unknown 1 (2.8) 
Stage 
 Locally advancedb 3 (8.3) 
 Metastatic 33 (91.7) 
Biopsy site characteristic 
 Skin 10 (27.8) 
 Non-skinc 26 (72.2) 
Biopsy location type 
 Primary 16 (44.4) 
 Metastatic 20 (55.6) 
Tumor fraction 
 ≥80% 30 (83.3) 
 <80% 6 (16.7)d 
Line of treatment for anti-PD-1/anti-PD-L1 
 1 15 
 2 10 
 ≥3 
 Not applicable 
DemographicsN (%)
Age 41–83 (median, 63) 
Gender 
 Male 18 (50) 
 Female 18 (50) 
Primary site 
 Acral 17 (47.2) 
 Mucosal 10 (27.8) 
 Cutaneous 5 (13.9) 
 Othersa 3 (8.3) 
 Unknown 1 (2.8) 
Stage 
 Locally advancedb 3 (8.3) 
 Metastatic 33 (91.7) 
Biopsy site characteristic 
 Skin 10 (27.8) 
 Non-skinc 26 (72.2) 
Biopsy location type 
 Primary 16 (44.4) 
 Metastatic 20 (55.6) 
Tumor fraction 
 ≥80% 30 (83.3) 
 <80% 6 (16.7)d 
Line of treatment for anti-PD-1/anti-PD-L1 
 1 15 
 2 10 
 ≥3 
 Not applicable 

aOther primary sites include dura, testicle, and conjunctiva.

bDefinite staging was not available in the 3 patients because primary sites were atypical (dura, conjunctiva, and rectum) without definite distant metastasis. All 3 patients experienced local recur.

cNon-skin biopsy sites include liver, lymph node, lung, bone, mucosa of conjunctiva, nasal cavity, gastrointestinal, and genital tract.

dSamples of 2 patients had tumor fraction of less than 50%.

Distribution of oncogenic alterations

The SNUH FIRST-Cancer panel sequencing was performed on 36 collected tumors to an average median read depth of 539 × (range 164–1138 ×). The panel assay detected 332 SNVs, 13 INDELs, 73 CNV genes, and 4 SVs. We evaluated the effect of these alterations and considered 106 oncogenic alterations including 46 SNVs, 4 INDELs, and 56 CNV genes for further analysis (Supplementary Table S2 and S3). We identified TCGA subtypes classified as RAS-mutated, BRAF-mutated, NF1-mutated, and triple-wild–type (WT) in the Korean patients with advanced melanoma and analyzed the distribution of oncogenic alterations across those subtypes (Fig. 1A-D; ref. 35).

Figure 1.

Distribution of oncogenic alterations in melanoma patient samples. A, Each column represents individual patient sample. The colors corresponding to the alterations are annotated at the bottom. B, Frequency of alterations in recurrently altered genes among patients (>5%). Tumor suppressor genes are indicated by asterisks. C, Frequency of altered oncogenic pathways in patients (>10%). D, Proportion of patients with melanoma altered in oncogenic pathways.

Figure 1.

Distribution of oncogenic alterations in melanoma patient samples. A, Each column represents individual patient sample. The colors corresponding to the alterations are annotated at the bottom. B, Frequency of alterations in recurrently altered genes among patients (>5%). Tumor suppressor genes are indicated by asterisks. C, Frequency of altered oncogenic pathways in patients (>10%). D, Proportion of patients with melanoma altered in oncogenic pathways.

Close modal

Of the 36 patients with melanoma, 22% (n = 8) harbored NRAS mutations including Q61R/K/L (n = 4), G12D/S (n = 2), and G13D/R (n = 2). In another RAS family, HRAS Q61L mutation was observed in 2 patients. NF1 was one of the most frequently mutated tumor suppressor genes (n = 4, 11%) along with TP53 in this cohort (Fig. 1B). NF1 mutations included three stop-gain SNVs and one missense SNV (p.Y489C) known to be associated with aberrant splicing. KIT mutations were found only and most frequently (n = 4, 11%) in triple-WT subtypes. Although we excluded BRAF-positive samples in tumor collection step, SNUH FIRST-Cancer panel detected BRAF V600E mutation in 2 patients that had been previously diagnosed as negative for BRAF V600E mutation by IHC and Sanger sequencing. Variant allele frequencies of BRAF V600E in the 2 patients were 7.6% and 7.8%, respectively. One of the patients diagnosed as BRAF V600E–negative by Sanger sequencing was confirmed positive for BRAF V600E mutation with additional validation experiments using peptide nucleic acid (PNA)- mediated PCR clamping method.

In CNVs, 38 amplified genes and 18 deleted genes were shown to be oncogenic alterations in 20 patients. Cyclin-dependent kinase inhibitor 2A gene (CDKN2A, n = 8, 22%) and the adjacent gene CDKN2B (n = 7, 19%) were frequently deleted across all subtypes. Triple-WT–subtyped patients had a significantly higher frequency of amplified genes compared with the rest of the patients (P < 0.01). Half of the triple-WT subtypes harbored at least one copy number amplification of a gene, which included not only TERT amplification (n = 6), known to frequently occur in melanoma but also emerging therapeutic targets such as amplification of KRAS (n = 2) and KDR (n = 2).

Overall, all patients except three (92%) harbored at least one oncogenic alteration in key genes of melanoma reported by the recent landscape study (12). In pathway analyses of those genes, we found that MAPK, cell-cycle, receptor tyrosine kinases (RTK), p53 signaling, and PI3K pathways were altered in 53%, 39%, 22%, 17%, and 11% of tumors, respectively (Fig. 1C). Furthermore, about half of the patients (47%) had alterations in multiple pathways involved in the development of melanoma (Fig. 1D).

Estimated TMB

Overall, the median estimated TMB was 8.2 mutations per Mb, ranging from 0 to 146.67 mutations/Mb. NF1-mutated patients showed dominance in higher estimated TMB, with 3 patients exhibiting TMB over 40 mutations/Mb including the highest TMB among the whole samples (Fig. 1A). Patients with the second highest estimated TMB (74 mutations/Mb) harbored MSH6 T955fs mutation, which is known to affect DNA mismatch-repair (MMR; ref. 37).

Clinical application of NGS results

Twenty-five patients with melanoma (70%) had at least one actionable alteration known to serve as a biomarker for the standard of care or as a potential target for newer investigational drugs (Table 2). Seven patients were enrolled in the clinical trial based on the profiling results from NGS. Data on target gene alterations and administered medications in these patients are summarized in Table 3. Among the 7 patients, 4 had SD and PD was seen in the other 2 patients. The response of 1 patient was not evaluable because the patient died 2 weeks after initiation of trial medication. Median PFS was 12.6 weeks (range 1.9–100.4 weeks, Supplementary Fig. S1). Patient 5 harboring NRAS Q61K mutation had early metabolic response, as revealed by PET-CT images, after 1-month treatment with the RAF inhibitor (Fig. 2).

Table 2.

Actionable alterations in 36 patients with melanoma.

PathwayGeneAlteration# Cases
MEK BRAF V600E 
(n = 16, 44%) NRAS Q61R 
 NRAS G12D 
 NRAS G12S 
 NRAS G13D 
 NRAS G13R 
 NRAS Q61K 
 NRAS Q61L 
 HRAS Q61L 
 KIT D816V 
 KIT K642E 
 KIT K693N 
 MAP2K1 R49H 
 NF1 Q1499X 
 NF1 Q2213X 
 NF1 Q236X 
 NF1 Y489C 
 KRAS Amplification 
 MET Amplification 
Cell cycle CDKN2A R80X 
(n = 11, 30%) CDK4 Amplification 
 CDKN2A Deletion 
RTKs FGFR2 E619K 
(n = 8, 22%) KDR Amplification 
 PDGFRA Amplification 
PI3K/mTOR PIK3CA H1047R 
(n = 3, 8%) PTEN M1R 
 PTEN R173C 
 PTEN Splicing 
p53 (n = 2, 6%) MDM2 Amplification 
DNA repair (n = 1, 3%) MSH6 T955fs 
PathwayGeneAlteration# Cases
MEK BRAF V600E 
(n = 16, 44%) NRAS Q61R 
 NRAS G12D 
 NRAS G12S 
 NRAS G13D 
 NRAS G13R 
 NRAS Q61K 
 NRAS Q61L 
 HRAS Q61L 
 KIT D816V 
 KIT K642E 
 KIT K693N 
 MAP2K1 R49H 
 NF1 Q1499X 
 NF1 Q2213X 
 NF1 Q236X 
 NF1 Y489C 
 KRAS Amplification 
 MET Amplification 
Cell cycle CDKN2A R80X 
(n = 11, 30%) CDK4 Amplification 
 CDKN2A Deletion 
RTKs FGFR2 E619K 
(n = 8, 22%) KDR Amplification 
 PDGFRA Amplification 
PI3K/mTOR PIK3CA H1047R 
(n = 3, 8%) PTEN M1R 
 PTEN R173C 
 PTEN Splicing 
p53 (n = 2, 6%) MDM2 Amplification 
DNA repair (n = 1, 3%) MSH6 T955fs 
Table 3.

Genomic profiles and clinical information of patients who were enrolled into clinical trial.

IndexGeneAA changeAllele frequency (%)Copy numberTMBDrugs in clinical trialsBest responsePFS (weeks)
Patient 1 KRAS Amplification — 14 4.0 RAF inhibitor NE 1.9 
Patient 2 NRAS G12D 88.8 — 4.9 RAF inhibitor PD 3.0 
Patient 3 NRAS G13D 89.2 — 2.0 RAF inhibitor PD 28.1 
Patient 4 NRAS Q61L 75 — 8.0 RAF inhibitor SD 15.3 
Patient 5 NRAS Q61K 45.8 — 3.3 RAF inhibitor SD 7.3 
Patient 6 KIT/KDR/PDGFRA Amplification — 133 5.0 Multiple receptor tyrosine kinase inhibitor SD 12.6 
Patient 7 MSH6 T955fs 31.8 — 74.0 Anti-PD-L1 antibody SD 100.4 
IndexGeneAA changeAllele frequency (%)Copy numberTMBDrugs in clinical trialsBest responsePFS (weeks)
Patient 1 KRAS Amplification — 14 4.0 RAF inhibitor NE 1.9 
Patient 2 NRAS G12D 88.8 — 4.9 RAF inhibitor PD 3.0 
Patient 3 NRAS G13D 89.2 — 2.0 RAF inhibitor PD 28.1 
Patient 4 NRAS Q61L 75 — 8.0 RAF inhibitor SD 15.3 
Patient 5 NRAS Q61K 45.8 — 3.3 RAF inhibitor SD 7.3 
Patient 6 KIT/KDR/PDGFRA Amplification — 133 5.0 Multiple receptor tyrosine kinase inhibitor SD 12.6 
Patient 7 MSH6 T955fs 31.8 — 74.0 Anti-PD-L1 antibody SD 100.4 

Abbreviation: NE, not evaluable.

Figure 2.

Metabolic response observed from PET images in Patient 5 who harbored NRAS Q61K mutation and received RAF inhibitor. The left panel shows pretreatment PET images: whole-body image in the left-upper, right middle lobe mass (SUV 9.3) in the right-upper, thoracic spine mass (SUV 13.3) in the left-lower, and left kidney mass (SUV 13.0) in the right-lower. The right panel shows PET images taken 1 month after treatment with RAF inhibitor: whole-body image in the left-upper, right middle lobe mass (SUV 2.0) in the right-upper, thoracic spine mass (SUV 2.8) in the left-lower, and left kidney mass (SUV 4.0) in the right lower. SUV, standardized uptake value.

Figure 2.

Metabolic response observed from PET images in Patient 5 who harbored NRAS Q61K mutation and received RAF inhibitor. The left panel shows pretreatment PET images: whole-body image in the left-upper, right middle lobe mass (SUV 9.3) in the right-upper, thoracic spine mass (SUV 13.3) in the left-lower, and left kidney mass (SUV 13.0) in the right-lower. The right panel shows PET images taken 1 month after treatment with RAF inhibitor: whole-body image in the left-upper, right middle lobe mass (SUV 2.0) in the right-upper, thoracic spine mass (SUV 2.8) in the left-lower, and left kidney mass (SUV 4.0) in the right lower. SUV, standardized uptake value.

Close modal

The patient with BRAF V600E confirmed by NGS and PNA-PCR clamping received dabrafenib/trametinib combination after progression from anti-PD-1 inhibitor and cytotoxic chemotherapy and experienced PR. The patient's disease has not progressed for 10 months and is currently on follow-up at the time of analysis. The other BRAF V600E–positive patient had not received systemic treatment after NGS because the disease was stable.

Estimated TMB and anti-PD-1/PD-L1 antibody response

Thirty-three patients received anti-PD-1/PD-L1 antibody for the treatment of advanced melanoma. ORR was 12.1% (4 PRs) and CBR was 39.4% (N = 13). Median PFS was 15.9 weeks [95% confidence interval (CI), 11.0–35.0], ranging from 0.4 to 138.1 weeks. Estimated TMBs tended to associate with PFS of anti-PD-1/PD-L1 antibody (HR, 0.96; 95% CI, 0.92–1.00; P = 0.07). Estimated TMBs in samples of patients who responded to anti-PD-1/PD-L1 antibody tended to be higher than the others (median estimated TMB 12.5 vs. 8.2, P = 0.33). Also, estimated TMBs in samples of patients who experienced clinical benefit was higher than in others (median estimated TMB 12.0 vs. 8.1, P = 0.26). With the estimated TMB cutoff of 13, which was the best value to discriminate treatment response, the high-TMB group experienced significantly longer PFS (median 34.0 weeks, 95% CI, 15.1–NA) compared with low-TMB group (median 11.0 weeks, 95% CI, 8.1–35.0; P = 0.04; Fig. 3). The patient with the longest PFS (100.4 weeks) had an MSH6 mutation.

Figure 3.

Progression-free survival of anti-PD-1/PD-L1–treated patients according to estimated TMB. Kaplan–Meier curve shows PFS of patients who received anti-PD-1/PD-L1 antibodies. Red line denotes the low-TMB group (estimated TMB < 13), and blue line denotes the high-TMB group (estimated TMB ≥ 13).

Figure 3.

Progression-free survival of anti-PD-1/PD-L1–treated patients according to estimated TMB. Kaplan–Meier curve shows PFS of patients who received anti-PD-1/PD-L1 antibodies. Red line denotes the low-TMB group (estimated TMB < 13), and blue line denotes the high-TMB group (estimated TMB ≥ 13).

Close modal

In this study, we analyzed molecular alterations across SNVs, INDELs, CNVs, and SVs, and also estimated TMBs in the patients with BRAF V600E–negative melanoma by using SNUH FIRST-Cancer panel. Although genomic features of melanoma development are well characterized in large-scale studies with whole-exome (35) and -genome (12) sequencing, clinical applications of FFPE sample sequencing without matched normal samples were not adequately described. We found a prevalence of NRAS mutation and CDKN2A copy number deletion, as expected in our cohort diagnosed as BRAF WT. We classified 4 tumors into genomic subtypes described by TCGA. We observed high TMB with NF1 mutation and prevalence of copy number amplification in triple-WT subtype in accordance with previous reports (35, 38). These results suggest that clinical NGS panel screen is feasible to discover various types of molecular alterations and to help estimate TMBs in melanoma.

We discovered that most of the patients (92%) harbored at least one oncogenic alteration in key genes of melanoma. Pathway analysis showed half the patients with oncogenic alterations in more than one signaling pathway of melanoma. These results are encouraging with respect to finding potential therapeutic targets for patients with BRAF WT, and also highlight the complexity of advanced melanoma and challenges in clinical treatment modalities.

Overall, 70% of patients had at least one actionable alteration, and seven of these patients were enrolled in clinical trials of targeted therapies. The observed clinical benefit was modest considering no observed objective response and median PFS of 12.6 weeks. Although there are many factors that may influence clinical outcomes, it is plausible that the modest benefit to patients from the implemented targeted therapies may be ascribed to redundancy of oncogenic signaling mechanisms and interaction of multiple pathways. Nevertheless, we observed metabolic response with RAF inhibitor in 1 patient showing promise of newer investigational drugs.

NGS enables identification of multiple oncogenic alterations to design effective treatments by massive parallel sequencing of numerous target genes instead of sequential testing by conventional methods. Thus, NGS panel enhances the potential to design relevant therapies in time and cost-effective manner for patients with advanced melanoma. As targeted therapies and combinatorial regimens are being increasingly developed, we expect that clinical NGS panel such as SNUH FIRST-Cancer panel will emerge as a powerful tool to guide clinical decisions for optimal outcomes in advanced melanoma.

We found 2 patients positive for BRAF V600E mutations with allele frequencies of 8%, who were initially diagnosed as BRAF V600E–negative using IHC assay and Sanger sequencing methods. This is consistent with the previous finding that although conventional molecular tests have high reliability, they also have drawbacks and limited sensitivities (39). One of the above 2 patients responded to dabrafenib/trametinib combination treatment, even though the allele frequency of BRAF V600E was not high. These results indicate that high-depth sequencing of clinical NGS panel allows to accurately detect mutations at low allele frequencies, which is likely to provide clinical benefits.

TMB estimated by the SNUH FIRST-Cancer panel also showed the ability to predict response to immunotherapy in patients with melanoma. Estimated TMB seems to associate with PFS of anti-PD-1/PD-L1 treatment, although statistically not significant, which might have been due to small sample size and diverse lines of treatments. Median PFS was 15.9 weeks, quite shorter than that reported previously (2). This result could be due to a higher proportion of acral lentiginous melanoma subtype prevalence in the Korean population, showing lower TMB than that of superficial-spreading melanoma (22). With the estimated TMB cutoff of 13 or above, the high-TMB group has longer PFS than the low-TMB group. The adopted cutoff of 13 is consistent with the TMB cutoffs in the range 7–20 reported in the literatures (18, 40, 41). However, our TMB estimation was done with somatic SNV prediction model rather than with matched germline genomic data and the cutoff we set is therefore arbitrary. Further independent validation of the estimated TMB is necessary to assess the response to immunotherapies.

It is remarkable that the patient with MSH6 mutation had the second highest estimated TMB of 74 and experienced durable response (PFS 100.4 weeks) to immunotherapy. The patient harbored MSH6 T955fs mutation, which is known to be associated with MMR defects (37). MMR defects are associated with genomic instability (microsatellite-high tumor) and commonly has a higher mutation burden, which is associated with favorable response to immunotherapy (15, 42, 43). We show supporting evidence to the above premise in 1 patient who harbored MSH6 T955fs mutation and who responded favorably to immunotherapy. This implicates that using NGS for detecting MMR defects in patients with melanoma may be effective.

Because of the small sample size and exclusion of patients with BRAF-mutant melanoma, the results in this study may not reflect genomic demographics of patients with advanced melanoma. (21) Even though, we observed major genomic profiles and promising applicability of clinical NGS panel to these patients, which warrant further validation with larger number of patients in prospective clinical studies.

In summary, our NGS panel identified key oncogenic gene alterations and estimated TMB in patients with melanoma. These alterations showed association with clinical outcomes. Application of NGS panel provides patients with melanoma with alternative therapeutic options including stratification of patients for clinical trials using the genetic analysis method described and tailored immunotherapy for better treatment outcomes.

M. Kim has an unpaid consultant/advisory board relationship with AstraZeneca. T.M. Kim has an unpaid consultant/advisory board relationship with AstraZeneca, Novartis, and Takeda. D.-W. Kim has an unpaid consultant/advisory board relationship with AstraZeneca/MedImmune, Lilly, Yuhan, Boehringer Ingelheim, Chong Keun Dang, Pfizer, and Novartis. No potential conflicts of interest were disclosed by the other authors.

Conception and design: C. Park, M. Kim, M.J. Kim, H. Kim, T.M. Kim, D.S. Heo

Development of methodology: C. Park, M. Kim, M.J. Kim, H. Kim, B. Keam, T.M. Kim, D.S. Heo

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C. Park, M. Kim, M.J. Kim, H. Kim, B. Keam, T.M. Kim

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C. Park, M. Kim, M.J. Kim, H. Kim, C.-Y. Ock, B. Keam, T.M. Kim, D.-W. Kim, J.-I. Kim, D.S. Heo

Writing, review, and/or revision of the manuscript: C. Park, M. Kim, M.J. Kim, H. Kim, C.-Y. Ock, B. Keam, T.M. Kim, D.-W. Kim, J.-I. Kim

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Kim, H. Kim, B. Keam, J.-I. Kim

Study supervision: M. Kim, M.J. Kim, D.S. Heo

This study was supported by a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number: HI14C1277).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Sandru
A
,
Voinea
S
,
Panaitescu
E
,
Blidaru
A
. 
Survival rates of patients with malignant melanoma
.
J Med Life
2014
;
7
:
572
6
.
2.
Wolchok
JD
,
Chiarion-Sileni
V
,
Gonzalez
R
,
Rutkowski
P
,
Grob
JJ
,
Cowey
CL
, et al
Overall survival with combined nivolumab and ipilimumab in advanced melanoma
.
N Engl J Med
2017
;
377
:
1345
56
.
3.
Schachter
J
,
Ribas
A
,
Long
GV
,
Arance
A
,
Grob
JJ
,
Mortier
L
, et al
Pembrolizumab versus ipilimumab for advanced melanoma: final overall survival results of a multicentre, randomised, open-label phase 3 study (KEYNOTE-006)
.
Lancet
2017
;
390
:
1853
62
.
4.
Davies
H
,
Bignell
GR
,
Cox
C
,
Stephens
P
,
Edkins
S
,
Clegg
S
, et al
Mutations of the BRAF gene in human cancer
.
Nature
2002
;
417
:
2
4
.
5.
Besmer
P
,
Murphy
JE
,
George
PC
,
Qiu
F
,
Bergold
PJ
,
Lederman
L
, et al
A new acute transforming feline retrovirus and relationship of its oncogene v-kit with the protein kinase gene family
.
Nature
1986
;
320
:
698
9
.
6.
Satzger
I
,
Schaefer
T
,
Kuettler
U
,
Broecker
V
,
Voelker
B
,
Ostertag
H
, et al
Analysis of c-KIT expression and KIT gene mutation in human mucosal melanomas
.
Br J Cancer
2008
;
99
:
2065
9
.
7.
Larkin
J
,
Ascierto
PA
,
Dréno
B
,
Atkinson
V
,
Liszkay
G
,
Maio
M
, et al
Combined vemurafenib and cobimetinib in BRAF-mutated melanoma
.
N Engl J Med
2014
;
371
:
1867
76
.
8.
Robert
C
,
Karaszewska
B
,
Schachter
J
,
Rutkowski
P
,
Mackiewicz
A
,
Stroiakovski
D
, et al
Improved overall survival in melanoma with combined dabrafenib and trametinib
.
N Engl J Med
2015
;
372
:
30
9
.
9.
Dummer
R
,
Ascierto
PA
,
Gogas
HJ
,
Arance
A
,
Mandala
M
,
Liszkay
G
, et al
Overall survival in patients with BRAF-mutant melanoma receiving encorafenib plus binimetinib versus vemurafenib or encorafenib (COLUMBUS): a multicentre, open-label, randomised, phase 3 trial
.
Lancet Oncol
2018
;
19
:
1315
27
.
10.
Hodi
FS
,
Corless
CL
,
Giobbie-Hurder
A
,
Fletcher
JA
,
Zhu
M
,
Marino-Enriquez
A
, et al
Imatinib for melanomas harboring mutationally activated or amplified kit arising on mucosal, acral, and chronically sun-damaged skin
.
J Clin Oncol
2013
;
31
:
3182
90
.
11.
Shain
AH
,
Yeh
I
,
Kovalyshyn
I
,
Sriharan
A
,
Talevich
E
,
Gagnon
A
, et al
The genetic evolution of melanoma from precursor lesions
.
N Engl J Med
2015
;
373
:
1926
36
.
12.
Hayward
NK
,
Wilmott
JS
,
Waddell
N
,
Johansson
PA
,
Field
MA
,
Nones
K
, et al
Whole-genome landscapes of major melanoma subtypes
.
Nature
2017
;
545
:
175
80
.
13.
Serrati
S
,
Simona De
S
,
Pilato
B
,
Petriella
D
,
Lacalamita
R
,
Tommasi
S
, et al
Next-generation sequencing: advances and applications in cancer diagnosis
.
Onco Targets Ther
2016
;
9
:
7355
65
.
14.
Siu
LL
,
Conley
BA
,
Boerner
S
,
LoRusso
PM
. 
Next-generation sequencing to guide clinical trials
.
Clin Cancer Res
2015
;
21
:
4536
44
.
15.
Chalmers
ZR
,
Connelly
CF
,
Fabrizio
D
,
Gay
L
,
Ali
SM
,
Ennis
R
, et al
Analysis of 100,000 human cancer genomes reveals the landscape of tumor mutational burden
.
Genome Med
2017
;
9
:
1
14
.
16.
Garofalo
A
,
Sholl
L
,
Reardon
B
,
Taylor-Weiner
A
,
Amin-Mansour
A
,
Miao
D
, et al
The impact of tumor profiling approaches and genomic data strategies for cancer precision medicine
.
Genome Med
2016
;
8
:
1
10
.
17.
Chan
TA
,
Yarchoan
M
,
Jaffee
E
,
Swanton
C
,
Quezada
SA
,
Stenzinger
A
, et al
Development of tumor mutation burden as an immunotherapy biomarker: utility for the oncology clinic
.
Ann Oncol
2019
;
30
:
44
56
.
18.
Hellmann
MD
,
Ciuleanu
TE
,
Pluzanski
A
,
Lee
JS
,
Otterson
GA
,
Audigier-Valette
C
, et al
Nivolumab plus ipilimumab in lung cancer with a high tumor mutational burden
.
N Engl J Med
2018
;
378
:
2093
104
.
19.
Reiman
A
,
Kikuchi
H
,
Scocchia
D
,
Smith
P
,
Tsang
YW
,
Snead
D
, et al
Validation of an NGS mutation detection panel for melanoma
.
BMC Cancer
2017
;
17
:
1
7
.
20.
Siroy
AE
,
Boland
GM
,
Milton
DR
,
Roszik
J
,
Frankian
S
,
Malke
J
, et al
Beyond BRAFV600: clinical mutation panel testing by next-generation sequencing in advanced melanoma
.
J Invest Dermatol
2015
;
135
:
508
15
.
21.
Erdmann
F
,
Lortet-Tieulent
J
,
Schüz
J
,
Zeeb
H
,
Greinert
R
,
Breitbart
EW
, et al
International trends in the incidence of malignant melanoma 1953-2008-are recent generations at higher or lower risk?
Int J Cancer
2013
;
132
:
385
400
.
22.
Seo
S
,
Oh
CK
,
Ko
HC
,
Kim
MB
,
Kwon
KS
,
Bae
YC
. 
A 10-year review of cutaneous melanomas in Korea: clinicopathologic features and changes by multidisciplinary team approach
.
J Eur Acad Dermatology Venereol
2009
;
23
:
1463
5
.
23.
Ascierto
PA
,
Kirkwood
JM
,
Grob
JJ
,
Simeone
E
,
Grimaldi
AM
,
Maio
M
, et al
The role of BRAF V600 mutation in melanoma
.
J Transl Med
2012
;
10
:
1
9
.
24.
Lee
HW
,
Song
KH
,
Hong
JW
,
Jeon
SY
,
Ko
DY
,
Kim
KH
, et al
Frequency of BRAF mutation and clinical relevance for primary melanomas
.
Korean J Pathol
2012
;
46
:
246
52
.
25.
Bolger
AM
,
Lohse
M
,
Usadel
B
. 
Trimmomatic: a flexible trimmer for Illumina sequence data
.
Bioinformatics
2014
;
30
:
2114
20
.
26.
Li
H
,
Durbin
R
. 
Fast and accurate long-read alignment with Burrows-Wheeler transform
.
Bioinformatics
2010
;
26
:
589
95
.
27.
Cibulskis
K
,
Lawrence
MS
,
Carter
SL
,
Sivachenko
A
,
Jaffe
D
,
Sougnez
C
, et al
Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples
.
Nat Biotechnol
2013
;
31
:
213
9
.
28.
Costello
M
,
Pugh
TJ
,
Fennell
TJ
,
Stewart
C
,
Lichtenstein
L
,
Meldrim
JC
, et al
Discovery and characterization of artifactual mutations in deep coverage targeted capture sequencing data due to oxidative DNA damage during sample preparation
.
Nucleic Acids Res
2013
;
41
:
1
12
.
29.
Wang
K
,
Li
M
,
Hakonarson
H
. 
ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data
.
Nucleic Acids Res
2010
;
38
:
1
7
.
30.
Talevich
E
,
Shain
AH
,
Botton
T
,
Bastian
BC
. 
CNVkit: genome-wide copy number detection and visualization from targeted DNA sequencing
.
PLoS Comput Biol
2016
;
12
:
1
18
.
31.
Rausch
T
,
Zichner
T
,
Schlattl
A
,
Stütz
AM
,
Benes
V
,
Korbel
JO
. 
DELLY: structural variant discovery by integrated paired-end and split-read analysis
.
Bioinformatics
2012
;
28
:
333
9
.
32.
Thorvaldsdóttir
H
,
Robinson
JT
,
Mesirov
JP
. 
Integrative genomics viewer (IGV): high-performance genomics data visualization and exploration
.
Brief Bioinform
2013
;
14
:
178
92
.
33.
Robinson
JT
,
Thorvaldsdottir
H
,
Winckler
W
,
Guttman
M
,
Lander
ES
,
Getz
G
, et al
Integrative genomics viewer
.
Nat Biotechnol
2011
;
29
:
24
6
.
34.
Chakravarty
D
,
Gao
J
,
Phillips
S
,
Kundra
R
. 
OncoKB: a precision oncology knowledge base
.
JCO Precis Oncol
2017
;
1
:
1
16
.
35.
Paklina
OV
,
Bristow
CA
,
Holt
RA
,
Tan
D
,
Lee
JE
,
Ramirez
NC
, et al
Genomic classification of cutaneous melanoma
.
Cell
2015
;
161
:
1681
96
.
36.
Eisenhauer
EA
,
Therasse
P
,
Bogaerts
J
,
Schwartz
LH
,
Sargent
D
,
Ford
R
, et al
New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1)
.
Eur J Cancer
2009
;
45
:
228
47
.
37.
Melloni
GEM
,
Mazzarella
L
,
Bernard
L
,
Bodini
M
,
Russo
A
,
Luzi
L
, et al
A knowledge-based framework for the discovery of cancer-predisposing variants using large-scale sequencing breast cancer data
.
Breast Cancer Res
2017
;
19
:
1
14
.
38.
Cirenajwis
H
,
Lauss
M
,
Ekedahl
H
,
Torngren
T
,
Kvist
A
,
Saal
LH
, et al
NF1-mutated melanoma tumors harbor distinct clinical and biological characteristics
.
Mol Oncol
2017
;
11
:
438
51
.
39.
Cheng
L
,
Lopez-Beltran
A
,
Massari
F
,
Maclennan
GT
,
Montironi
R
. 
Molecular testing for BRAF mutations to inform melanoma treatment decisions: a move toward precision medicine
.
Mod Pathol
2018
;
31
:
24
38
.
40.
Goodman
AM
,
Kato
S
,
Bazhenova
L
,
Patel
SP
,
Frampton
GM
,
Miller
V
, et al
Tumor mutational burden as an independent predictor of response to immunotherapy in diverse cancers
.
Mol Cancer Ther
2017
;
16
:
2598
608
.
41.
Morrison
C
,
Pabla
S
,
Conroy
JM
,
Nesline
MK
,
Glenn
ST
,
Dressman
D
, et al
Predicting response to checkpoint inhibitors in melanoma beyond PD-L1 and mutational burden
.
J Immunother Cancer
2018
;
6
:
1
12
.
42.
Nebot-Bral
L
,
Brandao
D
,
Verlingue
L
,
Rouleau
E
,
Caron
O
,
Despras
E
, et al
Hypermutated tumours in the era of immunotherapy: the paradigm of personalised medicine
.
Eur J Cancer
2017
;
84
:
290
303
.
43.
Le
DT
,
Uram
JN
,
Wang
H
,
Bartlett
BR
,
Kemberling
H
,
Eyring
AD
, et al
PD-1 blockade in tumors with mismatch-repair deficiency
.
N Engl J Med
2015
;
372
:
2509
20
.