Comprehensive molecular and clinical characterization of Asian melanoma patients treated with anti-PD-1 antibody

Background Clinical features of Asian melanoma patients are distinct from those of Western patients. This study was designed to determine the molecular and clinical characteristics of Asian melanoma patients treated with anti-PD-1 antibody. Methods Patients with recurrent or metastatic melanoma who began anti-PD-1 antibody therapy between January 2015 and April 2018 were retrospectively reviewed. Patients who underwent next-generation sequencing were also analyzed. Results A total of 152 patients were included. The median age was 61 years, and 53% of patients were female. A total of 56 patients (37%) received immunotherapy as second-line or greater chemotherapy. Primary sites were acral (38%), mucosal (31%), cutaneous (24%), uveal (2%), and unknown (5%). The overall response rate was 17% (95% CI, 11–22%), and disease control rate was 60% (95% CI, 52–68%). The median progression-free survival (PFS) was 4.2 months (95% CI, 1.8–6.6 months), and median overall survival (OS) was 32.9 months (95% CI, 20.0–45.7 months). However, BRAFV600 and KIT mutational statuses were not associated with response or survival. High neutrophil-lymphocyte ratio (NLR) was associated with poor PFS (median PFS 6.9 vs. 2.4 months, p = 0.015) and OS (median OS NR vs. 10.4 months, p < 0.001). In multivariate analysis, high NLR independently predicted poor survival. Conclusion This study includes the largest set of integrated genomic data analyzing Asian patients with melanoma treated with immunotherapy. BRAF V600 and KIT mutational statuses were not associated with response or survival, and high NLR was a strong predictor of poor response to and survival with anti-PD-1 therapy.


Background
The role of anti-PD-1 therapy is well established in malignant melanomapembrolizumab and nivolumab have been approved as first-line therapy in advanced melanoma. Such advances in immunotherapy have significantly improved the response rates and survival outcomes in patients with advanced melanoma [1,2]. Melanoma was traditionally classified based on histologic growth pattern superficial spreading, lentigo maligna, nodular, and acral lentiginous melanoma. However, with increasing data on distinct molecular aberrations and primary locations, a novel classification has been proposedcutaneous (with or without chronic sun-induced damage), acral, and mucosal melanoma. Previous studies have shown that acral and mucosal melanoma, which have higher frequency of KIT mutation [3,4], are the most prevalent subtypes in Asian populations [5,6]. Conversely, cutaneous melanoma is the predominant subtype in Caucasian populations, which have higher incidence of BRAF mutation [4,7].
Despite the increasing incidence of malignant melanoma in Asia, the absolute incidence remains small [6,8,9], and there are limited data available on immunotherapy treatment outcomes in Asian patients with melanoma. The effect of mutation status on response to immunotherapy is poorly understood. Despite the role of anti-PD-1 therapy as a first-line agent, the use of biomarkers for patient selection is an area of ongoing debate. In search of a readily available biomarker, the ratio of neutrophils to lymphocytes (NLR) has been evaluated in many solid cancers, including melanoma [10][11][12], and has emerged as an important biomarker to predict response to immunotherapy.
The aim of the present study was to evaluate the treatment efficacy of anti-PD-1 therapy in Asian patients with melanoma. Additionally, we sought biomarkers to predict treatment response to anti-PD-1 antibody in patients with melanoma.

Patients
A total of 152 consecutive patients with recurrent or metastatic melanoma who began anti-PD-1 (nivolumab or pembrolizumab) therapy between January 2015 and April 2018 were retrospectively analyzed. Baseline characteristics including age, sex, ECOG performance status, previous therapies, melanoma subtype, disease stage, metastatic sites, baseline CBC, LDH, treatment response, adverse events, and survival outcomes were obtained through medical records and tumor imaging review. This study was approved by the Institutional Review Board of Samsung Medical Center (IRB No. 2018-07-080), and informed consent was waived. All genomic analyses using cancer panel were used with consent.

Treatment and response
All patients received pembrolizumab 2 mg/kg IV every 3 weeks or nivolumab 3 mg/kg IV every 2 weeks until progression, unacceptable toxicity, or patient refusal. Patients were evaluated at baseline and every 6-9 weeks after starting treatment. Response categories were assessed using RECIST 1.1 [13]. In addition to response defined by RECIST, efficacy was also defined by durable clinical benefit (DCB), which included complete response (CR), partial response (PR), and stable disease (SD) lasting for more than 6 months. Adverse events were graded based on the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0 (CTCAE v5.0, National Cancer Institute, Bethesda, MD, USA).

Statistical analysis
Descriptive statistics were used to summarize patient and treatment characteristics. NLR was defined as the quotient of baseline absolute neutrophil count divided by absolute lymphocyte count. Each nominal variable was compared using Fisher's exact test or X 2 -test. PFS was defined as the time from initiation of anti-PD-1 therapy to documentation of disease progression or death. OS was defined as the time from initiation of anti-PD-1 therapy to death from any cause. Survival curves of categorical variables were calculated using the Kaplan-Meier method and compared using the log-rank test.
Univariate/multivariate models of patients for tumor characteristics in association with PFS and OS were based on Cox proportional hazards regression analyses. Results were presented as hazard ratio (HR) with 95% confidence interval (CI). A p-value less than 0.05 was considered statistically significant, and all analyses were performed using IBM SPSS Statistics 24 (Armonk, NY, USA).
BRAF V600 and KIT mutational statuses were evaluated in 133 and 98 patients, respectively, including 59 patients who underwent NGS. The incidence of BRAF V600 and KIT mutants was 23 of 133 patients (17%) and 14 of 98 patients (14%), respectively.

Response and survival
The overall response rate (ORR) and disease control rate (DCR) were 17% (95% CI, 11-22%) and 60% (95% CI, 52-68%), respectively ( Table 2). The median time to response was 2.0 months, and the median duration of response was 6.2 months. The median OS was 32.9 months (95% CI, 20.0-45.7 months), and the median PFS was 4.2 months (95% CI, 1.8-6.6 months). There was no significant difference in response rates or survival outcomes according to type of anti-PD-1 therapy received.
Patients with BRAF V600 mutant (mBRAF V600 ) who were previously treated with BRAF/MEK inhibitors demonstrated poor PFS (1.2 vs. 8.0 months, p = 0.039) and OS (1.2 vs. 32.9 months, p = 0.002) compared to patients with mBRAF V600 and no previous therapy with BRAF/ MEK inhibitors. No patients with mBRAF V600 who underwent previous BRAF/MEK inhibitor treatment demonstrated a clinical response to anti-PD-1 therapy.

Cox regression models on survival outcomes
Univariate and multivariate analyses using a Cox model were performed including NLR in addition to potential prognostic factors of age, ECOG PS, histologic subtype, M stage, mutational status, and LDH (Table 3). Low LDH level and low NLR (< 2.10) level were associated with

Discussion
To our knowledge, this is the largest integrated genomic analysis of Asian patients with melanoma treated with anti-PD-1 therapy. In this study, the ORR and DCR were 17% (95% CI, 11-22%) and 60% (95% CI, 52-68%), respectively. These values are slightly lower than those of previous prospective studies, which reported ORR ranging from 21 to 40%. However, our study included patients with poor performance status and brain metastases. This is more representative of patients with advanced melanoma in the clinical setting and possibly explains the low response rates. The median PFS and OS of 4.2 and 32.9 months, respectively, were comparable to those of Caucasian patients [1,2,15,16]. These findings suggest that immune checkpoint blockades are viable treatment options for Asian patients, especially considering the limited utilization of BRAF/MEK inhibitors due to low BRAF V600 mutant prevalence. In exploratory analysis, mBRAF V600 patients who were previously treated with BRAF/MEK inhibitors were analyzed for response rates and survival outcomes. They demonstrated poor PFS and OS compared to mBRAF V600 patients without previous therapy with BRAF/MEK inhibitors. It is also notable that no mBRAF V600 patients with previous BRAF/MEK inhibitor therapy experienced a clinical response to immunotherapy. Although we should be cautious about drawing conclusions from a small subset population, the results suggest that early immunotherapy in mBRAF V600 patients is more effective than early targeted therapies. Two randomized trials (NCT02224781 and NCT02631447) are currently investigating the optimal treatment sequence in a larger mBRAF V600 cohort, and the results are highly awaited.
The incidence of KIT mutations in our study population was relatively low compared to that of previous reports [3,17]. This is due to the low accessibility of the c-KIT inhibitor imatinib in the real-world setting. Consequently, clinicians are reluctant to pursue genomic testing for KIT mutational status.
NRAS mutation was detected in 19% of our cohort who underwent NGS, an incidence similar to that of previous reports [18]. Johnson et al. reported that NRAS-mutant patients demonstrate a superior response to immunotherapy (ORR 32 vs. 20%, p = 0.07), particularly to anti-PD-1 and anti-PD-L1 agents (ORR 64 vs. 30%, p-value not provided), which is possibly related to higher PD-L1 expression [19]. Horn et al. reported that melanoma cell lines with chromosomal loss of CDKN2A associated with JAK2 deletion are prone to immunotherapy resistance. However, neither NRAS mutation nor CDKN2A deletion status had statistically significant effects on response rate or survival outcome with anti-PD-1 therapy in our cohort. Whether this conflicting outcome is related to ethnic differences requires further validation in future prospective studies.
Studies regarding response outcomes to immunotherapy according to PD-L1-positivity are confounding [1,15]. There is an increasing need for an easily accessible and affordable biomarker to predict anti-PD-1 therapy responses, and NLR has emerged as a promising option. The tumor microenvironment is characterized by chronic inflammation, and neutrophils reflect the host inflammatory status in patients with cancer [20]. Although their role is multifactorial, neutrophils have been shown to contribute to tumor initiation, angiogenesis, proliferation, and metastatic spread [21]. However, the utility of NLR as a predictive biomarker in Asian patients with melanoma has not been previously validated.
In our study, NLR was the only independent factor, other than LDH, associated with superior PFS and OS on both univariate and multivariate Cox regression analyses. This is consistent with findings from previous studies [22][23][24], highlighting the role of NLR as an important biomarker to predict response and survival outcomes to immunotherapy in patients with melanoma. A cut-off value of 2.10 was utilized in this study; however, similar trends were demonstrated with cut-off values of 2.0, 3.0, and 5.0. Further studies are needed to validate the best cut-off value for utilization of NLR.

Conclusion
This study is the largest integrated genomic analysis of Asian patients with melanoma treated with anti-PD-1 Table 4 Treatment-related adverse events The only case with treatment discontinuation therapy. The heterogeneous patient population in our study reflects the real-world efficacy and safety of anti-PD-1 therapy in patients with advanced melanoma. Despite different distributions in subtypes, BRAF V600 or KIT mutational status does not affect response to immunotherapy. Low NLR is a strong predictor of higher response and longer survival in response to immunotherapy, and it may be useful as a biomarker.