Efficacy of nivolumab monotherapy against acral lentiginous melanoma and mucosal melanoma in Asian patients

Dear Editor, Early clinical trials showed that treatment with anti‐programmed death (PD)‐1 agents, such as nivolumab and pembrolizumab, led to marked improvements in the prognosis of patients with advanced melanoma.1,2 However, most large‐scale studies analysed mainly the efficacy of anti‐PD‐1 agents in white populations. As the clinical subtypes of melanoma differ widely among populations,3,4,5 and the response to anti‐PD‐1 agents is known to vary among the clinical subtypes of the disease,6,7,8 the efficacy of anti‐PD‐1 agents in Asian populations remains unclear. Herein we report the clinical outcomes of patients with advanced melanoma in Japan who received nivolumab monotherapy at our department. This is the largest detailed retrospective study to date of the efficacy of nivolumab monotherapy against each clinical subtype of melanoma among Japanese patients. This retrospective study included 71 patients with advanced melanoma who received nivolumab monotherapy at our institution from September 2014 to May 2018. Of these patients, 68 were included in the analysis. Clinical data were collected, including information about the patients’ age and sex, tumour locations, presence or absence of visceral metastases, BRAF mutation status, objective response rate (ORR), overall survival (OS) and progression‐free survival (PFS). Tumour response was defined according to the Response Evaluation Criteria in Solid Tumors 1·1. All statistical analyses were performed using Microsoft Excel 2016. The significance of differences in the ORR was estimated using Fisher's exact test. OS was calculated from the administration of the first dose of nivolumab to death from any cause in the observation period or the last follow‐up. PFS was calculated from the administration of the first dose of nivolumab to the date of radiological progression, death from any cause in the observation period, or the last follow‐up. OS and PFS were estimated using the Kaplan–Meier method. P‐values < 0·05 were considered statistically significant. This research was approved by the ethics committee of the Tokyo Metropolitan Cancer and Infectious Diseases Center Komagome Hospital. Data regarding the patients’ attributes and treatment efficacy are summarized in Table 1 (details of OS and PFS available on request). We classified the patients into three groups according to the locations of their tumours. The numbers of cases with and without visceral metastases are shown in Table 1, along with the numbers of BRAF mutations. The ORR was 29%, with two of 68 patients achieving complete responses, 18 achieving partial responses, 11 having stable disease and 37 having progressive disease. The ORR was 43% for cutaneous melanoma, 21% for mucosal melanoma (MCM) and 19% for acral lentiginous melanoma (ALM). The ORRs for tumours involving visceral metastases were 42%, 18% and 8%, respectively. Among the cases of melanoma involving visceral metastases, ALM and MCM tended to exhibit worse ORRs than cutaneous melanoma (13% vs. 42%, P=0·028). In the survival analysis, the mean estimated OS and PFS periods were 711 days and 401 days for cutaneous melanoma, 422 days and 226 days for MCM and 421 days and 197 days for ALM, respectively. Although the OS and PFS for ALM and MCM tended to be worse than those for cutaneous melanoma, no significant differences were observed. In addition, there was no significant difference in OS or PFS among the BRAF mutation‐positive and ‐negative cases. The patients’ attributes and clinical outcomes for each clinical subtype ALM, acral lentiginous melanoma; MCM, mucosal melanoma; F, female; M, male; CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; ORR, objective response rate. aBRAF mutations were detected in 10 of 23 cutaneous melanomas, one of 15 ALMs and none of the 18 MCMs subjected to BRAF mutation analysis. The patients’ attributes and clinical outcomes for each clinical subtype ALM, acral lentiginous melanoma; MCM, mucosal melanoma; F, female; M, male; CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; ORR, objective response rate. aBRAF mutations were detected in 10 of 23 cutaneous melanomas, one of 15 ALMs and none of the 18 MCMs subjected to BRAF mutation analysis. In early clinical trials of anti‐PD‐1 agents, the ORR was reported to be 40·0% for nivolumab and 32·9–33·7% for pembrolizumab.1,2 However, in our study the ORR for nivolumab was 29% for all melanomas, 19% for ALM and 21% for MCM. As for the poor prognoses of MCM and ALM, Hayward et al. revealed that the mutation burden, which is a prognostic factor for the response to immune checkpoint inhibitor treatment, was lower in the ALM and MCM groups than with the other clinical subtypes.8 In an analysis of pooled data from five clinical trials, D'Angelo et al. concluded that nivolumab monotherapy produced worse clinical outcomes in patients with MCM.6 In the latter study, the ORR for nivolumab was 23% for MCM and 41% for cutaneous melanoma. In contrast, another analysis of data from clinical trials reached a different conclusion. Shoushtari et al. studied the efficacy of anti‐PD‐1 agents in patients with ALM or MCM. As a result, they found that the ORR was 32% in ALM and 23% in MCM, and concluded that their results were comparable with those of early clinical trials involving other clinical subtypes.7 As ALM and MCM are rare subtypes of melanoma worldwide, there are limited clinical data about these conditions, and no definitive conclusion about the efficacy of anti‐PD‐1 agents against these subtypes has been reached. Taken together with the results of our study, these facts suggest that anti‐PD‐1 agents are less effective in Japanese patients than in white patients. While ALM and MCM account for only a small percentage of melanomas in white populations,3,4 these clinical subtypes account for approximately 50% of melanomas in Japan.4,5 The differences in the frequencies of each clinical subtype of melanoma among populations might explain the worse ORRs seen in Japanese reports. In conclusion, nivolumab monotherapy is less effective in Japanese patients with melanoma than in white patients, apparently because of the tendency for the response rates of ALM and MCM, which account for a high percentage of Japanese melanoma cases, to be low. Funding sources: none. Conflicts of interest: none to declare.