Response to immune checkpoint inhibitors in acral melanoma: A nationwide cohort study

Olivier J. van Not *, Melissa M. de Meza , Alfons J.M. van den Eertwegh , John B. Haanen , Christian U. Blank , Maureen J.B. Aarts , Franchette W.P.J. van den Berkmortel , Jesper van Breeschoten , Jan-Willem B. de Groot , Geke A.P. Hospers , Rawa K. Ismail , Ellen Kapiteijn , Djura Piersma , Roos S. van Rijn , Marion A.M. Stevense-den Boer , Astrid A.M. van der Veldt , Gerard Vreugdenhil , Han J. Bonenkamp , Marye J. Boers-Sonderen , Willeke A.M. Blokx , Michel W.J.M. Wouters , Karijn P.M. Suijkerbuijk b


Introduction
The introduction of immune checkpoint inhibitors (ICIs) has improved the survival of patients with advanced melanoma, especially for cutaneous melanoma (CM), in both trial and real-world settings [1e3]. The most frequently occurring melanoma type is CM, most commonly originating from the hair-bearing skin [4]. However, melanoma can also arise at other sites such as mucosal surfaces (mucosal melanoma), the uvea of the eye (uveal melanoma), or at the non-hair-bearing glabrous skin on the palms and soles and nail apparatus (acral melanomas; AM) [5]. AM are distinct from other CMs since they have a different genetic profile [6,7] and a lower tumour mutational burden (TMB) [8]. Previous studies have demonstrated a poor prognosis of patients diagnosed with advanced AM compared to nonacral CM. Still, data on the effectiveness of checkpoint inhibitors in the Western population is limited [9,10]. AM have a relatively low incidence in the Caucasian population [11], and studies investigating the response to ICIs in AM in Caucasians include limited sample sizes [12,13]. This nationwide study aimed to investigate the differences in objective response rate (ORR), progression-free survival (PFS), and overall survival (OS) to ICIs in a large cohort of Dutch patients with advanced AM and CM. We hypothesise that patients with AM have a worse outcome compared to patients with CM and that the effectiveness of ICIs in patients with AM is lower in the advanced setting.

Materials and Methods
Data were retrieved from the Dutch Melanoma Treatment Registry (DMTR). The DMTR prospectively registers data of all patients with unresectable stage III and IV melanoma in the Netherlands since 2012. Jochems et al. [14] have shown the high quality of this registry. The medical ethical committee approved research using DMTR data, and this research was not deemed subject to the Medical Research Involving Human Subjects Act in compliance with Dutch regulations. For this study, the dataset cut-off date was 2nd September 2021.

Patients
We included all patients with stage III and IV melanoma, 18 years or older, with AM or non-acral CM receiving firstline anti-PD-1 monotherapy or ipilimumab-nivolumab for irresectable stage III and IV. Patients that were treated on clinical trials and patients that received prior adjuvant treatment were excluded. All melanomas registered as acral lentiginous melanoma in the DMTR, and melanomas located on the glabrous skin of the hand and feet or subungual melanomas were considered AMs.

Characteristics
The following patient and tumour characteristics at diagnosis were registered for all patients: age at diagnosis, gender, Eastern cooperative oncology group performance status (ECOG PS), lactate dehydrogenase levels (LDH), primary melanoma location, type of melanoma, Breslow thickness, ulceration, mutation (BRAF, NRAS, KIT, GNAQ, GNA11, other or wild type), liver metastasis, brain metastasis, number of organ sites with metastases, median time from primary melanoma to metastatic disease and AJCC staging system 8th edition [15].

Outcomes
For patients with advanced melanoma who received systemic therapy, outcomes were calculated for anti-PD-1 monotherapy and ipilimumab-nivolumab separately. Patients who did not have a response evaluation were excluded from the analysis. First-line ORR and ORR in any treatment line were calculated separately. The treating physicians determined response evaluation in line with Response Evaluation Criteria in Solid Tumours (RECIST) version 1.1 [16]. PFS and OS were calculated for first-line treated patients only. A Cox proportional hazards model was used to perform a multivariate regression analysis to assess individual factors associated with PFS and OS.

Statistical analysis
Baseline characteristics were analysed using descriptive statistics. Pearson's chi-squared test was used to compare categorical variables and the t-test for continuous variables. The median follow-up time was estimated from the first visit to the melanoma centre using the reversed KaplaneMeier method [17]. Covariates used in the multivariable regression analysis for patients with advanced melanoma were age, gender, ECOG PS, LDH level, liver metastasis, brain metastasis, number of organ sites with metastasis and type of systemic therapy. The ORR was defined as the proportion of evaluable patients who were tumour-free or achieved a complete response or partial response. Patients were deemed not evaluable if they died from a non-melanoma-related cause before their first evaluation of response or did not have a response registered in the DMTR. The median PFS and OS were calculated using the KaplaneMeier method. PFS was calculated from the start of systemic therapy until progression, death or the last moment of follow-up. OS was calculated from the start of systemic therapy until death by any cause or the last moment of follow-up. Patients not reaching the end-point were right-censored at the date of the last contact. A Cox proportional hazards model was used to perform a multivariable regression analysis to assess factors associated with PFS and OS. Comparisons were considered statistically significant for two-sided P-values 0.05. Data handling and statistical analyses were performed using R studio (version 4.0.2) [18], packages tidyverse [19], tableone [20], survival [21], and survminer [22].

Results
From 2013 to 2021, 2580 patients with advanced melanoma (unresectable stage III-IV) were registered in the DMTR who received ICIs as their first-line treatment. We excluded 374 patients with melanoma of unknown primary, 26 patients with uveal melanoma, 101 patients with mucosal melanoma and 21 patients because information regarding the location of their primary melanoma was lacking. In total, 2058 patients treated with first-line ICIs met the inclusion criteria. Eightyeight patients treated with ICIswere diagnosed with AM, of which 70 received anti-PD-1 monotherapy and 18 ipilimumab-nivolumab. Of the 1970 patients with CM treated with ICIs, 1402 patients received anti-PD-1 monotherapy and 568 patients received ipilimumabnivolumab. The median follow-up was 32.2 months for patients treated with anti-PD-1 and 23.9 months for patients treated with ipilimumab-nivolumab.

Patient and tumour characteristics
Patients with AM had higher Breslow thickness, more ulcerated melanomas but lower AJCC metastatic stage and higher T-stage at primary diagnosis. BRAF mutations were less frequent in the AM group than in the CM group (10% versus 42%; P < 0.001) and KIT mutations were seen more often in patients with AM (7% versus 2%; P Z 0.001). All baseline characteristics are shown in [ Table 1]. Except for mutation status, no significant differences at baseline existed between the acral and cutaneous ipilimumab-nivolumab patients.

ORR
ORR on first-line anti-PD-1 was 34% for AM and 54% for CM. In any treatment line, ORR was 31% for AM and 52% for CM. For ipilimumab-nivolumab, first-line ORR was 31% for AM and 52% for CM and 23% for AM and 43% for CM in any treatment line [ Table 2].

PFS
The median PFS was significantly longer for patients with CM than for patients with AM receiving anti-PD-1 monotherapy.

Overall survival
The median OS was significantly lower for patients with AM than for CM in both treatment cohorts [ Fig. 3a Fig. 4].

Subsequent treatment
Of the 88 AM patients receiving first-line checkpoint inhibitors, 35 received a second treatment line. Thirteen of these patients received ipilimumab, eight received anti-PD-1 monotherapy and five received BRAF/MEK inhibitors. All subsequent treatment lines can be seen in [ Supplementary Tables 1a and 1b].

Discussion
To our knowledge, this real-world, population-based study is the largest study to demonstrate the lower effectiveness of ICIs in patients with AM in direct comparison to patients with CM. ORR, median PFS and median OS were all significantly lower in the advanced AM group, despite relatively lower AJCC stages. The acral subtype was significantly associated with a higher hazard for progression and death in multivariable analysis. Although we describe the largest ipilimumab-nivolumab-treated cohort of patients with advanced AM thus far, the number of ipilimumab-nivolumab-treated patients is still small. Moreover, a direct comparison of anti-PD-1 and ipilimumab-nivolumab-treated patients is potentially hampered by selection bias. In both the AM and CM cohorts, patients with AM and CM treated with ipilimumabnivolumab had higher LDH levels and more often had liver and brain metastases than patients with anti-PD-1 (data were not shown), which possibly explains limited response rates in this treatment group. Therefore, even though one might tend towards using dual checkpoint inhibition in advanced AM patients based on the limited effectiveness of anti-PD-1 monotherapy, our data do not provide conclusive evidence to support or discourage this. The lower effectiveness of ICIs in AM than in CM might be due to the lower TMB which can be explained by its non-UV-related pathogenesis. Furney et al. [23] demonstrated a lower TMB in AM than in CM. Furthermore, the frequency of tumour-infiltrating lymphocytes, a known predictive factor for ICI-response, has been shown to be lower in AM [24]. Table 2 Objective response rate in first and any treatment line in advanced melanoma stratified by melanoma location. Compared to CM, AM patients harboured fewer BRAF mutations and more KIT mutations. The higher incidence of BRAF mutations in patients with CM partially explains the higher OS in this group since BRAF mutated patients are eligible for BRAF or BRAF/MEK inhibitors after progressing on ICI treatment [25]. Fifty percent of patients with CM receiving a second treatment line were treated with either BRAF or BRAF/MEK inhibitors versus fourteen percent of patients with AM. Earlier studies have reported a higher incidence of KIT mutations in AM [26]. Hode et al. [27] did not find a response to KITinhibitors in AM, but a systematic review by Steeb et al. [28] found an ORR of 22% for AM. In the DMTR, KIT-inhibitors are listed as other systemic therapy. Therefore, we could not retrieve the exact number of patients treated with KIT-inhibitors.
Our data are in contrast to the results of the phase II CheckMate 172 study by Nathan et al. [29] which included 55 patients with AM treated with nivolumab after progression on ipilimumab. In this study, a median OS of 25.8 months for patients with AM was reported, which was similar to patients with CM (25.3 months) and compares very favourably to our results. A recent study by Nakamura et al. [10] investigated the response to anti-PD-1 in any line in 193 Japanese patients. They reported an ORR of 16.6%, which is lower than the ORR we found in our cohort. They found a PFS of 3.5 months and an OS of 18.2 months, comparable to the PFS of 3.1 months and OS of 18.6 months for anti-PD-1 monotherapy in our study. The variation in response could be due to differences in the study design but are more likely to reflect differences in ICI effectiveness in Asian and Caucasian populations. Shoushtari et al. [30] investigated the effectiveness of ICIs in acral and mucosal melanoma in patients with advanced melanoma, of which most already received prior treatment. This study included 25 patients with AM treated with nivolumab or pembrolizumab. Patients with AM had an ORR of 32%, a median PFS of 4.1 months and a median OS of 31.7 months. The reported ORR and PFS of AM were comparable to our cohort. However, we found a shorter median OS of 18.6 months for anti-PD-1 treated patients. A single-centre cohort study by Rose et al. [31] included 230 patients with advanced melanoma treated with anti-PD-1 AE anti-CTLA4. Their cohort included 18 AM patients, of whom 11 were treated with anti-PD-1 monotherapy and 7 with ipilimumab-nivolumab. They found a median PFS of 3.5 months and a median OS of 14.6 months, similar to our findings. Zheng et al. [32] recently published a systematic review, including mostly Asian studies, investigating anti-PD-1 monotherapy in advanced melanoma. They included 12 studies with a total of 494 patients with AM treated with anti-PD-1, reporting an ORR ranging from 14 to 40%. The median PFS ranged from 3.2 to 9.2 months, and the median OS was over 14 months.
The present study does have some limitations. The use of observational data can cause bias by indication. Additionally, due to the retrospective analysis of our study, we cannot rule out the presence of residual confounding. The DMTR does not contain information about the ethnic background of patients, so it is unknown what proportion of our treatment cohort consists of patients with a non-Caucasian background.
A strength of our study is that we report on the largest nationwide cohort of Western AM patients. DMTR data registration is performed by independent data managers who are annually trained. The patients' treating oncologist checks the registered data to ensure data quality. The online registry in which patients are registered also includes warnings on inconsistent or missing data. Our study confirms decreased effectiveness of anti-PD-1 monotherapy as well as ipilimumab-nivolumab for advanced AM. Future studies should focus on alternative treatment strategies for this subgroup of patients with melanoma.

Conflict of interest statement
The authors declare the following financial interests/ personal relationships which may be considered as potential competing interests: AvdE has advisory relationships with Amgen, Bristol Myers Squibb, Roche, Novartis, MSD, Pierre Fabre, Sanofi, Pfizer, Ipsen, Merck and has received research study grants not related to this paper from Sanofi, Roche, Bristol Myers Squibb, Idera and TEVA and has received travel expenses from MSD Oncology, Roche, Pfizer and Sanofi and has received speaker honoraria from BMS and Novartis. JH has advisory relationships with Achilles Thera-