The efficacy of immune checkpoint blockade for melanoma in-transit with or without nodal metastases e A multicenter cohort study

Purpose: Guidelines addressing melanoma in-transit metastasis (ITM) recommend immune checkpoint inhibitors (ICI) as a first-line treatment option, despite the fact that there are no efficacy data available from prospective trials for exclusively ITM disease. The study aims to analyze the outcome of patients with ITM treated with ICI based on data from a large cohort of patients treated at international referral clinics. Methods: A multicenter retrospective cohort study of patients treated between January 2015 and December 2020 from Australia, Europe, and the USA, evaluating treatment with ICI for ITM with or without nodal involvement (AJCC8 N1c, N2c, and N3c) and without distant disease (M0). Treatment was with PD-1 inhibitor (nivolumab or pembrolizumab) and/or CTLA-4 inhibitor (ipilimumab). The response was evaluated according to the RECIST criteria modified for cutaneous lesions. Results: A total of 287 patients from 21 institutions in eight countries were included. Immunotherapy was first-line treatment in 64 (22%) patients. PD-1 or CTLA-4 inhibitor monotherapy was given in 233 (81%) and 23 (8%) patients, respectively, while 31 (11%) received both in combination. The overall response rate was 56%, complete response (CR) rate was 36%, and progressive disease (PD) rate was 32%. Median PFS was ten months (95% CI 7.4–12.6 months) with a one-, two-, and five-year PFS rate of 48%, 33%, and 18%, respectively. Median MSS was not reached, and the one-, two-, and five-year MSS rates were 95%, 83%, and 71%, respectively. Conclusion: Systemic immunotherapy is an effective treatment for melanoma ITM. Future studies should evaluate the role of systemic immunotherapy in the context of multimodality therapy, including locoregional treatments such as surgery, intralesional therapy, and regional therapies.


Introduction
Approximately, 5-10% of patients with high-risk cutaneous melanoma will develop intransit metastases (ITM) [1], a form of tumor spread within the lymphatic channels between the primary tumor site and regional lymph nodes [2]. According to the American Joint Committee on Cancer (AJCC) staging system, patients with ITM are considered to have stage III disease with a corresponding high risk of both locoregional and systemic recurrence [3]. When there are few lesions and all can be removed with an excision, surgical resection is recommended, but for patients with unresectable ITM (e.g., multiple, bulky, or quickly recurrent), there are numerous locoregional treatment options available. These include isolated limb perfusion (ILP) [4], isolated limb infusion (ILI) [5], intralesional injections with talimogene laherparepvec (TVEC) [6] and electrochemotherapy [7]. These locoregional therapies offer very high overall response rates (ORR) (70-90%) and complete response (CR) rates (50-70%). The majority of the literature on locoregional treatments are retrospective series, with few prospective clinical trials, and there are no studies directly comparing these locoregional treatments with each other or to systemic immunotherapy [8].
There have been major advances recently in the treatment of advanced melanoma with the introduction of effective systemic treatments, including targeted therapies with BRAF-MEK inhibitors [9][10][11] and immune checkpoint inhibitors (ICI) [12][13][14]. These agents have proven to be effective in unresectable disease as well as in the adjuvant and neoadjuvant settings. Many guidelines addressing ITM recommend ICI as an option for first-line treatment, despite the fact that there are no data available from prospective trials on the efficacy of these agents for ITM. When the pivotal registration trials of PD-1 inhibitors for patients with advanced melanoma were analyzed, no patients with exclusively ITM were identified [15]. Two small retrospective analyses of patients with ITM treated with systemic immunotherapy, showed an ORR of 31-54% with a CR rate of 13-26% [16,17].
The aim of this study was to analyze the outcome of patients with ITM treated with immune checkpoint blockade using PD-1 inhibitors and/or a CTLA-4 inhibitor based on real-world data from a large international cohort of patients from melanoma centers in Australia, Europe, and the USA.

Study design
A multicenter retrospective cohort study that included patients treated between January 2015 and December 2020. All participating study sites are national or regional referral clinics for the treatment of melanoma. The study was conducted in adherence to the ethical principles of the Declaration of Helsinki, Good Clinical Practice (GCP) and ICH Guidelines with the approval of the Swedish Ethical Review Authority (477-18 and 2021-02315). Sahlgrenska University Hospital served as the coordinating center and participating institutions obtained appropriate ethical approval and performed independent data extraction and provided data to the coordinating center in compliance with individual institutional requirements and negotiated data use agreements.

Study cohort
This study included cutaneous melanoma patients treated with ICIs for in-transit melanoma metastases (AJCC8 stages IIIB-D), with or without concurrent nodal involvement (AJCC8 N1c, N2c, and N3c). Patients were included if they were treated with a PD-1 inhibitor (nivolumab or pembrolizumab) and/or CTLA-4 inhibitor (ipilimumab); with or without previous treatment for in-transit metastases, e.g., surgery, locoregional, and systemic treatments. Patients having received concomitant therapies parallel to ICI treatment were excluded. All in-transit lesions were required to be evaluable by RECIST criteria, but modified so that lesions not visible by radiology were measured by calipers instead. Patients with previous or synchronous distant metastases (stage IV disease) were excluded.

Outcomes
The primary endpoint was CR rate, with best overall response, time to local progression (TTLP), time to nodal progression (TTNP), time to systemic progression (TTSP), progression-free survival (PFS), melanomaspecific survival (MSS), and overall survival (OS) as secondary endpoints. MSS and OS were calculated from the start of immunotherapy to death or end of follow-up, and PFS was calculated from the start of immunotherapy to progression or death. The response was evaluated as the best response during follow-up according to the RECIST criteria modified for cutaneous lesions (allowing for caliper measurement if lesions were not visible on radiology). To be considered a CR, all lesions had to disappear. Partial response (PR) was defined as a decrease of more than 30% of the total tumor burden, measured as a number of lesions or shrinkage in the largest tumor diameter. Progressive disease (PD) was defined as an increase of more than 25% in existing lesions or the appearance of new lesions. Stable disease (SD) was defined as when criteria for CR, PR, or PD were not met.

Statistical analysis
Survival was estimated using the Kaplan-Meier method and analyzed using the log-rank test. Adjusted analyses were performed using binary logistic regression for CR and Cox regression for TTLP, PFS, and MSS, with results reported as hazard ratios (HR) with 95% confidence intervals (CIs) and p-values. Statistical significance was set at p < 0.05. Covariates for adjustment were selected a priori (age, gender, performance status, geographic region, T-status, mutational status, having received adjuvant therapy, anatomical site of ITM, number of ITM, size of largest ITM>Z30 mm, lymph node status, and having received previous treatment for ITM). Statistical analysis was performed using SPSS version 24.0 (SPSS Inc., Chicago, IL, USA).

Results
A total of 287 patients with melanoma ITM from 21 institutions in eight different countries treated with immunotherapy were included in the analysis (Supplementary Table 1). Overall, 161 patients (56%) were male, and the median age at the start of ICI was 69 years (IQR 59-78). The median time from melanoma diagnosis to the first in-transit recurrence was 13 months (IQR 5-30). The most common site for in-transit metastasis was the lower limb, both for females (95/124, 77%) and males (74/157, 47%), and 127 (44%) had concomitant lymph node metastasis. A BRAF mutation was present in 84 (29%) and an NRAS mutation in 57 (20%) of the tumors. The most common histological subtypes of the primary melanomas were nodular (35%) and superficial spreading (32%), while acral lentiginous was present in only 8% of the patients. Ulceration was present in 122 (42.5%) of primary tumors (Table 1). Data on disease stage at diagnosis of the primary melanoma were available for 88% (n = 253) of patients, with 1% (n = 2) stage 0, 9% (n = 25) stage I, 24% (n = 60) stage II, and 66% (n = 166) stage III.
The median time to local progression was 12 months (95% CI 6.0−18.0) and occurred in 160 patients (56%). There was a significant difference in time to local progression depending on response, where the median time was not met for CR, and for patients with a PR or SD, 23 and 21 months, respectively (p < 0.001) (Fig. 1). A total of 72 patients (25%) developed regional nodal progression after a median time of five months (95% CI 2.9−7.1). A total of 89 patients (31%) progressed with distant metastases after a median time of nine months (95% CI 6.7−11.3).
Out of the 104 patients that experienced an initial CR, 24 patients (23%) later progressed with further ITM after a median time of ten months, ten patients (10%) progressed with lymph node metastases after a median time of seven months, and 16 patients (15%) progressed with distant metastases after a median time of ten months ( Table 2).
The median PFS from the start of ICI was ten months (95% CI 7.5−12.5 months) with a one-, two-, and five-year PFS rate of 47%, 33%, and 19%, respectively. When dividing PFS based on the type of immunotherapy received, there was no significant difference with a median PFS of ten months for single-agent PD-1 inhibitor, nine months for single-agent CTLA-4 inhibitor, and nine months for the combination treatment of PD-1 and CTLA-4 inhibitor (p = 0.77) (Fig. 2). In multivariate analysis, only ECOG performance status 2 (HR 2.16, 95% CI 1.04−4.50, p = 0.04) was found to be predictive for PFS (Supplementary Table  2).

Discussion
This is the largest pooled multi-institutional analysis to date presenting real-world (outside of clinical trials) outcomes for patients with in-transit metastasis treated with systemic immunotherapy. The topic is clinically challenging and a pooled analysis is timely, particularly as patients with ITM only were not included as a subgroup for analysis in the previous registration trials for unresectable stage III or stage IV melanoma [15]. We report a CR rate of 36%, which is higher than the 26% and 13% recently reported by Nan Tie et al. and Zaremba et al respectively [16,17]. Nan Tie et al. presented a retrospective analysis of 54 patients with in-transit metastasis treated with PD-1 inhibitor (nivolumab or pembrolizumab) CTLA-4 inhibitor (ipilimumab) at three tertiary hospitals in Australia. Their cohort (partly included in this study) is comparable to ours in patient characteristics and treatment regimes, except that our study includes more patients with ulcerated primary tumors and more patients with concurrent nodal metastasis at the start of immunotherapy. We report a higher one-and two-year OS, while ORR and PFS are comparable.
More recently, Zaremba et al. retrospectively reported on 191 patients with unresectable stage IIIB-D in-transit and satellite metastases from 16 centers in Germany. Notably, they excluded patients who had received prior treatments followed by adjuvant immunotherapy and only included patients receiving PD-1 ±CTLA4-inhibitor as first-line treatment. Patient characteristics were, again, roughly comparable to this cohort, but the overall CR rate for patients receiving PD-1 inhibitor was only 13% [17].
Our CR rate is also higher than the response data from the registration trials investigating immunotherapy for unresectable stage III and stage IV patients. The CR rates in these registration trials range from a CR rate of 2% for CTLA4-inhibitor monotherapy, up to 22% for combination PD-1 and CTLA4-inhibitors [12,14,[18][19][20][21][22][23]. Though the direct comparison is difficult, we also see a significantly higher MSS in our CR patient group. We also see an unexpectedly high CR rate of 30% for single-agent CTLA4 in this study. This is significantly higher than reported from other cohorts and is possibly an overestimate due to the small sample size of only twenty-three patients in this subgroup. However, it should be noted that direct comparisons of response rates are complicated by the lack of a standardized method of judging response in cutaneous lesions. We used RECIST criteria for the definition of response, with the modification that the lesions did not have to be visible on radiology, but could be measured by calipers. This is a strength, since the response rates can be compared to other publications, and using calipers is the most common way in-transit lesions are measured. Of note, how the response was evaluated was not stated at all in the Zaremba manuscript [17], and the Nan Tie manuscript [16] used the WHO criteria for PR (50% cut-off). In addition, any biological particularities of in-transit metastases underlying response and progression patterns are currently unknown.
Other treatment options for patients with ITM include, e.g., TVEC, ILI, or ILP, with reported CR rates of up to 62%, 38%, and 65%, respectively [4,[24][25][26][27]. Owing to the heterogeneous follow-up between these different studies, any comparison of response duration is not possible. In the current cohort, there was no benefit of having received locoregional treatments prior to immunotherapy. This finding can be the result of a selection bias due to our study design, where only patients progressing after a previous locoregional treatment were included. Patients with limited and resectable in-transit metastases may undergo a resection of the metastases followed by adjuvant systemic therapy, supported by data from the CheckMate-238 study showing a benefit in four-year RFS favoring nivolumab over ipilimumab with a HR of 0.63.
Another important finding was the rate of patients with no response and primary resistance; the rate of PD as the best response in this real-world cohort was 32% compared to 30% in Nan Tie et al. [16] and 43% in Zaremba et al. [17]. These are all higher PD rates than reported for locoregional therapies, e.g., 3% for ILP [4], 7% for ILI [5], 18% for TVEC [28] and 3% for electrochemotherapy [29]. We, therefore, hypothesize that for patients with multiple, bulky, or quickly recurrent ITM, a combination of locoregional with systemic treatment might be appealing. This strategy might combine the very high ORR and low PD rates of T-VEC, ILP, or ILI, together with a systemic treatment that would also treat undetected micrometastatic disease outside of the area treated with locoregional therapy.
Hypothetically, the combination of locoregional and systemic treatment might not just have effects independent of each other but might also act in synergy. It has, for example, been shown that ILP leads to an activation of tumor-specific T-cells, which potentially can be enhanced by the addition of a PD-1 inhibitor [30,31]. A phase I study of TVEC together with pembrolizumab showed a potential synergistic effect [32], and the following placebocontrolled Masterkey-265 phase III trial showed an increased response rate (18% vs. 12%), however, without any benefit in PFS or OS after a median of 31 months of follow-up [33]. On the other hand, TVEC has been shown to be an effective salvage therapy with ORR of 51% and CR rates of 37% after the failure of systemic immunotherapy for ITM [34]. Another phase III trial (ILLUMINATE-301), examined the synergistic effect of intratumoral injections of the Toll-like receptor 9 agonist tilsotolimod together with ipilimumab, however, there was no difference in ORR compared to ipilimumab alone, 8.8% vs 8.6% (data not published). In summary, the optimal combination and sequence of these different treatment modalities for patients with ITM needs further clarification through prospective trials. Two examples of ongoing studies are the Nivo-ILP (ClinicalTrials.gov: NCT03685890) and NIVEC (ClinicalTrials.gov: NCT04330430) studies, which are examining the combination of ILP and nivolumab and T-VEC and nivolumab for patients with ITM, respectively.
We demonstrate a five-year OS of 63%, and in comparison, 44% for patients treated with ILI as shown by Kroon [4][5][6]14]. However, any direct comparisons are not possible as some of these studies were undertaken before effective systemic therapies were available, and the patient selection criteria/baseline characteristics, patient prognosis, methodologies, and follow-up times are also very different.
The relatively low rate of BRAF positive patients in the current cohort, 32% (84/266) when excluding missing values, can be compared to 42% (71/170) in the material by Zaremba et al., and 19% (10/52) in the material by Nan Tie et al. Hypothetically, patients developing in-transit metastasis might have a different proportion of BRAF mutations, or there may be a selection bias where patients with a BRAF-mutation to a larger extent have received targeted therapies rather than immunotherapy. Unfortunately, based on the current data, we cannot draw any conclusions concerning this, but it is indeed a question that warrants further investigation.
A strength of this study is the large, international, and multi-center cohort. This results in what we believe to be an accurate overview of current practice and outcomes internationally. There were regional differences in treatment regimens when comparing centers by geographical location, and it is very possible that these differences reflect dissimilarities in either treatment availability or management approach between countries. Further, the study is limited by the retrospective design, inherent selection bias, and relatively small subgroups. Also, as has been noted previously, measurement of treatment response in this patient group is difficult as the disease in this stage may not be accurately evaluable by RECIST [35] or WHO [36] criteria. Many institutions report using RECIST criteria modified for cutaneous lesions, but do not specify the modifications further. In our view, there is currently a lack of objective and standardized criteria by which to assess in-transit lesions sufficiently accurately, and future research should be focused on this aspect.

Conclusions
In summary, systemic immunotherapy with checkpoint inhibition is an effective treatment for melanoma in-transit metastases. Future studies should be conducted to establish not only the optimal sequencing, timing, and role of surgery and/or other locoregional therapies in combination with systemic immunotherapy but also the mechanisms of clinical immune responses making such treatments effective.

Supplementary Material
Refer to Web version on PubMed Central for supplementary material. Local progression-free survival (in-transit metastasis) based on response of treatment from start of immunotherapy. Melanoma-specific survival by best response from the start of ICI.  Table 1 Baseline patient characteristics at start of ICI for ITM.   Table 2 Response rate based on the type of treatment and subsequent disease progression.   Table 3 Logistic regression analysis for complete response.