Enrollment Trends Among Patients with Melanoma Brain Metastasis in Active Clinical Trials

Abstract The CNS is a common site for distant metastasis and treatment failure in melanoma patients. This study aimed to evaluate the inclusion rate of patients with melanoma brain metastases (MBM) in prospective clinical trials. 69.3% of trials excluded MBM patients based on their CNS disease. In univariate analysis, trials not employing immunotherapy (p = 0.0174), inclusion of leptomeningeal disease (p < 0.0001) and non-pharmaceutical sponsor trials (p = 0.0461) were more likely to enroll patients with MBM. Thoughtful reconsideration of clinical trial designs is needed to give patients with MBMs access to promising investigational agents and improve outcomes for patients with MBM.


Introduction
Melanoma is the fifth most common cancer in the United States (1).Melanoma brain metastasis (MBMs) are frequent in advanced melanoma patients, with an estimated incidence of 60% (2), and historically portend a poor prognosis with estimated 1-year survival of 10-20% and a median overall survival of 4 months (3,4).Approximately 20% of all melanoma patients already have brain metastases at the time of de novo metastatic diagnosis and 28% will eventually develop brain metastases (4-6).Autopsy series report the prevalence of brain metastases to be 50-75% in patients who died of melanoma further highlighting the propensity of melanoma to spread to the brain, a multifactorial association with complex underpinnings that have not been fully elucidated (7-9) Central nervous system (CNS) spread is a frequent site of systemic treatment failures in melanoma, owing to a hypothesized distinct intrinsic tumor biology of metastatic lesions that may then confer resistance to particular therapies (7, [10][11][12].Traditionally employed systemic therapies, such as cytotoxic chemotherapy or interferons for systemic control offered minimal efficacy in improving overall survival (OS) for patients with MBM, often due to challenges with blood-brain barrier penetration or intolerable toxicities (4, [13][14][15].While some studies have individually noted good local control for MBMs with approaches like stereotactic radiosurgery and surgical resection (15)(16)(17), a recent pooled analysis demonstrated a lack of evidence accurately comparing survival benefit outcomes, underscoring the need for more standardized trials to properly establish an evidence-based approach to management (18).Fortunately, the landscape of available systemic treatment options for metastatic melanoma has transformed over the past decade following the advent of targeted therapies (BRAF and MEK inhibitors) and immune checkpoint inhibitors (ipilimumab, nivolumab, pembrolizumab, relatlimab), which have yielded significant improvements in CNS response rates and OS in patients with MBMs (18)(19)(20)(21)(22)(23).
Historically, patients with MBMs have been excluded from clinical trials including the early trials of modern systemic therapies, likely due to uniform acceptance of a historically poor prognosis of this patient population or due to concerns regarding CSF penetration of systemic therapy (12,24).Early drug development trials have classically been structured to first evaluate clinical safety and efficacy primarily in patients with extracerebral advanced stage disease only.Similar paradigms were followed in the development of BRAF/MEK inhibitor targeted therapy, as well as immune check-point inhibitor therapy in melanoma.Such linear approach to early drug development can lead to a delayed investigation and evaluation of benefit of viable therapies for MBM patients, as exemplified in the pivotal studies that led to FDA-approval of the modern day aforementioned immune and targeted therapies for management of advanced melanoma (25,26).
Treatment paradigms have undoubtedly evolved in recent years which have led to some smaller MBM-specific trials and inclusion of subsets of patients with brain metastases in some early trials-albeit still contingent upon strict inclusion criteria.As discussion surrounding how to best standardize the incorporation of MBM patients early in the drug development process continues, it is imperative to evaluate where we stand currently with the representation of patients with brain metastases in prospective melanoma research.A source of consolidated evidence adequately capturing this does not presently exist.Our study therefore aimed to examine ongoing clinical trials of systemic therapies for advanced melanoma and establish the current state of MBM patient inclusion in ongoing studies.Certain characteristics of trial design and eligibility criteria were analyzed for associations predictive of CNS disease inclusion to elucidate broader patterns across studies, with an aim to identify areas of unmet need that could guide future clinical trial design and a framework for patient advocacy.

Methods
Trial data were collected from ClinicalTrials.gov on March 12, 2022, by using the primary search term "Melanoma".This search was further limited by the terms "recruiting" and "interventional" studies, as well as to adult (ages 18-64 years) and older adult (65þ) populations.The search terms were similar to those of a prior retrospective analysis of non-small cell lung cancer clinical trials (27).Trials were excluded if they were specific for other disease types, excluded advanced disease, or did not include a potentially active systemic anticancer pharmaceutical.We obtained the following information from each trial: (1) phase of clinical trial (I, I-II, II, III, or IV); (2) location (United States only, international only, or global); (3) melanoma-specific or multi-site malignancy; (4) melanoma subtype (cutaneous, uveal, mucosal, not specified); (5) anticancer therapy (targeted therapy, immunotherapy (IO), radiation, chemotherapy, combination therapy, or other); (6) sponsor type (pharmaceutical or non-pharmaceutical); (7) criteria for inclusion of CNS disease [MBM and/or leptomeningeal disease (LMD)] (always included, included only if no evidence of active disease/ symptoms, included if treated/untreated, strictly excluded, or not specified).Accuracy of trial data was confirmed by two independent investigators.
Inclusion of CNS disease was the primary end point of the analysis.Contingency tables were created to illustrate the frequency distributions of trial characteristics.Associations between trial characteristics and CNS inclusion were evaluated using Fisher's exact test.Inclusion of active CNS disease in trials was modeled using logistic regression, and univariate analyses only considered covariates with known endpoint information.Statistical analysis was performed using SPSS version 28 (IBM Corp, Armonk, New York, United States of America), and p-values less than 0.05 were considered statistically significant.

Results
A total of 475 trials were identified using the key search terms described above.387 trials met the inclusion criteria (Figure 1).Of the included trials, 147 were phase I, 86 were phase I/II, 130 were phase II, 3 were phase II/III, 19 were phase III, and 2 were phase IV (Table 1).199 studies (51.4%) were conducted exclusively in the United States (US), 118 (30.5%) were international, and 70 (18.1%)were global.A total of 204 (52.7%) trials focused on melanoma patients specifically, while 183 (47.3%) trials investigated patients with multiple malignancies including melanoma.Of the trials that were melanoma-specific, 17.3% included only cutaneous melanoma, 12.4% included cutaneous or mucosal melanoma patients, 3.1% included only mucosal melanoma, 5.7% included only uveal melanoma, 4.7% included all types of melanomas, and 56.8% did not specify melanoma sub-type.Treatment modalities included -targeted therapy in 19.4% trials, immune therapy (IO) in 53.5%, radiation in 2.6%, chemotherapy in 1.8%, combination therapy in 20.9%, and other treatment modalities in 1.8% trials.Pharmaceutical companies sponsored 187 (48.3%) studies, while non-pharmaceutical sources sponsored 200 (51.7%).
Trial inclusion of patients with MBM was reported based on disease activity, symptoms, and treatment status.Overall, 268 trials (69.3%) excluded patients solely based on CNS disease status (active, symptomatic, symptomatic and active, and treated only), and 50 (12.9%)trials did not specify whether patients with CNS disease were included (Figure 2).More specifically, any form of CNS disease was excluded in 69 tri-   2).There were no statistically significant associations between the odds of a study including patients with active MBM and trial characteristics such as study phase, location, melanoma-specific site, or sponsor.After removing other variables that were not significant at the  0.05 level, only inclusion of leptomeningeal disease remained significant at the 0.05 level for the final model therefore no multivariate analysis was performed (p < 0.0001).Given that only one variable was found to be significant in univariate analyses, no multivariate analysis was performed.

Discussion
Expanding prospective clinical trial eligibility for patients with brain metastases has been recognized as a priority in oncology, especially in the settings of melanoma, breast cancer, and lung cancer where CNS metastasis is a common cause of disease related morbidity and entails a poor prognosis (12,28).The goal of this study was to methodically outline the current status of clinical trial enrollment criteria for MBM patients in interventional systemic therapy clinical trials.We examined ongoing clinical trials of systemic therapies for melanoma to assess patterns of inclusion of patients with CNS metastasis.Our analyses revealed the vast majority of ongoing clinical trials (69.3%) continue to exclude patients with melanoma CNS metastasis.Furthermore, nearly all trials that allowed MBM study participation required stable or asymptomatic CNS disease and only roughly half allowed enrollment of patients who had undergone prior CNS-directed therapy for their MBMs.In the present study, 14% of trials allowed participation for patients with treated CNS disease only, which is lower than the corresponding 41% inclusion rate across non-small cell lung cancer (NSCLC) trials.To compare to metastatic breast cancer (MBC), Costa et al reported a similar trend and found that 81% of 1,474 early phase trials published between the years of 1992 and 2016 excluded patients with CNS involvement or simply did not document inclusion (52%) criteria (29).Historical exclusion of patients with CNS disease was often attributed to unknown pharmaco-kinetic/-dynamic properties of investigational agents, uniformly accepted poor prognosis of patients, or lack of tools to assess CNS response criteria.Research has since been undertaken to better define CNS response criteria, elucidating key molecular pathways of efficacy and CSF penetration and reinvigorating hope for inclusion of patients with CNS metastasis.More recent studies examining enrollment trends in lung cancer trials has shown trends towards lower rates of strict CNS metastasis exclusion (24,27).Ongoing efforts of advocacy groups, investigators, and organizational support groups, such as the US Food and Drug Administration and ASCO/Friends of Cancer Research/FDA working groups, will be pivotal in expanding clinical trial enrollment for patients with solid tumor brain metastasis (10,22).
A much smaller percentage of trials (4.4%) allowed for leptomeningeal disease, possibly reflecting a scarcity of proven treatments and general difficulty with assessing treatment efficacy, often deemed "non-measurable" per RECIST 1.1 guidelines.Notably, the inclusion of LMD in trial eligibility was significantly associated with the inclusion of patients with active CNS disease.Given the close association between LMD and CNS parenchyma metastasis (30,31), it is conceivable that investigators specifically exploring the CNS efficacy of investigational drugs are more likely to include LMD patients, but only a handful of ongoing targeted and immunotherapy trials are allowing inclusion of patients with LMD, due to predefined restrictive exclusion criteria.
We observed that non-pharmaceutical company sponsored studies had greater odds of including patients with active CNS disease.This agrees with previous findings demonstrating industry-sponsored trials in general have higher rates of strict exclusion of patients with brain metastasis (28,32).Seeking trial funding from diverse sources, including philanthropic resources may be needed to boost enrollment of patients with active CNS disease.Additionally, should companies be concerned that poor outcomes in MBM patients may lead to poorer perception of their medication amongst practitioners, it is reasonable to include MBM patients but have them analyzed separately due to intrinsic differences between CNS and non-CNS disease.It should be noted however that the significance of sponsor did not remain true once all non-significant variables were utilized.
In our univariate analyses, melanoma trials utilizing non-immunotherapy agents were more likely to include active CNS disease compared to immunotherapy-alone trials.This is comparable to findings in NSCLC trials, where immunotherapy investigational agent studies had higher odds of strict CNS disease exclusion (28).While risks of toxicity and adverse neuro-inflammatory events with newer immunotherapies are certainly relevant to consider for safe MBM patient enrollment, they have not proven to be of significant concern in most trials to date, which have instead demonstrated improved intracranial disease response rates and patient outcomes.For example, a follow up of the Checkmate 204 trial showed a 3-year intracranial progression free survival of 54% in asymptomatic MBM and 3-year overall survival of 72% and 37% for patients with asymptomatic and symptomatic MBM, respectively (33).A more nuanced approach to trial design rather than strict exclusion should therefore be encouraged.It should be noted however that like pharmaceutical sponsor, the significance of IO therapy vs non IO therapy diminished once all non-significant variables were utilized.
Our study has several limitations.First, data was solely collected from the ClinicalTrials.govregistry.Data accrual and review of each trial were therefore dependent upon publicly entered and available information.Furthermore, since this study sought to record numerous characteristics of trial design, it is reasonable that not all variables would be reported for all included studies, limiting our ability to perform a full univariate analysis without exclusions.Additionally, since we aimed to assess the present status of inclusion in clinical trials, we did not subcategorize trials by year of commencement at the time of query.As a result, any temporal trends in trial design across studies were not assessed.Drug target-specific trends were also not captured because of unknown or potentially overlapping spectrum of investigational agents and multi-drug trials.
The findings of this study reveal that despite advances in therapeutic approaches for MBM, clinical trials have not simultaneously adapted to expand inclusion criteria or modify trial design to expand clinical trial enrollment options for patients with MBMs.Consequently, patients with MBMs are drastically underrepresented in current clinical trials unlike other similar disease states with common metastases to the brain such as NSCLC and metastatic breast cancer.Analogous to the long lag that occurred between initial approval and subsequent confirmation of CNS efficacy for currently FDA approved systemic therapies for advanced melanoma, promising therapies currently under investigation will continue to experience such delays in evaluation of CNS efficacy unless efforts are made to boost inclusion of MBM patients in early phase investigational studies.Approaches such as inclusion of patients with MBMs, provision of simultaneous cohorts with and without known CNS metastasis with adoption of appropriately standardized response criteria for each cohort should be encouraged to expand clinical trial availability and streamline drug development processes for these patients (26,27).
We do however acknowledge that broad inclusions of all patients with MBM is also not the answer.Therapeutic drugs come with substantial treatment related adverse events and should not entirely replace current standard of care measures for the treatment of MBM patients before being clinically validated.Despite this, progress cannot be made in determining alternative treatment modalities for patients with MBM without protocolized investigations and trialing new therapies.Possible modalities to include more MBM patients include using new treatment options in conjunction with standard of care locoregional therapies or in patients without alternative treatment options who have already progressed on standard of care therapeutics.

Conclusion
This study found that the majority of the ongoing clinical trials evaluating systemic therapies in melanoma still exclude patients with CNS metastases.Future studies, like the present, should be conducted in 5 to 10 years to identify updated trends and assess whether progress in has been made in boosting clinical trial participation for patients with MBM early on in drug development process.

Figure 1 .
Figure 1.Consort diagram depicting the criteria used to identify trials used in this analysis.

Figure 2 .
Figure 2. Histogram depicting rate of inclusion of CNS disease among studied trials.

Table 1 .
Clinical trial characteristics recorded from ClinicalTrials.gov.from trials within this study.

Table 2 .
Univariate analysis evaluating associations between trial characteristics and CNS inclusion and odds ratio estimates of active melanoma CNS metastasis inclusion.
Odds Ratio Estimates for modeling trial inclusion of patients with active melanoma brain metastases