Leptomeningeal disease in melanoma: an update on the developments in pathophysiology and clinical care

Abstract

Leptomeningeal disease (LMD) remains a major challenge in the clinical management of metastatic melanoma patients. Outcomes for patient remain poor, and patients with LMD continue to be excluded from almost all clinical trials. However, recent trials have demonstrated the feasibility of conducting prospective clinical trials in these patients. Further, new insights into the pathophysiology of LMD are identifying rational new therapeutic strategies. Here we present recent advances in the understanding of, and treatment options for, LMD from metastatic melanoma. We also annotate key areas of future focus to accelerate progress for this challenging but emerging field.

Introduction

Multiple therapeutic advances have substantially improved the prognosis for patients with metastatic melanoma. However, metastasis to the central nervous system (CNS) remains a common, morbid, and highly lethal challenge. Leptomeningeal disease (LMD), which is diagnosed in 5-8% of metastatic melanoma patients (up to 20% at autopsy), is the rarest but most deadly manifestation of melanoma in the CNS (Harstad et al., 2008; Le Rhun et al., 2013; Leal et al., 2011; Pape et al., 2012; Raizer et al., 2008; Wolf et al., 2017). LMD is characterized by the spread of tumor cells to the cerebrospinal fluid (CSF) space and/or the leptomeninges, which are the membranes covering the brain and spinal cord (Leal et al., 2011). Many attribute the first report of LMD to Karl Joseph Eberth in the 1870’s, who wrote a case report about a 47-year-old woman who was admitted to an insane asylum for “nonsense” only to be later diagnosed with chronic hydrocephalus by Eberth and moved to a medical facility (Eberth, 1869). In relation to the etiology of the woman’s affliction, Eberth simply offered that “it should be mentioned that the patient is said to have had a glass of schnapps here and there and that she lived in very poor circumstances.” Following debilitating disease progression, her eventual death, and post-mortem investigation, Eberth provided one of the first detailed pathological descriptions of LMD (in this case associated with lung cancer), acknowledging how very little they knew about this phenomenon despite the published investigations from Rudolf Virchow, who wrote a series of case reports describing LMD some 20 years previously (Virchow, 1855). It has now been nearly 170 years since Virchow’s initial investigations, and even today, we can mirror Eberth’s statement: “Yes, one can boldly assert that since Virchow's investigations in 1855 our knowledge of this new formation has not undergone any appreciable enrichment.” Still, very little is known about the biology of LMD, diagnosis remains a challenge, and efficacious therapies remain limited. LMD patients may progress rapidly, even in the setting of active systemic therapy (Virchow, 1855). This clinical course often results in significant neurological morbidity and, ultimately, a dismal prognosis, with survival generally measured in mere weeks.

Since 2015, the Melanoma Research Foundation has been active in bringing melanoma researchers together to focus on the biology and treatment of melanoma brain metastases (MBM). Similar to the recent report summarizing the current state of research and management for MBMs, here we present an update on the most recent developments in the understanding and treatment of melanoma LMD.

The current state of diagnosis and prognosis

Establishing the diagnosis of LMD remains a challenging clinical issue. The “gold standard” for the diagnosis of melanoma LMD is positive CSF cytology, which can be falsely negative in 50% of the cases (Le Rhun et al., 2013). However, the CSF of LMD patients often displays other abnormalities, including elevated protein and WBC levels, even in the absence of tumor cells on cytopathology (Bönig et al., 2019; Murray et al., 1983). In the absence of positive cytology, LMD diagnosis may be determined through the detection of leptomeningeal enhancement on MRI imaging of the neuroaxis (Glitza et al., 2020; Leal et al., 2011). However, MRI can also be false positive, with meninges enhancing from other processes such as encephalitis, hemorrhage or chemical irritation by intrathecal chemotherapy (Gil et al., 2016; Paakko et al., 1990; Straathof et al., 1999; Sze et al., 1989). Additionally, elevated intracranial pressure (detected with lumbar puncture) may indicate alterations in CSF flow dynamics that can also be associated with LMD. (Le Rhun et al., 2013; Taillibert & Chamberlain, 2018) Prompt diagnosis of LMD is critical in order to initiate therapy, hence reliable tools to assist with diagnosis can potentially improve outcome. As such, several liquid biopsy approaches are currently being developed to aid in LMD diagnosis and to interrogate the biology of the disease, including the quantification of circulating tumor cells (CTCs) and the assessment of cell-free (cf)DNA in the CSF (Bale et al., 2021; Fedorenko et al., 2016; Li et al., 2016; Lin et al., 2017; Melms et al., 2018; Smalley et al., 2020; Wijetunga et al., 2021).

i. Circulating tumor cells

A case report of LMD in a patient with uveal melanoma demonstrates the ability to identify and quantify CSF-CTCs based on an anti-high-molecular-weight melanoma-associated antigen-positive and CD45/CD34-negative staining pattern using the Veridex CellSearch platform(Fedorenko et al., 2016). The largest study utilizing quantification of CSF-CTCs using the CellSearch platform for the diagnosis of LMD from epithelial tumors demonstrated a sensitivity of 93%, specificity of 95%, positive predictive value 90%, and negative predictive value 97% using the EPCAM+ and CD45-negative 1 CSF-CTC/mL cutoff (Lin et al., 2017). Another group examined the quantification of CSF-CTCs following proton craniospinal irradiation therapy and found that the number of pre-treatment CSF-CTCs predicted response to therapy, and improved survival was observed in patients with lower pre-therapy CSF-CTC counts (Wijetunga et al., 2021).

ii. Cell-free DNA

Cf DNA is a fragment of DNA released after cell apoptosis that carries genome-wide DNA information. Circulating tumor DNA (ctDNA) is a subset of cfDNA which can be derived from malignances. Studies across several pathologies have reported using ctDNA (in blood) to aid in early diagnosis, help predict recurrence, and monitor treatment response for systemic (non-CNS) metastatic disease (Bettegowda et al., 2014; Newman et al., 2014; Volik et al., 2016). Additional recent studies have indicated that ctDNA may also improve the diagnostic yield of CSF for LMD. Zhao et al., highlighted this in a study comparing the sensitivity of CNS imaging, CSF cytology, and CSF ctDNA for the diagnosis of LMD across several solid tumor types. The authors reported that 71% and 63% of LMD cases had positive cytology and CNS imaging, respectively. However, next-generation sequencing of CSF ctDNA detected cancer-associated mutations in 100% of cases (Zhao et al., 2019). Another study including 30 patients with LMD mostly from solid malignancies (2 patients with melanoma) reported improved sensitivity and accuracy of ctDNA analysis over CSF cytology (White et al., 2021). The authors did note 3 false positive results from patients with brain parenchymal tumors abutting CSF, suggesting that CSF ctDNA analysis should not be used for diagnosis of LMD in these patients (White et al., 2021). In reference to melanoma specifically, studies have reported utilizing ctDNA from CSF as a diagnostic tool for LMD. In their evaluation of seven melanoma LMD patients utilizing droplet-digital PCR and next-generation sequencing of ctDNA in CSF, Ballester et al., demonstrated that ctDNA in CSF may have superior sensitivity to CSF cytology and be able to detect melanoma-associated mutations (Ballester et al., 2018). In a rare case report of a primary leptomeningeal melanoma, clinicians were able to utilize ctDNA analysis to guide diagnosis and treatment decisions by identifying the targetable BRAF^V600E mutation in CSF (Melms et al., 2018). Another group demonstrated that molecular profiling of cell-free DNA was superior to cell-pellet genomic DNA from CSF, with analysis of ctDNA picking up at least one tumor-associated mutation more than twice as frequently as analysis of the cell pellet DNA (Bale et al., 2021). The authors noted that 44.6% of the samples which were positive for mutation using ctDNA analysis were then negative using cell pellet DNA or failed sequencing (Bale et al., 2021). ctDNA may also be used to track clinical progression of melanoma LMD. Li et al evaluated CSF specimens for BRAF^V600E mutations and demonstrated that the fraction of mutant DNA corresponded well with the patient's clinical response. Further, the fraction of mutant DNA decreased during a period of clinical improvement, then returned to high levels upon relapse of symptoms (Li et al., 2016). Overall, these studies highlight the potential for utilizing ctDNA in CSF for augmenting CSF diagnostic yield, identifying actionable targets, and potentially monitoring tumor progression/treatment responses.

In addition to analyzing the tumor cells or ctDNA in CSF from patients, there is some preclinical rationale for using mass-spectrometry analysis of CSF for diagnostic purposes. Smalley et al. showed that patients with melanoma LMD have a distinct proteomic profile in their CSF compared to CSF from patients without LMD (Smalley et al., 2020). Furthermore, the study demonstrates that proteomic profiling can be used to predict prognosis or monitor response. Proteomic signatures detected in the serial specimens of CSF from a rare patient who responded to therapy for over three years were distinct from those of patients with the typical short survival (Smalley et al., 2020). These proteomic signatures are described in more detail in section “Proteomic landscape of the CSF microenvironment in melanoma LMD” below. Overall, there is a growing effort to develop liquid biopsies for the diagnosis and prognosis of these aggressive tumors.

The Biology of melanoma LMD

The development of LMD

A full understanding of the pathogenesis of melanoma LMD remains elusive. However, there have been recent studies that have significantly added to our understanding on this topic. The typical scarcity of protein and metabolite products in healthy CSF would logically make it a very inhospitable place for energetically-demanding tumor cells to thrive (Fishman, 1992). One seminal study has uncovered the pivotal role of complement component 3 in facilitating cell invasion into the CSF space using mouse models of breast and lung cancer LMD (Boire et al., 2017). Complement C3 was found to be upregulated in mouse models of LMD and was shown to be necessary for tumor growth within the LMD microenvironment (Boire et al., 2017). In patients, it was also predictive of LMD progression (Boire et al., 2017). It has long been hypothesized that the choroid plexus, which is a specialized secretory epithelium structure in the ventricles responsible for secreting CSF and regulating the blood-CSF barrier, is a key entry point for tumor cells into the leptomeningeal space (Kokkoris, 1983). Mechanistically, C3 was found to activate the C3a receptor on the choroid plexus epithelium to disrupt the blood-CSF barrier and allow for additional mitogens to leak into the CSF environment to support tumor growth (Boire et al., 2017).

Metabolism of LMD

The metabolically harsh environment of the leptomeninges is evident both through the scarcity of micronutrients as well as the hypoxic environment of CSF (Fishman, 1992). It is therefore rational to expect that either the melanoma cells or the LMD microenvironment undergoes unique metabolic reprogramming to support rapid tumor growth. Chi et al have recently shown tumor cells to outcompete macrophages for iron by upregulating lipicalin-2, an iron scavenging protein, and its receptor SLC22A17 (Chi et al., 2020). Mass spectrometry analysis demonstrated an increase in total iron concentration and the proportion of iron bound to lipocalin-2 in CSF of LMD patients. Lipocalin-2 was shown to also support tumor cell growth in hypoxic conditions. Tumor growth supported by this mechanism was inhibited through iron chelation therapy, underscoring the potential for identifying metabolic vulnerabilities specific to melanoma LMD. Stemming from these discoveries, a new Phase I clinical trial (NCT05184816) has recently opened to test IT administration of iron chelation therapy deferoxamine in patients with LMD from solid tumors (Table 1). This is one of the first examples of the translation of research on LMD biology into a clinical trial for these patients.

Table 1:

Actively recruiting or soon to be recruiting clinical trials including melanoma patients with leptomeningeal metastasis

NCT#	PHASE	STUDY TITLE	INTERVENTIONS	Route of Administration	PRIMARY INVESTIGATOR STUDY DIRECTOR	LOCATION
NCT05598853	Phase I	Intrathecal Double Checkpoint Inhibition (IT-IO)	Nivolumab Ipilimumab	Intrathecal	Emilie Le Rhun, MD	University Hospital Geneva, Geneva, Switzerland Cantonal Hospital St Gallen, St Gallen, Switzerland University Hospital Zurich, Zurich, Switzerland
NCT03025256	Phase I/Ib	Intravenous and Intrathecal Nivolumab in Treating Patients With Leptomeningeal Disease	Nivolumab	Intravenous + Intrathecal	Isabella C. Glitza Oliva, MD	M D Anderson Cancer Center Houston, Texas, United States
NCT04543188	Phase I/Ib	A FIH Study of PF-07284890 in Participants With BRAF V600 Mutant Solid Tumors With and Without Brain Involvement	PF-07284890 Binimetinib Midazolam	Oral	Pfizer	72 study locations worldwide
NCT05026983	Phase II	Binimetinib and Encorafenib for the Treatment of Metastatic Melanoma and Central Nervous System Metastases	Binimetinib Encorafenib	Oral	Isabella C. Glitza Oliva, MD	M D Anderson Cancer Center Houston, Texas, United States
NCT05034497	Phase I	Intraventricular Administration of Rhenium-186 NanoLiposome for Leptomeningeal Metastases (ReSPECT-LM)	186RNL	Intraventricular	Marc Hedrick, MD	Universiy of Texas Southwestern Medical Center, Dallas, Texas, United States UT Health Science Center San Antonio/Mays Cancer Center, San Antonio, Texas, United States
NCT05184816	Phase I/Ib	A Study of Deferoxamine (DFO) in People With Leptomeningeal Metastasis	Deferoxamine (DFO)	Intrathecal	Adrienne Boire, MD	Memorial Sloan Kettering, New York, United States
NCT05289908	Phase I/II	Intrathecal Pemetrexed for Leptomeningeal Metastasis	Pemetrexed Folic Acid Vitamin B12 Dexamethasone	Intrathecal Pemetrexed Intrathecal Dexamethasone Oral Folic Acid Intramuscular Vitamin B12	Zhenyu Pan, PhD	The First Hospital of Jilin University Changchun, Jilin, China
NCT05305885		Intra-pemetrexed Alone or Combined With Concurrent Radiotherapy for Leptomeningeal Metastasis	Pemetrexed Dexamethasone Radiotherapy	Intrathecal, intraventricular or via lumbar puncture	Zhenyu Pan, PhD	The First Hospital of Jilin University Changchun, Jilin, China
NCT04192981	Phase I	GDC-0084 With Radiation Therapy for People With PIK3CA-Mutated Solid Tumor Brain Metastases or Leptomeningeal Metastases	GDC-0084 whole brain radiation therapy	Oral	Brandon Imber, MD, PhD	Memorial Sloan Kettering, New York, United States
NCT04729348	Phase II	Pembrolizumab And Lenvatinib In Leptomeningeal Metastases	Pembrolizumab Lenvatinib	Oral	Nancy Wang, MD, MPH	Massachusetts General Hospital Boston, Massachusetts, United States Dana Farber Cancer Institute Boston, Massachusetts, United States
NCT05459441	Early Phase I	Efficacy and Safety Evaluation of Percutaneous Ommaya Capsule Injection of Autologous Bi-dimensional Specific T Cells in the Treatment of Glioma and in Combination With Pemetrexed in the Treatment of Brain/Meningeal Metastases	Autologous progenitor expansion -T Pemetrexed	Intrathecal	Jun Ren, MD, PhD	Fudan University Pudong Medical Center Shanghai, Shanghai, China
NCT05112549	Phase I	Intrathecal Application of PD1 Antibody in Metastatic Solid Tumors With Leptomeningeal Disease (IT-PD1/ NOA 26) (IT-PD1)	Nivolumab	Intrathecal	Ghazaleh Tabatabai, Prof. Dr.	University Hospital Freiburg Freiburg, Germany University Hospital Heidelberg Heidelberg, Germany And 5 other locations in Germany
NCT03719768	Phase I	Avelumab With Radiotherapy in Patients With Leptomeningeal Disease	Radiotherapy Avelumab	Intravenous	Peter A Forsyth, MD	Moffitt Cancer Center Tampa, Florida, United States
NCT04511013	Phase II	SWOG S2000: A randomized phase 2 trial of encorafenib + binimetinib + nivolumab vs ipilimumab + nivolumab in BRAFV600- mutant melanoma brain metastases	Binimetinib Encorafenib Ipilimumab Nivolumab	Oral Binimetinib Oral Encorafenib Intravenous Ipilimumab Intravenous Nivolumab	Zeynep Eroglu, MD	NCI Cooperative group trial: SWOG/ECOG/NRG

The cellular landscape of melanoma LMD

With the differences in clinical prognosis between melanoma patients with LMD and those with parenchymal brain metastases (Hasanov et al., 2022; Raizer et al., 2008), it was clear early on that there was something very unique about the LMD tumors and their microenvironment. Single-cell RNA sequencing (scRNAseq) approaches have recently been utilized to elucidate and compare the immune landscapes of leptomeningeal metastasis, parenchymal brain metastases, and extra-cranial metastases. It was demonstrated that the typical melanoma LMD microenvironment is enriched for immune-suppressive myeloid cell subpopulations, including alternatively-activated macrophages, as well as a dysfunctional T cell landscape (Smalley et al., 2021). The authors report the presence of MDSC-like monocytes expressing high levels or TIMP1, MMP19 and IL1B in the CSF of poor responders. The macrophages present in these LMD samples showed higher expression of TGFβ, CD163 and IL10, while expressing lower levels of markers associated with classical activation such as IL1A, IL1B and IL6. CSF from patients with LMD predominantly contained subpopulations of T cells which lacked the expression of activation markers, such as IL2RA, CD69, GZMB and CD38, and those that expressed high levels of exhaustion markers, including PDCD1, LAG3 and CTLA4. Of note, the study also showed that a rare LMD exceptional responder who survived over three years on treatment had a CSF immune landscape more similar to patients without LMD involvement. The dynamics of various T cells, dendritic cells and macrophages in the CSF also correlated with the disease response versus progression over the course of various targeted and immune therapy treatments (Smalley et al., 2021). Another study tracked the cellular microenvironment of melanoma LMD over the course of two Phase II clinical trials of anti-PD1 and anti-CTLA-4 immune checkpoint inhibitors (NCT02886585, NCT02939300) (Brastianos et al., 2020; Brastianos et al., 2021). More CD8 T cells exhibiting higher levels of genes associated with effector function, including interferon gamma, were found in post-treatment specimens relative to pre-treatment samples (Prakadan et al., 2021). A correlation between the magnitude of the initial inflammatory response and clinical outcomes was also noted (Prakadan et al., 2021).

Proteomic landscape of the CSF microenvironment in melanoma LMD

In addition to possible diagnostic uses, mass spectrometry analysis of the CSF proteome can provide insights into drug resistance of tumors in the this microenvironment. Proteomic characterization of CSF from patients with melanoma LMD and those without LMD has demonstrated enrichment for proteins involved in innate immunity, protease-mediated damage, and IGF-related signaling (Smalley et al., 2020). Furthermore, it was shown that CSF from poor-responding patients contained pro-tumorigenic factors which can protect melanoma cells from MAPK inhibition in vitro. This occurred through upregulation of PI3K/AKT, integrin, B-cell activation, S-phase entry, TNFR2, TGFβ, and oxidative stress responses in the melanoma cells. Melanoma cells exposed to CSF from poor responders demonstrated an upregulation of TGFβ production and AKT phosphorylation at serine 473. It was shown that exogenous TGFβ protected melanoma cells from BRAF inhibitor-mediated cell death (Smalley et al., 2020).

Preclinical Models of LMD

The development of preclinical models of melanoma LMD is still in a work in progress (Table 2). Several current models utilize direct melanoma cell injection into a CSF space (intrathecal). This modelling technique has been conducted with the B16F-10 melanoma (Reijneveld et al., 1999; Schackert & Fidler, 1988). This model has the potential benefits of expedient LMD development, replication of clinical progression of LMD, and the creation of isolated LMD not confounded by concurrent systemic or parenchymal metastasis. A major limitation of this model is that it omits key steps in the natural metastatic process. Several of the recent insights on the metabolism and development of LMD utilized intracardially injected mouse models. These models are developed by passaging parental cell lines through multiple rounds of intrathecal injection via cisterna magna with ex vivo selection of GFP-positive tumor cells, and finally through intracardiac injection of the selected cells, which then may colonize the leptomeninges (Boire et al., 2017). This model takes more time to develop but has the advantage of capturing a more representative colonization process of cells seeding to the CSF. However, even in this model, the process of intracardiac injection skips several steps in the metastatic process.

Table 2:

Preclinical Models of LMD

Model	Intact immune system	Rapid LMD development	Replicate human disease	More natural metastatic process	References
Intrathecal injection of mouse melanoma cell lines	✓	✓			38, 39
Intracardially injected mouse cell lines selected for LMD	✓			✓	30, 32
Intrathecal injection of PDX		✓	✓		41
Spontaneously metastasizing PDX			✓	✓	40

Another group has recently established patient-derived xenografts from resected parenchymal brain tumors that spontaneously metastasize to the leptomeningeal space when injected subcutaneously in 21% of the cases (Faria et al., 2020). Going a step further, Law et al have developed in vitro and in vivo models of LMD from patient-derived cerebrospinal fluid circulating tumor cells (PD-CSF-CTCs) (Law et al., 2022). In this case, the cells were isolated from CSF of melanoma patients diagnosed with LMD. In this proof-of-concept study, scRNAseq identified IGF-1 signaling as a potential vulnerability of melanoma LMD. Combined inhibition of IGF-1 with MAPK-targeting therapy was then shown to inhibit the growth of both the in vitro and in vivo PD-CSF-CTCs models (Law et al., 2022). One major limitation of patient-derived models is the current requirements to utilize immune-deficient animals, thereby altering the immune component of the LMD microenvironment. This is a particularly notable limitation as immunotherapy is a critical component of contemporary therapies for this and other cancers.

We need better murine models of melanoma LMD for preclinical testing of novel therapeutic approaches. Since most patients with LMD have an Ommaya Reservoir placed to facilitate the administration of intrathecal chemotherapy and relieve symptoms of increased intracranial pressure easily, an analogous route of drug delivery is useful in murine models. Recently, a murine Ommaya-like reservoir was developed (Law et al., 2021) which allows the controlled administration of 3-7 microliter volumes using a Hamilton syringe and a motorized syringe pump. This provides therapeutic administration directly into the intra-ventricular space, is well tolerated for chronic use, and allows repeated IT administration of therapeutic agent(s). This overcomes limitations of repeated intrathecal treatments (inconvenient and stressful for the animals) such as via the intraventicular, cisterna magna, or lumbar cistern routes. In addition, very small volumes of CSF can be withdrawn using this device. It is very useful for the repeated administration of cellular therapies, therapeutic antibodies, chemotherapies and other therapies.

The current state of clinical care and clinical research efforts

The National Comprehensive Cancer Network^® (NCCN^®) guidelines for unresectable melanoma distant metastatic disease recommend systemic therapy as the preferred first line of treatment, including anti-PD1 monotherapy, combination checkpoint blockade, and combination targeted therapy (in the context of a BRAF V600-activating mutation) (Network, 2023). In cases of brain metastases, the guidelines remain similar but involve a multidisciplinary consultation (Network, 2023). Notably, the Food and Drug Administration recently approved a second-generation checkpoint inhibitor combination using a monoclonal antibody against LAG3 (relatlimab) in combination with anti-PD1 antibody nivolumab for the treatment of metastatic melanoma, offering a novel combination checkpoint blockade strategy in addition to the combination of anti-PD1 with anti-CTLA4 therapies ("FDA approves anti-LAG3 checkpoint," 2022). Typically, combination checkpoint blockade or anti-PD-1 monotherapy are the preferred first line of therapy for most patients, while BRAF/MEK targeting therapy is recommended for those with high-volume symptomatic disease and patients who are not appropriate candidates for immunotherapy ("FDA approves anti-LAG3 checkpoint," 2022). Although there are no NCCN^® guidelines specific for melanoma LMD, the NCCN guidelines for all patients with LMD primarily recommend supportive care and/or palliative radiation, particularly for patients with poor prognostic factors (i.e., Karnofsky performance scores (KPS) < 60; neurological deficits; active systemic disease or bulky CNS disease) (Network, 2022). Patients with high KPS, no major neurological deficits, minimal systemic disease and reasonable systemic treatment options are recommended to be treated with systemic therapy, intra-CSF systemic therapy, and/or radiation of bulky disease and neurologically symptomatic sites. Below is a summary of the several treatment modalities for melanoma patients with LMD, as well as completed and ongoing clinical trials for each. Table 1 summarizes the currently recruiting (or soon to be recruiting) trials which include melanoma LMD patients at this time.

Systemic Therapies

Systemic therapies discussed below refer to those administered either orally or via intravenous infusion.

i. Targeted therapy

Nearly half of all late-stage cutaneous melanoma patients have tumors with an activating mutation in the serine/threonine kinase BRAF (BRAF^V600E being most common), rendering it constitutively active (Fedorenko et al., 2015). Of melanoma patients with LMD, activating BRAF mutations are observed in a higher proportion (~65%) of patients than typically observed (~45%) in melanoma (Chorti et al., 2021; Fedorenko et al., 2015; Ferguson et al., 2019). For metastatic melanoma patients with activating BRAF mutations, systemic targeted therapy with combined inhibition of BRAF and MEK is the standard of care, due to superior efficacy and safety versus either inhibitor alone. The level of CSF penetrance of the majority of these agents has not been definitively determined, and the few studies that investigated their CNS distribution did not specifically assess drug levels in CSF but rather the distribution in brain tissue. In a study on blood-brain barrier penetrance, dabrafenib and trametinib have been shown to be substrates for p-glycoprotein and BCRP efflux pumps, resulting in reduced brain delivery (Mittapalli et al., 2013; Vaidhyanathan et al., 2014). Notably, dabrafenib has shown to have better CNS penetrance than vemurafenib (Mittapalli et al., 2013). However, this study only examined the brain concentration of the drugs, and it is unclear if CSF penetrance profiles would be similar. One study has specifically examined the concentrations of the BRAF inhibitor vemurafenib in CSF and showed that levels in CSF were usually markedly lower than extra-cranial sites (53.4±26.2 and 0.47±0.37 mg/l as the mean concentrations in plasma and CSF, respectively) and were shown to have patient-to-patient variability (Sakji-Dupré et al., 2015). Poor CSF penetration of dabrafenib and selumetinib were observed in a preliminary study in nonhuman primates (Rodgers et al., 2017). While no data currently exist for CSF levels of encorafenib, two studies are evaluating the safety and efficacy of a higher-than-FDA-approved dose of encorafenib in melanoma patients with CNS disease. The initial results of the Phase II POLARIS trial (NCT03911869), which tested high dose regimen of encorafenib (300 mg BID) plus binimetinib (45mg BID) in patients with melanoma brain metastases (excluding LMD), were presented at the Melanoma Bridge 2022 meeting. These results showed that the high-dose regimen was not well tolerated in the safety lead-in cohort, in which three of the nine evaluable participants experienced dose-limiting toxicity. It was also reported that the high-dose encorafenib regimen used, with BID dosing, had a lower C_max at steady state, suggesting a higher clearance than standard daily dosing. A second trial (NCT05026983) will evaluate encorafenib 600 mg QD and binimetinib 45 mg BID in metastatic melanoma patients with brain metastases and/or LMD. No results from this trial have been reported yet. Notably, radiotherapy may improve the blood-CSF penetration of targeted therapy and may be an effective alternative strategy to high-dose regimens (Sakji-Dupré et al., 2015; Sprowls et al., 2019).

As many targeted therapy trials exclude patients with leptomeningeal involvement due to fears of rapid clinical deterioration, prospective trial data for the efficacy of targeted therapy in BRAF-mutant melanoma patients with LMD does not exist. However, case reports have highlighted possible prolonged survival following treatment with targeted therapy for LMD (Abu-Gheida et al., 2019; Arasaratnam et al., 2018; Floudas et al., 2016; Glitza et al., 2017; Kim et al., 2015; Lee et al., 2013; Schafer et al., 2013; Wilgenhof & Neyns, 2015). Many of these reports describe a rapid response to therapy and well-managed control of neurological symptoms. Importantly, all patients in these case reports were BRAF/MEK inhibitor treatment-naïve. Once patients develop LMD while on targeted therapy, the LMD is often rapidly progressive and morbid. A potential future treatment option might be new BRAF inhibitors with enhanced CNS penetration (i.e., PF-07284890 and ABM-1310), which are both being evaluated in patients with brain metastases (NCT04543188 and NCT04190628). The clinical trial of PF-07284890 (NCT04543188) also allows for inclusion of patients with LMD.

ii. Immune checkpoint inhibitor therapy:

Immunotherapy has transformed the outcomes for late-stage melanoma patients with brain metastases. As with targeted therapy, almost all clinical trials investigating immune checkpoint inhibitor therapies, even in patients with parenchymal brain metastases, have traditionally excluded patients with LMD. The ABC trial, which evaluated both nivolumab and nivolumab + ipilimumab in melanoma patients with asymptomatic brain metastases, included a cohort (Cohort C) of patients with symptomatic brain metastases, previously treatmented brain metastases, or LMD, who were treated with single-agent nivolumab (NCT02374242). The cohort of 16 patients included 4 patients with LMD. It was reported that only one patient in the cohort responded, and that patient did not have LMD (Long et al., 2018). Despite lack of clinical response, two of the four patients with LMD exhibited a prolonged survival time of 49 and 86 weeks compared to 6 and 13 weeks for the remaining two LMD patients (personal communication with Georgina Long February 5^th 2023). The majority of published experiences with immune checkpoint inhibitor therapy in LMD are summarized in an increasing number of individual case reports. At least three such case reports describe patients with extended responses, some lasting over 2 years (Glitza & Bucheit, 2017; Smalley et al., 2016; Wu et al., 2020). A case report of primary melanoma leptomeningeal disease also described a response to anti-programmed cell death protein 1 (PD1) therapy with stable disease lasting over two years with no significant toxicity (Misir Krpan et al., 2020).

Several clinical trials have recently been reported or activated that are designed specifically for patients with LMD. A phase II trial of pembrolizumab for patients with LMD from solid tumors accrued thirteen of the planned sixteen patients (NCT03091478) (Naidoo et al., 2021). The study reported a 38% CNS response rate and a tolerable safety profile. Two treatment-sensitive patients demonstrated durable complete responses with overall survival of 9 months and 3+ years (Naidoo et al., 2021). However, no patients with melanoma were included in the study, and all patients were naïve to anti-PD1 checkpoint inhibitors. Another Phase II study evaluating the anti-cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) antibody ipilimumab and the anti-PD1 antibody nivolumab enrolled 18 patients with LMD, two of whom had melanoma (Brastianos et al., 2021). Patients with prior systemic treatment with an anti-CTLA4 antibody were excluded (Brastianos et al., 2021). The study met its primary endpoint of 3-month survival as 8 out of 18 patients were still alive (OS3 = 0.44; 90% CI: 0.24–0.66). A third of the patients reported grade 3 or higher adverse events, which is typical for the ipilimumab plus nivolumab combination treatments (Brastianos et al., 2021). Thus, there were no new or unexpected toxicities, and this profile is acceptable for the aggressiveness of LMD.

iii. Other combination therapies

Existing data demonstrates that combining targeted therapy and immune checkpoint inhibitors could take advantage of the rapid disease response observed with targeted therapy and the prolonged duration of responses seen with immune therapy, but the safety of such combinations may require careful management. The toxicity profiles of such combinations have already been reported in the treatment of non-LMD metastatic melanoma (Dummer et al., 2022; Gutzmer et al., 2020). Grade 3 or 4 toxicities occurred in 53-79% of patients treated with a combination of targeted and checkpoint inhibitor therapy. The most common adverse events associated with the combination are blood creatine phosphokinase increased, lipase increased, alanine aminotransferase increased, anemia, dermatitis acneiform, and amylase increased. One death due to limbic encephalitis occurred on the TRICOTEL study and was considered to be related to atezolizumab treatment (Dummer et al., 2022). Further information on the toxicities is provided within the study reports (Dummer et al., 2022; Gutzmer et al., 2020). A case of exceptional response and survival spanning over three years has been described in a patient treated with a combination of targeted (BRAF + MEK inhibition using dabrafenib/trametinib and then switched to encorafenib/binimetinib) and checkpoint inhibitor therapy (both nivolumab monotherapy and ipilimumab/nivolumab combination) (Smalley et al., 2021; Smalley et al., 2020). A new Phase II trial is testing the safety and efficacy of the triplet combinations of nivolumab with dabrafenib and trametinib, or nivolumab with encorafenib and binimetinib, in melanoma patients with unresectable metastatic disease, including melanoma patients with untreated LMD (NCT02910700). Initial data presented at the 2021 ASCO Annual Meeting on 27 patients enrolled in the study demonstrates the triple combination is well-tolerated, with promising clinical activity in patients with immunotherapy-refractory disease (Burton et al., 2021). A Phase II trial is investigating another combination of immune and targeted therapies for treatment of LMD using pembrolizumab and lenvatinib, a multi-receptor tyrosine kinase inhibitor (NCT04729348). The combination of systemic therapy with radiation may take advantage of the transient blood-CSF permeability following radiation therapy to improve CSF responses to systemic therapy administration (Minniti et al., 2021; Sakji-Dupré et al., 2015; Sprowls et al., 2019). A Phase Ib trial is testing the combination of intravenous anti-PD1 agent avelumab with 3000 centriGray whole brain radiotherapy (NCT03719768). Similarly, a Phase I trial is testing the combination of paxalisib (GDC-0084, inhibitor of PI3K/mTOR) with radiation therapy in patients with PIK3CA-mutated solid tumor brain metastases or leptomeningeal metastases (NCT04192981).

Intrathecal Therapies

Intrathecal (IT) drug infusion therapy for LMD usually involves the placement of an Ommaya reservoir, which enables direct and repeated drug administration into the CSF space through the ventricle, bypassing the blood-CSF barrier and maximizing the dose of drug available in the CSF space (Glitza et al., 2020; Ommaya, 1963).

i. Chemotherapy

Currently, two chemotherapy agents continue to be used via IT administration in melanoma LMD patients: thiotepa and topotecan. A multicenter phase II trial of IT topotecan in patients with LMD, which included three patients with melanoma (out of 62 total patients), reported a relatively safe toxicity profile but no therapeutic benefit over other intrathecal therapies (Groves et al., 2008). A randomized cooperative study of IT thiotepa vs IT methotrexate for patients with nonleukemic LMD showed neither therapy to provide significant neurologic improvement (Grossman et al., 1993). This study did not explicitly include melanoma patients, although there were 5 patients listed as “other solid” or “unknown” primary malignancy. Pape et al reported a prospective trial including nine patients with melanoma LMD who were treated with a combination of intrathecal chemotherapy and systemic chemotherapy (Pape et al., 2012). Of the 9 patients, two had overall survival of 168 and 60 weeks, demonstrating the potential benefit of combining intrathecal and systemic chemotherapy in some patients (Pape et al., 2012). One patient received intrathecal liposomal cytarabine and systemic fotemustine; the other received intrathecal liposomal cytarabine, then thiotepa and systemic fotemustine. However, to date, there has not been a single clinical trial showing a statistically significant benefit of IT chemotherapy, either as a single agent or in combination, in patients with LMD from solid tumors (Groves et al., 2008; Hitchins et al., 1987; Pape et al., 2012; Stewart et al., 1987). There are two active trials testing intrathecal pemetrexed plus dexamethasone in combination with radiotherapy or folic acid plus vitamin B12 (NCT05305885 and NCT05289908).

ii. Cytokine therapies

The IT mode of drug administration was first utilized with cytokine therapies the 1980s, mostly with recombinant interleukin-2 and interferon-alpha. The overview of the history of cytokine therapy for melanoma patients with LMD is well described in another review (Glitza et al., 2020). In summary, some patients experienced long-term survival with IT IL-2, but this treatment was associated with significant toxicity and led to performance status decline in all patients. Despite these limitations, the experience with IT IL-2 provided proof of concept that IT immunotherapy might durably benefit a subset of melanoma patients with LMD (Glitza et al., 2018). An important question is whether newer immotherapies can achieve this in more patients and/or with a better safety profile, similar to what has been seen in the general metastatic melanoma population.

iii. Immune checkpoint inhibitors

In addition to systemic administration of immune checkpoint inhibitors, there is now great interest in investigating IT administration of these therapies. A phase I/Ib trial tested the combination of IT and intravenous nivolumab (NCT03025256). Initial efficacy analysis on 23 patients showed the median OS was 63 % at 3 months, 42 % at 6 months and 30% at 12 months (John et al., 2021). No unexpected systemic or neurological toxicity was observed with the initially planned 5, 10 and 20 milligram IT doses, so a 50 mg dose was added. The trial is now closed to accrual and an updated analysis, including the results with the 50 mg IT dose cohort, are expected to be reported soon. Another Phase I study examining intrathecal administration of nivolumab is currently recruiting various tumors types with LMD (NCT05112549). Treatment with intrathecal nivolumab and ipilimumab combination immunotherapy is also being tested in a Phase I study (NCT05598853).

iv. Cellular Therapies

While intrathecal cellular anti-cancer therapy has been used for decades, and dates back to an initial report from 1987, recent advances in the use of cellular therapies, particularly chimeric antigen receptors (CAR) T cell therapy, has seen a significant rise in clinical trials using either intraventricular and/or direct intra-tumoral (within the CNS located) administration, particularly in gliomas (Elmadany et al., 2022; Shimizu et al., 1987). No prospective clinical trial has been reported for melanoma patients with LMD, but case reports indicate that this approach may be feasible and worth further investigation. The first report dates back to 1991, when a patient developed LMD years after initial melanoma diagnosis. The patient was treated with 5 × 10⁹ intraventricular lymphokine-activated killer cells (LAK), supported by IT IL-2 (Heimans et al., 1991). Unfortunately, the patient deteriorated due to LMD progression, and further treatment was held. A second patient was initially treated with radiation and IT IL-2, but progressed, and subsequently received three doses of IT cytotoxic T cells, again supported by systemic low dose IL-2 infusion (Clemons-Miller et al., 2001). Overall, the patient did well during treatment, with some improvement of neurological symptoms. The third patient reported in the literature was treated with 3 increasing doses of IT tumor infiltrating lymphocytes (TIL; 1^st IT TIL: 0.3x10⁹, 2^nd IT TIL: 1x 10⁹, 3^rd IT TIL: 3x 10⁹ ), each supported with IT IL-2 (Glitza et al., 2015). The patient had symptoms similar to the toxicities seen with IT IL-2, and unfortunately passed away 5 months after the IT TIL from systemic disease progression (Glitza et al., 2018). While these reports highlight the limited experience of such an approach in melanoma patients with LMD, they do show that this approach might be feasible for selected patients. The time from T cell harvest (either by pheresis or from tumor resection) to the time of IT administration, which could be as long as 6 weeks, might be prohibitive in the majority of patients due to rapid deterioration caused by their LMD. Nonetheless, an early Phase I trial is currently recruiting to test autologous bi-dimensional specific T cells in combination with pemetrexed (NCT05459441).

Radiation Therapies

i. Whole Brain Radiation Therapy (WBRT)

WBRT is typically used with palliative intent in melanoma patients with LMD. WBRT has been shown to be effective for neurologic symptom stabilization in LMD but does not improve overall survival (El Shafie et al., 2019). New strategies, including hippocampal-sparing radiation and concurrent treatment with memantine, can decrease the risk of neurocognitive decline following WBRT in patients with brain metastases (Brown et al., 2020; Brown et al., 2013). However, patients with LMD typically have diffuse involvement, either macroscopic or microscopic, and therefore hippocampal-sparing radiation is often considered inappropriate in these patients. As we continue to make advances in systemic and IT therapies for melanoma LMD, the optimal use of WBRT may also shift with synergistic strategies, such as combination approaches with immune checkpoint inhibitors (NCT03719768) (Arasaratnam et al., 2018).

ii. Stereotactic Radiosurgery (SRS)

SRS provides the advantage over WBRT in that only focal areas of the neuroaxis undergo irradiation, limiting some of the undesirable side effects and morbidity of diffuse radiation. In LMD this is generally of limited benefit, as the disease process can affect the complete craniospinal system. However, SRS may have a role in the management of focal LMD in select patients. A study of SRS for LMD in 16 patients with solid tumors, including one patient with melanoma, found evidence for successful local control and possible delay in the need for WBRT in some patients with focal disease (Wolf et al., 2017). Out of 14 patients with follow-up magnetic resonance imaging, focal LMD was stable in 5 and partially regressed in 8 patients at follow-up, while one patient had progression with hemorrhage 5 months after SRS. With improved craniospinal control, focal SRS may have an increasing role in the management of melanoma LMD.

iii. Proton cranial-spinal irradiation (CSI):

While traditional photon CSI has failed to gain widespread adoption for LMD management, recent data reveals proton craniospinal irradiation may increase CNS PFS and OS compared to involved field photon irradiation without an increase in grade 3 or 4 toxicities (Yang et al., 2022). Proton radiation has dosimetric superiority over photon radiation to the craniospinal axis by sparing the organs anterior to the spinal column from radiation dose (Yang et al., 2021). A phase II study from Yang et al enrolling 63 breast and lung patients revealed a median improvement in CNS PFS of 7.5 vs. 2.3 months with the use of proton CSI compared to photon IFRT and an approximately 4-month OS benefit with proton CSI. Six melanoma patients were enrolled in the exploratory pCSI group without randomization. It is difficult to assess the benefits of CSI in the heterogenous group of patients in the exploratory pCSI group, which also included ovarian cancer and esophageal cancer, among others. Therefore it is still unclear if proton CSI prolongs survival or improves quality of life in patients with melanoma LMD, (Yang et al., 2022) as this disease has frequently exhibited relative resistance to radiation. However, these promising results, including tolerable side effects, support the rationale for further study of CSI for the treatment of melanoma LMD.

iv. Radionuclide

The development of beta-emitting radionuclides is paving the way for innovative drugs for the visualization and treatment of disease (Chhabra & Thakur, 2022; Harris & Zhernosekov, 2022; Mehata et al., 2022; Wang et al., 2022). The use of rhenium-186-nanoliposomes promises delivery of radiation with high precision and exceptional tumor retention while sparing the surrounding normal tissue compared to external beam radiation therapy (Brenner et al., 2021; Phillips et al., 2022; Woodall et al., 2021). Based on promising preliminary results in the treatment of recurrent glioma (Brenner et al., 2021), a Phase I study Intraventricular Administration of Rhenium-186 NanoLiposome for Leptomeningeal Metastases (ReSPECT-LM) was initiated (NCT05034497). The first patient treated with 6.6 mCi186RNL via the Ommaya reservoir exhibited no adverse events. Enumeration of tumor cells in the CSF showed a gradual decline of tumor cells from 70.77 cells/ml (pre-treatment) to ~6 cells/ml over 48-168 hours after treatment administration (Brenner et al., 2022). Although this trial seems to be primarily focused on LMD from lung and breast cancer, the investigation of rhenium-186-nanoliposomes in melanoma LMD patients may warrant consideration.

Future directions

While the prognosis for melanoma patients with LMD remains poor, there has been a notable increase in research on LMD over the last few years. This research has already produced multiple novel and interesting insights, some of which are already leading to new clincal trials for patients. However, numerous areas of immediate unmet need remain in this challenging field. Here we summarize some of the key unmet needs and opportunities in both clinical and laboratory research to help accelerate progress and further improve outcomes for patients with LMD.

One of the key clinical challenges for patients with LMD has been their continued, nearly universal exclusion from prospective clinical trials. While this has often been done, at least in part, due to skepticism about feasibility of patient accrual due to the poor prognosis associated with LMD, the recently completed trials of systemic immunotherapies, intrathecal immunotherapy, and CSI therapy all repudiate this concern (John et al., 2021; Long et al., 2018; Yang et al., 2022). Thus, consistent with recent guidance from the US Food and Drug Administration (U.S. Department of Health and Human Services, 2020; U.S. Department of Health and Human Services, 2021), we advocate for the inclusion of patients with MBM and LMD in prospective clinical trials of novel agents, unless there is a specific conern for patient safety based on the features of the agent(s) being evaluated. However, we do acknowledge and recommend that patients with LMD should likely be enrolled and evaluated in distinct cohorts, based on the unique features of this disease, including the challenges of response assessment discussed above. The continued poor outcomes of these patients with currently approved agents underscores the rationale for allowing access to promising agents in early-phase clincial trials. Importantly, demonstration of signals of activity in these patients could suggest a novel indication for new agents. Further, longitudinal CSF pharmacokinetic and pharmacodynamic analyses may be critical for the design of future trials. Ultimately multi-center trials utilizing academic centers with experience in the multidisciplinary management of LMD will likely facilitate safe and accelerated testing of new strategies and can also provide the opportunity for standardized CSF collection and processing for robust translational analyses.

There is a also a need and opportunity for the development of novel systemic and intrathecal therapy approaches based on emerging insights into the unique biology and immunology of LMD. Additional checkpoint proteins appear to be critical for the immune-evasive biology of melanoma CNS metastases, including LAG3 and TIM3 (Wang et al., 2019). Along with PD1 and PD-L1, LAG3 and TIM3 have been detected on both intracranial and extracranial lymphocytes. Importantly, LAG3 staining is even more prominent on intracranial lymphocytes than extracranial, potentially highlighting a CNS-specific vulnerability of immune regulation (Wang et al., 2019). While Relativity-047 reported a favorable progression-free survival benefit of combined inhibition of LAG3 and PD1 versus PD1 inhibitor monotherapy (10.1 months vs 4.6 months), it remains to be seen if this combination has activity for CNS metastases (Tawbi et al., 2022). Second-generation chimeric antigen receptor (CAR) T cells have shown promising initial results for treating LMD from B cell malignancies (Bennani et al., 2019; Ghafouri et al., 2021; Karschnia et al., 2022; Li et al., 2020; Wu et al., 2021) and are currently being tested for LMD from breast cancer and glioblastoma (NCT03696030, NCT05063682). This cellular therapy approach might be adapted for melanoma LMD in the future, with a number of melanoma-specific CAR already under clinical investigation (Bahmanyar et al., 2022) (NCT04119024, NCT03893019, NCT02107963, and NCT02830724, among others). A possible limitation of these therapies could be neurotoxicity and/or cytokine release syndrome, which have been observed in patients treated systemically with CAR T-cells (Siegler & Kenderian, 2020). Additionally, despite the promising initial responses observed in non-Hodgkin lymphoma patients with LMD, CAR T-cell therapies often do not yield durable responses in these patients (Ghafouri et al., 2021; Li et al., 2020; Wu et al., 2021). The potential for more effective disease control may reside in new combinations of the most promising concepts. For example, combining WBRT with systemic immune checkpoint inhibition, as in NCT03719768, (Arasaratnam et al., 2018), may improve outcomes by ehancing the blood-CSF barrier permeability (Sprowls et al., 2019), but with the known risk of neurocognitive decline in patients that achieve prolonged survival. The recent safety and efficacy data reported or proton CSI therapy for patients with LMD suggests a promising alternative. Such innovative combinations will benefit from improved understanding of the on- and off-target effects of these treatments, the LMD microenvironment, and the biology of the tumor cells at this unique site.

Accelating progress also critically depends upon investment in translational research opportunities afforded by new clinical trials, continued development of models of melanoma LMD for functional testing, and expansion of basic science research focused on LMD biology and immunology. It is imperative that a systematic approach to sample collection and processing be standardized across the field. We recommend the establishment of a work group focused on this challenge and opportunity, similar to the unified effort to standardize response assessments in clinical trials for patients with CNS metastases (Chukwueke & Wen, 2019; Lin et al., 2015; Soffietti et al., 2022), and in melanoma for neoadjuvant clinical trials (Amaria et al., 2019; Menzies et al., 2021; van Akkooi et al., 2022). This can also build upon the lessons learned from the establishment of the International Neoadjuvant Melanoma Consortium (INMC), which published guidelines for both clincal trial design and pathology processing/review, ultimately facilitating pooled analyses of pathological response versus clincal outcomes across multiple studies. Key considerations for translational studies include types of samples to be collected, key timepoints to standardize, and standard operating procedures (SOPs) for samples processing- and associated clincal data and outcomes. Such samples will be critical to improving diagnostics, creating new models and understanding the biology of this devastating disease. With the low sensitivity of the “gold standard” cytopathology for the diagnosis of melanoma LMD, there is a dire need to improve the diagnostic approaches and the ability to monitor treatment response. A concerted sample collection effort will enable and expedite the evaluation and validation of promising new approaches, such as ctDNA profiling, mass spectrometry analysis, and scRNAseq of patient CSF. Ultimately, these efforts will expedite the implementation of of new diagnostic strategies that are desparately needed.

Another obvious need is the development of additional models of LMD. The includes the need for in vitro cell line models which incorporate key features of the LMD-specific microenvironment, as well as in vivo models, particularly with intact immune function. Together, such models will facilitate both the study of LMD biology and the testing, optimization, and prioritization of new therapeutic strategies. These tools will be very important becase tumor cells growing in the leptomeningeal space are uniquely difficult to study. For example, LMD is rarely resected due to the diffuse nature of the disease, and the quantity of analyzable cellular material available in the CSF is frequently small. With scRNAseq capabilities, a more thorough cellular and molecular characterization of the tumor at this site is finally possible through the interrogation of CSF. As we develop more sophisticated analysis approaches for these rich datasets, this technology will allow us to thoroughly map out the cellular ecosystem of this unique tumor microenvironment and hopefully result in a better understanding of the drivers of resistance to current therapies and identify rational strategies to overcome them.

Key Concusions

• Due to the poor prognosis of melanoma LMD patients and unique biology of this disease, LMD-specific therapeutic approaches are urgently needed

• Patients with melanoma LMD must be included in prospective clinical trials

• A standardized approach to sample collection and processing must be established

• Better diagnostics for melanoma LMD are critically needed

• New models of melanoma LMD and an investment into basic science research of LMD will fast-track the development of new therapies

• The development of a network of researchers focused on melanoma LMD would accelerate communication and progress

In short, although we have finally begun to unravel the mysteries of this devastating disease, significant momentum in the translational science and clinical research focused on melanoma LMD is urgently needed. It is now more important than ever to act on the last thought in Eberth’s case report (Eberth, 1869) “wünsche ich nur, es möchte anderen Forschern bald gelungen, die Entwickelung dieser Neubildung zu ermitteln,” translated as “I only wish that other investigators would soon be able to determine the development of this new formation.”