Symptomatic brain metastases in melanoma

Although clinical outcomes in metastatic melanoma have improved in recent years, the morbidity and mortality of symptomatic brain metastases remain challenging. Response rates and survival outcomes of patients with symptomatic melanoma brain metastases (MBM) are significantly inferior to patients with asymptomatic disease. This review focusses upon the specific challenges associated with the management of symptomatic MBM, discussing current treatment paradigms, obstacles to improving clinical outcomes and directions for future research.

metastases, followed by a rapid wean of corticosteroid, prior to initiation of ICI, while others support a less invasive approach with stereotactic radiation to multiple lesions.Whether any surgery or radiotherapy is required in asymptomatic brain metastases remains controversial, given the excellent intracranial response rate to doublet ICI, with more than 50% of patients achieving a disease response or stability at 6 months. 16This question is currently being tested in the ABC-X trial (NCT03340129) comparing doublet ICI with or without stereotactic radiosurgery (SRS).Below we discuss the merits and disadvantages of available therapies for the management of symptomatic MBM patients.

| SU RG ERY
Surgical management of symptomatic brain metastases is recommended by guidelines 13,14 due to the rapid symptom improvement and low morbidity it offers. 17Unfortunately, most data on surgery pre-date the era of ICI use, such that current recommendations on the role of surgery are largely empiric.Notably, surgery may be required to confirm diagnosis and to obtain tissue for molecular testing.
Data from the 1990s established that resection of a single brain metastasis could improve outcomes compared to whole brain radiotherapy (WBRT); comparing WBRT with or without resection of a single brain metastasis, surgical intervention resulted in improved survival and function. 18,19In one study, adding surgery to radiation gave faster improvement in functional status. 19Of note, a similar study showed no benefit of adding surgery, 20 with proposed confounding factors this this trial being poorer performance status and more extensive extracranial disease.A meta-analysis of these studies, including 195 patients, did not show a significant difference in overall survival, noting inter-trial heterogeneity. 21A large retrospective analysis from the pre-immunotherapy era showed that resection of MBM was associated with better prognosis, and that the presence of extracranial metastases was associated with worse outcomes. 22A study from the immunotherapy era reviewing the timing of surgery and ICI found that patients who had surgery prior to ICI had better survival than those who had ICI. 23e question of surgery in the setting of multiple metastases is more complex.There is limited retrospective data that removal of all lesions (in all cancer types) gives superior survival outcomes without an increase in surgical complication rates (8%-9%). 24Metastectomy of up to four lesions, post-operative WBRT, and improved recursive partitioning analysis (a score based on Karnovsky performance status, age, control of primary tumour and extracranial disease) were associated with better outcomes in patients with multiple brain metastases. 25However, with the advent of sophisticated radiation techniques, discussed next, this approach is seldom pursued in the modern era.In this setting, removal of the most symptomatic or critical metastases, followed by radiation would be a standard approach.
Rapidly progressing, symptomatic or larger lesions, or lesions in critical areas such as the posterior fossa may sway a multidisciplinary team to recommend up-front surgical management. 13

| Whole brain radiotherapy
WBRT was used historically for symptomatic MBM, but this is increasingly avoided due to poor efficacy, associated neurocognitive toxicity and the development of alternative treatment approaches.
In patients with symptomatic MBM, local therapy, such as surgery or SRS is recommended. 13In patients for whom this is not appropriate, such as those with leptomeningeal disease or lesions in functionally sensitive locations within the brain, WBRT, especially in conjunction with other therapies, may improve symptom control and survival. 26wever, adjuvant WBRT is now generally avoided in the management of symptomatic MBM.A randomised trial comparing WBRT with observation alone after local treatment (surgery and SRS) in ≤3 MBM showed no statistical difference in intracranial disease control, performance status or survival. 27Further, the addition of WBRT to SRS is associated with neurocognitive decline. 28Adding WBRT to surgical resection improved intracranial disease control in solid tumours but not survival, 29 and worsened cognitive outcomes, 30 although use of hippocampal sparing radiation techniques and memantine may mitigate neurocognitive decline. 31,32Overall, given the potential delays to systemic therapy incurred with longer treatment courses, and the potential toxicities, the role for WBRT is very limited in symptomatic patients.

| Stereotactic radiosurgery
With improving imaging resolution and sophistication of radiation techniques, SRS had become the preferred radiation modality for MBM.Highly focused SRS treatment reduces toxicity to surrounding brain tissue and can be performed in ambulatory settings.As a primary treatment, SRS is favoured for fewer lesions (up to 4), 33 to a maximal size of 3-4 cm. 13,33There is evidence to support a recommendation for SRS in 5-10 brain metastases in patients with good performance status. 33,34For lesions 4-6 cm in diameter, surgery is generally preferred 33 as SRS to larger tumour volume is associated with decreased rates of local control and survival, and higher risks of radiation necrosis. 35Individual metastasis control rates in MBM at 12 months are reported as 76%-91%, [36][37][38] and Christ et al. 39 reported similar retrospective outcomes between patients treated with surgery plus SRS and SRS alone.
Surgery can be combined with SRS, with adjuvant treatment to the surgical cavity.Mahajan et al. demonstrated in a randomised controlled trial that SRS following surgical resection of one to three brain metastases reduced the rate of local recurrence, and proposed this as an effective alternative to adjuvant WBRT. 40e risk of radiation necrosis is of special consideration in patients who are to receive ICI, as it can result in symptoms that may be confused with disease progression or that may require corticosteroids.In a retrospective analysis, large brain metastases (greater than 2 cm) treated with single fraction SRS caused higher rates of radiation necrosis than multi-fraction SRS at 1 year (20% vs. 8%). 41lti-fraction SRS also conferred higher rates of local metastasis control in MBM at 1 year (67% vs. 55%) but no difference in survival.
This accords with a meta-analysis of largely retrospective data comparing single and multi-fraction SRS, 42 that showed improvements in 1-year local control from 77% to 93%, and reductions in the rates of radionecrosis from 23% to 7% in brain metastases of 2-3 cm.In MBM greater than 3 cm, there was a reduction of radionecrosis with multi-fraction SRS from 12% to 6.5%, but no improvement in local control (77% with single fraction and 79% with multi-fraction).
Intralesional haemorrhage of MBM is associated with poorer local control rates and survival. 43A study found that in larger MBM, the use of single fraction SRS (rather than multi-fraction) was associated with higher rates of post-treatment haemorrhage. 44In the immunotherapy era, avoiding haemorrhage and the associated symptoms and need for corticosteroid use is of critical importance, and supports the use of multi-fraction SRS in larger MBM, in line with the current American Society for Therapeutic Radiology and Oncology (ASTRO) clinical practice guidelines. 33 note, the location of MBM may affect the risk of radiation necrosis.One study showed that deep white matter metastases had higher rates of radionecrosis after SRS, 45 possibly as these locations are watershed areas of perfusion and therefore at increased risk of impaired tissue viability from radiation damage.

| Combining SRS with systemic therapy
For symptomatic brain metastases that are not amenable to resection, SRS followed by ICI may be a well-tolerated option. 46Treating metastases with SRS concurrently with ICI (the latter commencing within 4 weeks of SRS) gave local control rates of 87% in a study which included 21% symptomatic MBM patients. 47In this study, 5% of patients developed symptomatic radionecrosis requiring intervention, with a median time to onset of 10.5 months after SRS.The improved efficacy of local control when combining ICI with SRS is reflected in several studies, and in all cancers, 48 and in melanoma. 37,49,50ross tumour-types, 51 and particularly in MBM treated with CTLA-4 inhibitors, 52 radiation necrosis is more likely when combining SRS with ICI combination therapy versus SRS alone.A meta-analysis suggested rates in ICI plus SRS of 16% versus SRS alone of 6.5% (OR 2.35), 53 with most cases diagnosed within the first 18 months after SRS.The consequences of radiation necrosis can overlap with those of progressive disease, including the development of symptoms, repeated imaging, biopsies, corticosteroid use, or even surgery.
Concurrent SRS with BRAF-inhibitors has not been prospectively studied, but potentially does not add substantially to local control rates and is associated with increased risk of toxicities.Such combinations have been shown to prolong local control (SRS with BRAF-inhibitors), however much of the outcome data is confounded by the fact that comparator cohorts were typically BRAF wild-type melanoma patients and received different systemic therapies, including some that were observed without receiving systemic therapy. 54There is an increased risk of both intracranial haemorrhage, 54 and radiation necrosis 55 when combining targeted therapy with SRS.In the unusual instances in which combined SRS and BRAF-MEK inhibition is preferred for MBM, such as in cases of resistance to or contraindication of ICI, withholding the inhibitors is recommended for 2-3 days before and after SRS.

| SYS TEMI C THER APY
3][64][65] There are few prospective ICI trials that include patients with symptomatic MBM, Table 1 includes a summary.CheckMate 204 9,16 studied the use of ipilimumab plus nivolumab in two cohorts: symptomatic (n = 18) and asymptomatic patients (N = 101).Of the symptomatic patients, 12 were taking corticosteroids, and 17% responded to treatment.The symptomatic cohort had a median survival of 8.7 months (compared to median OS not reached, 12-month OS 82% in the asymptomatic cohort), and a clinical benefit rate (complete response, partial response or stable disease ≥6 months) of 22% versus 58% in the asymptomatic cohort.
The ABC trial 8 was a randomised phase II study which included a cohort of patients with symptomatic, leptomeningeal MBM, or those that had failed systemic therapy (n = 16).One patient (6%) showed a response to nivolumab alone.Of note, 75% of this cohort had previously been exposed to BRAF-MEK inhibitors. 66al-world results in retrospective analyses of responses to ICI in symptomatic MBM are similarly poor.A study of 29 patients requiring corticosteroids showed an overall response rate of 28% and median OS of 5.5 months to ipilimumab plus nivolumab. 57Parakh et al. found response rates to nivolumab alone were 21% in symptomatic MBM, with a median OS of 5.7 versus 13.0 months favouring asymptomatic MBM. 58 intracranial response rates to treatment with targeted therapies are short-lived, even in patients with BRAF mutations, the best chance of long-term control of MBM remains with ICI.Table 2 summarises outcomes in symptomatic and asymptomatic patients.
A prospective phase II trial (COMBI-MB) assessing dabrafenib and trametinib in BRAF V600-mutant MBM showed response rates of 59% in 17 symptomatic patients.However, the median duration of response was 4.5 months and at 6 months only 13% of patients had not progressed 59 Interestingly, re-analysing the cohorts in this study together, the trial also demonstrated that corticosteroid use was associated with a worse intracranial response rate (39% vs. 63% in those not taking corticosteroid) and shorter survival (median 9.3 vs. 13.5 months), 67 although this was at least partly because nonresponding patients were more likely to require corticosteroids for symptoms.It is critical to note that intracranial response rates to ICI after progression on BRAF-MEK inhibitor therapy are as low as 4.8%, demonstrated in a retrospective study which included 44% of TA B L E 1 Outcomes of patients with symptomatic melanoma brain metastases treated with immunotherapy.patients who had symptomatic MBM at presentation, 68 with similarly reduced response rate in extracranial disease. 69The DREAMseq study, which only permitted treated MBM, has demonstrated superior survival when using first line ipilimumab plus nivolumab over BRAF-MEK inhibitors. 70Therefore, extrapolating from the data for asymptomatic MBM, our preference is to give ipilimumab plus nivolumab first line in patients with BRAF-mutated MBM, where possible.For patients on high dose corticosteroids, other options are discussed below.

Intervention
The idea of a bridging period of treatment to reduce corticosteroids prior to commencing ICI is intriguing and may have an increasing role in future practice.One method to reduce corticosteroid use before commencing ICI is the use of the antiangiogenic vascular endothelial growth factor (VEGF) inhibitor, bevacizumab, also used to treat radiation necrosis. 71A case series of 12 patients supports this approach in MBM, demonstrating the ability to wean corticosteroids and reduced peri-tumoural oedema on imaging. 72There are also postulated synergistic effects of bevacizumab and ICI, related to VEGF's role in immune regulation. 73ere is limited data surrounding the use of a BRAF-MEK inhibitor 'induction' period as a bridge to ICI and to reduce corticosteroid use. 74,75The SECOMBIT study assessed this approach in extracranial disease, noting high response rates and clinical benefit with 'sandwiched' short course targeted therapy and a planned switch to ICI, followed by targeted therapy again at progression. 76The TRICOTEL phase II study looked at adding atezolizumab after 28 days of vemurafenib and cobimetinib in BRAF V600 mutated melanoma, and atezolizumab and cobimetinib in BRAF wild-type melanoma, and included over a third of patients with symptomatic MBM.In the BRAF mutation cohort, the response rate was 42%, versus 27% in the BRAF wild-type cohort.Of 11 patients with BRAF mutations on more than 2 mg of dexamethasone per day, four were able to rapidly taper their corticosteroid to zero over 28 days (prior to ICI initiation), five continued on greater than 2 mg per day during ICI initiation, and two discontinued treatment early. 77mbining BRAF-MEK inhibitors with PD-1/PD-L1 inhibitors has been trialled and considered as a bridging option in symptomatic MBM, but has limited data to support it as routine practice.A case series of five patients, all with symptomatic brain metastases, demonstrated that the combination of BRAF-MEK inhibitors and PD-1 inhibitor (and radiation) could produce durable responses, with two patients experiencing a complete response, and more than 3 years without progression. 78Initial results from the TRIDeNT phase II study, including ten MBM patients (some symptomatic), treated with nivolumab, dabrafenib and trametinib, reported an intracranial response of 57%. 79

| MA JOR OPEN QUE S TI ON S: OBS TACLE S TO IMPROVING OUTCOME S
The poor clinical outcomes of patients with symptomatic MBM are exacerbated by difficulties in conducting clinical trials in this patient TA B L E 2 Outcomes of patients with symptomatic metastatic brain metastases treated with targeted therapy.population.Trials are often limited by strict eligibility criteria, which may exclude active brain metastases or the use of corticosteroids.
In our opinion, the only way to meaningfully impact the clinical outcomes of patients with symptomatic MBM is to design clinical trials specifically tailored to this patient population.In contrast to many existing studies, these trials would necessitate thoughtful screening assessments and timelines, and must be maximally permissive in allowing multi-modal therapy as required and concurrent corticosteroid use.

| MA JOR OPEN QUE S TI ON S: FUTURE RE S E ARCH
There are promising novel strategies being trialled in melanoma that have implications for the management of symptomatic MBM, represented in Figure 1.Corticosteroids are used in MBM to reduce symptoms caused by vasogenic oedema.Unfortunately, corticosteroids are associated with multiple toxicities, as well as poorer melanoma survival outcomes 64 Corticosteroid toxicities occurring in >20% of patients include sleep disturbances, lipodystrophy, adrenal suppression, metabolic syndrome, hypertension, vertebral fractures and weight gain. 80Severe complications such as myocardial infarction, opportunistic infections and severe gastrointestinal adverse events occur in a further 1%-5% of patients. 80Corticosteroids affect all immune cells 81 and impair the proliferation and differentiation of naïve T cells thereby potentially abrogating the efficacy of ICI, particularly in immunotherapy-naïve patients, who have yet to develop a population of primed and expanded T cells against tumour antigens. 82 3][64][65] In a study comparing the outcomes of patients with MBM in the pre-immunotherapy era and the current immunotherapy era, the presence of symptoms in the pre-immunotherapy era was not associated with a worse The use of autologous tumour-infiltrating lymphocytes (TILs) in melanoma has received FDA approval for patients with immunotherapy-resistant melanoma, and may be an option for MBM in the future.Adoptive cell therapy involves harvesting TILs from fresh tumour and expanding them ex vivo before reinfusion.A cohort study 83 demonstrated that 18 patients with small, asymptomatic untreated MBM had an overall response in 28% to TIL therapy, but had worse progression-free survival and overall survival than those without MBM.The authors concluded that local therapies would be required in most MBM patients.A phase II study reported the efficacy of a single infusion of lifileucel, an adoptive cell therapy, after lymphodepletion in 78 patients, including 11% with asymptomatic brain lesions. 84The overall response rate of 36% and disease control rate of 80% were extremely promising in this otherwise treatment-refractory population.The subgroup of patients with brain metastases appeared to have a similar magnitude of benefit. 84Most toxicity related to the nonmyeloablative chemotherapy conditioning regimen, and largely resolved by 6 months after treatment. 83

CO N FLI C T O F I NTE R E S T S TATE M E NT
There are no conflicts of interest which would influence the integrity of this manuscript; however, the authors are involved in the BETTER clinical trial (NCT06163820), investigating the use of bevacizumab as a corticosteroid-sparing agent in patients with symptomatic melanoma brain metastases.The authors have no financial or commercial disclosures.

outcome 10 ;
probably because almost all MBM patients fared poorly prior to the widespread use of ICI.Taken together, these data suggest that the inferior clinical outcomes of patients with symptomatic MBM may at least be partly due to corticosteroid use.The development of strategies to minimise corticosteroid use may thus improve outcomes.Towards this, the forthcoming phase I/II BETTER trial (NCT06163820) is assessing the use of bevacizumab in symptomatic MBM patients.The study aims to assess the safety and tolerability of combining this treatment with ipilimumab plus nivolumab and SRS, with the hope of sparing corticosteroid use and improving response rates in symptomatic MBM.

F I G U R E 1 8 |
A phase III study comparing TIL therapy with ipilimumab in first or second line (88% had progressed on previous PD-L1 directed ICI, and allowed up to two asymptomatic MBM) showed a median PFS benefit of 4.1 months, and OS benefit of 6.9 months,85 although the comparator arm of ipilimumab alone is unlikely to be used in the Current and future management of symptomatic melanoma brain metastases.Top: corticosteroids used to reduce oedema also inhibit T-cell function; left: using monoclonal antibodies (such as VEGF inhibitors) to target areas of oedema and increase T cell infiltration; middle: isolating tumourinfiltrating lymphocytes, expanding this population ex vivo, and re-infusing them; right: using chimeric antigen receptor (CAR) T-and CAR NK-cells to target brain metastases.current era.A pilot study of lifileucel is underway in patients with asymptomatic MBM who have progressed on prior PD-L1 therapy (clini caltr ials.gov ID: NCT05640193).Another early phase study (clini caltr ials.gov ID: NCT05588453) is assessing the safety and dosing of natural killer (NK) cell therapy in combination with temozolomide.This study allows corticosteroids and aims to exploit the innate anti-tumour activity of donor NK cells to induce anti-tumour immune response.Exploring novel immune targets is a high priority in developmental therapeutic strategies for symptomatic MBM.In addition, further understanding of the biology of MBM could unlock new therapeutic strategies.For example, an analysis of the mutation signatures of different sites of melanoma metastases demonstrated enrichment of PTEN and CDKN2A alterations in MBM. 60Given that response rates to targeted therapies are less affected by corticosteroids, PI3K pathway inhibition could theoretically be used to target the downstream effects of aberrant PTEN signalling in MBM.Another study 86 comparing intracranial to extracranial metastases demonstrated increased BRAF and PI3K/AKT mutations in brain lesions, confirming the potential of this signalling pathway for targeted therapy.Further elucidation of the biological underpinnings of the inferior outcomes of patients with symptomatic MBM may guide the development of rationally designed clinical trials targeting MBM biology.CON CLUS I ON S AND PER S PEC TIVE S Although significant strides have been made in the management of melanoma and melanoma brain metastases, patients with symptomatic MBM still have poor clinical outcomes.The current standard of care for this patient population involves multidisciplinary management.In treatment of MBM, a combination is recommended of surgical management of larger, more symptomatic metastases, stereotactic SRS for optimising local control at each site, and/or ICI therapy.The poor clinical outcomes in symptomatic MBM may at least partly be contributed to by the need for corticosteroids to control vasogenic oedema, which directly impair cellular immunity.Consequently, the development of therapeutic strategies tailored to this group of patients is an urgent unmet need.Novel combinations of known melanoma drugs, as well as innovative immune and targeted therapies may provide the treatment transformation required to improve outcomes.However, we may never realise this goal if patients with symptomatic MBM continue to be systematically excluded from clinical trials.AUTH O R CO NTR I B UTI O N S A. Knox: writing-original draft, writing-review and editing.T. Wang: writing-original draft, writing-review and editing.M. Shackleton: conceptualisation, writing-original draft, writing-review and editing.M. Ameratunga: supervision, conceptualisation, writing-original draft, writing-review and editing.ACK N OWLED G EM ENTS None.Open access publishing facilitated by Monash University, as part of the Wiley -Monash University agreement via the Council of Australian University Librarians.