Efficacy and safety of the combined use of ipilimumab and nivolumab for melanoma patients with brain metastases: a systematic review and meta-analysis

Abstract Context Immune checkpoint inhibitors have advanced immunotherapy for melanoma patients. Objective: This study evaluates efficacy and safety of ipilimumab and nivolumab combination (IN) for melanoma brain metastases (MBM) patients. Materials and methods Literature search was conducted in electronic databases and studies were included if they reported efficacy and safety of IN in MBM patients or prognostic information related to brain metastases. Outcomes evaluated were objective response rate (ORR), complete remission/stable disease/progressive disease rates, progression-free survival (PFS), overall survival (OS), incidence rates of adverse events, and hazard ratios of disease progression or mortality between IN-treated patients with and without brain metastasis. Results Intracranial ORR was higher in IN-treated MBM patients than with control therapies (nivolumab or ipilimumab plus fotemustine). IN treatment led to longer PFS and OS in than control treatments. Five-year OS of IN-treated MBM patients was up to 51% compared to 34% for nivolumab. Outcomes were better for treatment naïve and asymptomatic patients. Whereas many studies reported significantly higher mortality or progression risk with IN treatment in MBM patients compared to non-MBM melanoma patients, many others did not find this risk significant. Incidence of grade 3/4 adverse events in IN-treated MBM patients was: diarrhea or colitis (16%), hepatitis (15%), rash (8%), increased alanine transaminase (8%), increased aspartate aminotransferase (7%), increased lipase (6%), increased amylase (4%), fatigue (3%), hypophysitis (2%), pneumonitis (2%), headache (2%), nausea or vomiting (1%), and neutropenia (1%). Conclusion IN is an efficacious and safer treatment option for MBM patients, especially for asymptomatic and treatment naïve patients.


Introduction
Melanoma is a cancer of melanocytes.It is relatively uncommon (accounts for 1% of skin cancers) but a more dangerous cancer because of its strong spreading potential.It can develop anywhere in the body, but common sites are the neck, face, trunk (in males), and legs (in females).Having dark skin lowers the risk of melanoma incidence.In the USA alone, the annual incidence of melanoma is estimated at over 97000 individuals, and about 8000 individuals die each year with this cancer [1].Approximately 1.3 million individuals are surviving melanoma in the USA.The incidence of melanoma is increasing at younger ages.In individuals between 30-40 years of age, melanoma is the fifth most common form of cancer.In individuals under 30 years of age, the incidence of melanoma is increasing faster [2].
After lung and breast cancers, melanoma is the most common source of brain metastases.Melanoma brain metastasis (MBM) is a common complication that can develop in as high as 70% of patients and is associated with a worse prognosis and earlier mortality [3].Less than 1% of patients have brain metastasis at melanoma diagnosis, and the incidence is about 28% at the diagnosis of metastatic melanoma [4].Melanoma at the head or neck, ulceration of the tumor, and Breslow thickness are identified as the risk factors for the development of brain metastases [5,6].The poor prognosis of MBM patients is partly due to the resistance of the tumor to external radiation therapy and poor penetration of systemic agents into the brain [7].Female sex, a lower number of brain metastases, normal lactic dehydrogenase levels, and treatment are identified as positive prognostic factors for MBM patients [8,9].
Immunotherapy with checkpoint inhibitors has profoundly improved the outcomes of patients with advanced melanoma.A meta-analysis of 85 clinical trials revealed that immunotherapy led to a higher complete remission (CR) rate than other treatments and the highest CR rate was observed for immune checkpoint inhibitors [10].Among the important checkpoint inhibitors, ipilimumab, which is a recombinant human immunoglobulin (IgG4) monoclonal antibody, binds to cytotoxic T lymphocyte-associated protein-4 (CTLA-4) and increases the activation of T-cells.Nivolumab, an IgG4 monoclonal antibody, is another checkpoint inhibitor that binds to the programmed death 1 (PD-1) receptor and activates T-cell response to tumors [11].Whereas their monotherapies have improved the outcomes of advanced melanoma patients [12], the combined therapy with ipilimumab and nivolumab (IN) provides a synergistic effect.A randomized controlled trial (RCT) that compared the outcomes of IN versus ipilimumab or nivolumab in untreated advanced melanoma patients found median overall survival of 60 months in IN group, 37 months in the nivolumab group, and 20 months in the ipilimumab group [13].
Historically, MBM patients are treated with surgery, radiotherapy, and stereotactic radiosurgery [14][15][16].Among the systemic treatments, targeted therapies with BRAF inhibitors and MEK inhibitors provide short-term benefits [17,18].However, combinational therapies with checkpoint inhibitors are found to provide better survival outcomes to MBM patients.Whereas fewer studies have evaluated the efficacy of IN therapy in MBM patients, many others have reported the prognostic role of IN in the management of MBM.However, there is no synthesis of these outcomes.This systematic review aimed to refine the evidence of the efficacy and safety of IN for the treatment of MBM and to examine the prognostic role of brain metastases in IN-treated melanoma patients.

Inclusion and exclusion criteria
For inclusion, a study had a) to be retrospective or prospective in design; b) recruited melanoma patients with and without brain metastases to treat them with IN therapy; c) reported response and survival rates of IN therapy in MBM patients; and d) reported a prognostic endpoint related to the outcomes of IN therapy in MBM patients.Exclusion criteria were: a study a) compared either ipilimumab or nivolumab, or used both these drugs in a sequential design; b) used additional agents with IN therapy; c) performed radiotherapy with IN therapy; d) compared ipilimumab with either nivolumab or pembrolizumab without differentiation of outcomes; and e) case reports.

Literature search
Electronic databases (Embase, Google Scholar, Ovid, Science Direct, Springer, and Wiley) were searched for peer-reviewed research articles using the most relevant key terms.The original research strategy was 'Melanoma AND brain OR central nervous system OR intracranial AND metastases AND ipilimumab AND nivolumab AND combined therapy' .After identifying relevant articles, reference lists of retrieved articles were screened to strengthen the literature search.The search encompassed original research articles published in English from the inception date of each database till January 2023.

Data analysis
Demographic information, melanoma type, tumor dimensions, tumor stage, performance status of patients, therapeutic history, clinical markers, and brain metastases data were extracted from the research articles of included studies.Information about the study sample size, design, follow-up, and dosage schedules was also extracted from the respective research articles of selected studies and was organized on datasheets.The quality of included studies was assessed with Cochrane's Tool for the Assessment of the Quality of Randomized Controlled Trials or the Newcastle-Ottawa Tool for Assessing the Quality of Non-Randomized Studies in Meta-analysis.
Outcomes of interest were the objective response rate (ORR; CR rate plus partial remission rate), CR rate, stable disease (SD) rate, progressive disease (PD) rate, progression-free survival (PFS), overall survival (OS), and incidence rates of adverse events.An additional outcome of interest was the hazard ratio of disease progression or mortality between IN-treated patients with versus without brain metastasis.
Meta-analyses of proportions were performed to estimate overall ORR, rates of CR, SD, and PD, and the incidence rates of adverse events by using numerical data.These meta-analyses were incorporated with Freeman Tukey's double arcsine transformation of proportions for variance stabilization.Meta-analyses of odds ratios were performed to examine the risk of new brain metastasis development during treatment and the incidence of adverse events between IN and control therapies.Statistical analyses were performed with Stata software (StataCorp, College Station, Texas).
The quality assessment of the included studies is presented in Tables S2a&b.The quality of included RCTs was moderate in general.Unclear bias in allocation concealment and blinding of participants/personnel and outcomes were the main constraints in randomized studies.In non-randomized studies, all studies fulfilled the criteria of representativeness, ascertainment, comparability, and follow-up.
Intracranial SD rate was 0% with both first-line and second-line IN treatments [22] and was 4% and 0% in patients with asymptomatic and symptomatic MBM respectively [32].The intracranial PD rate was 25% with first-line IN treatment but 96% with second-line IN treatment [22], and was 30% in patients with asymptomatic MBM but 61% in patients with symptomatic MBM [32].Details of response rates observed in the included studies are presented in Tables S3a-e.
Many studies have reported the survival outcomes of MBM patients who received IN therapy.A synthesis of these data is presented in Table 1.In general, IN treatment was associated with longer PFS and OS in comparison with control treatments [21,24].Five-year OS of IN-treated patients was up to 51% compared to 34% for nivolumab treatment [25].Better survival outcomes were observed in treatment naïve than in previously treated patients [22], and in asymptomatic than in symptomatic MBM patients [32].Survival was also longer with BRAF-mutant compared to wild-type genotypes in IN-treated MBM patients [21].
Several studies have reported the hazard ratios depicting the prognostic role of IN therapy in MBM patients (Table 2).From these prognostic data, it was evident that IN treatment led to better survival outcomes in comparison with control treatments.However, whereas many studies reported statistically significantly higher mortality or progression risk with the IN treatment in MBM patients than in melanoma patients without brain metastases, many others did not find this risk statistically significant.
Although the odds of new brain metastases were not statistically significantly different between IN-treated and control patients (OR 3.34 [95% CI: 0.21, 53.68];I 2 = 92%; p = 0.394), the number of new brain metastases developed during treatment was slightly less in patients treated with IN in comparison with controls (11% vs 17%).
Treatment with IN was associated with considerably high adverse events in MBM patients (Figure S1a&b).A meta-analysis showed that 83% [95% CI: 64, 96] of patients experienced at least one adverse event.Adverse events observed in 2 or more studies were: rash (55% [95% CI:

Discussion
In this review, we included studies that recruited MBM patients to treat them with IN as well as studies that recruited melanoma patients and sought a prognostic role of IN in patients with versus without brain metastasis.We found that IN treatment profoundly improved the outcomes of patients with MBM.Although fewer studies reported the comparative efficacy and safety of IN, the response and survival rates were better with IN than with comparators.These outcomes were also endorsed by the hazard ratios between IN-treated and control MBM patients.However, whereas many studies reported a statistically significant higher hazard of mortality or progression with the IN treatment in patients with versus without brain metastases, many others did not find this relationship statistically significant.Patients with symptomatic brain metastases have a poor prognosis.In a study comparing the outcomes of IN therapy between asymptomatic and symptomatic MBM patients, although, in general, outcomes were poor for symptomatic MBM patients, at least in some patients with symptomatic MBM durable responses were observed.Authors postulated that steroid dependence at IN therapy start may interfere with the activity of immunotherapy [32,33].Patients with brain metastases often use steroids for symptom relief or prophylaxis.The immunosuppressive effects of corticosteroids are well known.Steroids upregulate PD1 and CTLA4, thus blocking T-cell proliferation.However, the administration of CTLA4 inhibitors may partially prevent the immunosuppressive effects of steroids.In the presence of steroids, CTLA-4 inhibition is accompanied by a partial reduction in T-cell proliferation [36].
Corticosteroid administration impairs anti-PD1 responses by causing persistent systemic lymphodepletion, especially it leads to significant reductions in CD4 + and CD8 + T cells.However, steroids manifest differential effects in intracranial and extracranial environments [37].It has been reported that the presence of extracranial tumors is necessary for the intracranial activity of anti-PD1/CTLA4 drugs.This is because antigens inside the brain are less accessible for their transportation to lymph nodes and the immune response created by the checkpoint inhibition relies on antigens outside the brain [38].
Based on the observation that the best survival outcomes were noted for treatment-naive BRAF mutation-positive patients, whereas outcomes were dismal for BRAF mutation-positive patients who progressed from BRAF targeted therapy (ORR 30%; median PFS 3 months), Serra-Bellver et al. [31] suggested that IN therapy should be a first-line treatment option for BRAF mutation-positive patients.Di Giacomo et al. [21], who in a post-hoc analysis found better efficacy of IN in patients with BRAF mutation in comparison with wild-type genotypes, suggested that these outcomes should be investigated further in future prospective trials.They also suggested that first-line IN therapy in MBM patients is a feasible and manageable strategy.Long et al. [24] who also found a good prognosis of untreated stage IV MBM patients with IN treatment suggested that IN therapy should be considered as first-line therapy for MBM patients.
In the trials designed to evaluate the efficacy and safety of IN for MBM patients, the vast majority of IN-treated patients experienced at least one adverse event during treatment including rash, fatigue, diarrhea or colitis, hepatitis, nausea or vomiting, pruritus, increased ALT, AST, ALP, lipase, or amylase, hyperthyroidism, hypothyroidism, hypophysitis, headache, cough, pneumonitis, arthralgia, myalgia, vitiligo, and adrenal insufficiency.However, grade 3 or 4 events were relatively fewer.These included: diarrhea or colitis, hepatitis, rash, increased ALT, increased AST, increased lipase, increased amylase, fatigue, hypophysitis, pneumonitis, headache, nausea or vomiting, and neutropenia.It has been observed that patients receiving fewer cycles of IN had a lower incidence of grade 3/4 adverse events [31].Some studies have shown that adverse events may be beneficial in prolonging survival in patients treated with immune checkpoint inhibitors [28,[39][40][41].Thus, the incidence of adverse events may be a predictor of a better prognosis [28].However, a meta-analysis of 56 trials involving 14 types of cancer patients found that the risk of early death (within 3 months of treatment start) was higher with immune checkpoint inhibitors compared to combinational therapies.Since the overall incidence of treatment-related death in trials is <2%, therefore, such an impact can be of lesser significance [42].
Among the ongoing trials, the German noninterventional study, NICO (NCT02990611), is intended to enroll 750 MBM patients to investigate the survival outcomes of IN or nivolumab-treated patients in a follow-up period of 5 years.Interim results involving 623 patients indicate that both regimens are effective [43].Among others, a phase I trial (NCT05704933) intends to evaluate the effectiveness of a single dose of nivolumab with ipilimumab or relatlimab in MBM patients who can undergo surgery for removal of brain metastases 7 to 10 days post-treatment.Another phase II trial (NCT04511013) intends to compare encorafenib, binimetinib, and nivolumab with IN in BRAF-V600 mutant MBM patients.Such studies will further refine the present-day evidence.
Immunotherapy is a rapidly evolving branch of pharmacology, and it can be hoped that the prognosis of MBM patients will further improve.Exploration of more effective combinations of checkpoint inhibitors and their use with surgery or radiotherapy remains a challenging area of pharmaceutical sciences.A focus on patients with symptomatic brain metastases and leptomeningeal tumors is also important [44].Another important area will be the determination of the preference threshold for systemic treatment with IN over stereotactic radiosurgery.Moreover, targeted therapies are also needed to be developed by targeting pathways that lead tumors to metastasize in the brain [45].Selection of a suitable therapy, optimized scheduling, appropriate sequencing, and tailoring for prognostic factors remain a challenge for oncologists.Besides IN, several other combinational immunotherapies are being developed that may help individualize treatment based on tumor characteristics and microenvironment [8].For maximizing the outputs of immunotherapy, lifestyle factors may also play a role.Recently, immune checkpoint inhibition in association with the Mediterranean diet has been found to yield a better response to the therapy in advanced melanoma patients [46].
The present study collates outcomes of IN therapy in MBM patients from 14 studies to summarize present-day evidence of IN efficacy.However, it is constrained by some important factors including heterogeneity in the outcome measures and data size.Among the foremost limitations of the present study, less availability of comparator-controlled data is more important.Also, the survival rates could not be pooled to achieve overall estimates because of the differing comparative designs of the included studies.Moreover, the survival endpoints of many studies were not reached during the study duration.Less availability of data also precluded us to seek prognostic information about the roles of age, sex, brain metastases characteristics, lactic dehydrogenase levels, etc.
In conclusion, immunotherapy with immune checkpoint inhibitors has been found effective in MBM patients, a subgroup of melanoma patients who were previously excluded from trials because of their poor prognosis.Objective response and 5-year survival rates have now exceeded the 50% mark.Whereas many studies found a statistically significant higher hazard of mortality or progression with the IN treatment in patients with brain versus without brain metastases, many others did not find this risk significantly.Although a high incidence of adverse events is observed with IN treatment, grade 3 and 4 events were relatively less.These results encourage the use of IN as a first-line treatment for MBM patients.

Figure 1 .
Figure 1. a flowchart of study screening and inclusion process.

Figure 2 .
Figure 2. Forest graph showing the response rates of In treated and control patients.

Figure 3 .
Figure 3. Forest graph showing the odds ratios of adverse events between In treated and control patients.

Table 1 .
Survival rates of mBm patients observed in the included studies.

Table 2 .
hazard ratios depicting the efficacy of In in patients with melanoma brain metastases.