The purpose of this research project is to test the effectiveness of nivolumab versus nivolumab together with ipilimumab for the treatment of melanoma brain metastases.

Patients are eligible to join this study if they are aged 18 years or above and have been diagnosed with melanoma with brain metastases.

BACKGROUND Brain metastases are a common and serious complication of metastatic melanoma. Up to 50% of patients develop brain metastases during the course of their illness, and approximately 20% of patients have them at first presentation with metastatic disease. The prognosis for melanoma patients with brain metastasis is generally very poor with a median overall survival ranging from 2.8 to 4 months and a large proportion of up to 95% of these patients will ultimately die as a direct result of their brain metastases. The prognosis of patients with melanoma brain metastases has not changed between 1986 and 2007. The five year survival rate is 15% for late stage disease.

LOCAL TREATMENT OPTIONS There are 3 treatment modalities used for the local management of brain metastases: surgery, stereotactic radiosurgery (SRS), and whole brain radiotherapy (WBRT). Surgery and SRS are directed to the tumour volume only, whereas WBRT delivers lower-dose radiotherapy to the whole brain, including areas of no overt tumour. Surgery is not limited by the size of the tumour; however, it is dependent on the lesion being surgically accessible. In appropriately selected patients (lesions >3 cm and minimal symptoms), SRS is considered equivalent to surgical resection; however, this has been addressed in only 1 randomised controlled trial that was stopped early because of poor accrual. No statistically significant difference was found in survival or local tumour control. In solid tumours, WBRT remains the treatment modality of choice in those where surgery or SRS are contraindicated because of tumour size, number, or location. The role of WBRT is to palliate symptoms because, despite response rates of 60%, median survival is less than 5 months.

SYSTEMIC THERAPY OPTIONS Systemic chemotherapy has shown little benefit in the treatment of metastatic melanoma, including those with brain metastases. It is usually reserved for those who have central nervous system (CNS) progression despite surgery and/or radiotherapy, or in those patients with rapidly progressing or symptomatic extracranial disease. The response rate in the brain for the most active chemotherapy agents temozolomide and fotemustine is <10% in large clinical trials, thus patients with brain metastases have been excluded from most systemic therapy clinical trials.

Ipilimumab was the first systemic treatment proven to extend survival in patients with metastatic melanoma, and has activity in progressing brain metastases in patients who are not taking corticosteroids for neurological symptoms.

Dabrafenib and vemurafenib are potent selective BRAF inhibitors proven to increase survival in patients with V600 BRAF-mutant metastatic melanoma and have activity in brain metastases. Although there is unprecedented proven activity of the BRAF inhibitor dabrafenib in a large clinical trial of patients with untreated V600 BRAF-mutant brain metastases, this represents only 40% of the metastatic melanoma population, and responses are rarely durable. Brain metastases remain a major clinical problem, and an unmet medical need for patients with both BRAF-mutant and wild-type metastatic melanoma. And yet, all major clinical trials continue to exclude such patients.

IMMUNOTHERAPY Cancer immunotherapy rests on the premise that tumours can be recognized as foreign rather than as 'self' and can be effectively attacked by an activated immune system. An effective immune response in this setting is thought to rely on immune surveillance of tumour antigens expressed on cancer cells that ultimately results in an adaptive immune response and cancer cell death. This functions by aborting the emergence of tumours as they arise and/or causing tumour shrinkage where it is present. Meanwhile, tumour progression may depend upon the acquisition of traits that allow cancer cells to evade immune surveillance and an effective immune response. This evasion may occur by exploiting any of the checkpoints that control the regulatory immune response, including the display of antigens and control of co-stimulatory pathways that affect the proliferation of cells involved in immunity. Current immunotherapy efforts attempt to break the apparent tolerance of the immune system to tumour cells and antigens by either introducing cancer antigens by therapeutic vaccination or by modulating regulatory checkpoints of the immune system - either directly by stimulation of immune cells by antibodies directed to receptors on T and B cells or indirectly by cytokine manipulation. T-cell stimulation is a complex process involving the integration of numerous positive, as well as negative, co-stimulatory signals in addition to antigen recognition by the T-cell receptor (TCR). Collectively, these signals govern the balance between T-cell activation and tolerance to antigens.

NIVOLUMAB Nivolumab is a fully human monoclonal antibody directed against the negative immunoregulatory human cell surface receptor PD-1 (programmed death-1 or programmed cell death-1/PCD-1) with immunopotentiation activity. The efficacy and safety of nivolumab is being explored in 3 ongoing melanoma phase 3 trials in previously treated or treatment naïve patients.

IPILIMUMAB In a study investigating the safety and activity of ipilimumab specifically in patients with brain metastases, ipilimumab was shown to have activity in some patients with advanced melanoma and brain metastases, particularly when metastases are small and asymptomatic and in patients who do not need corticosteroid treatment. Furthermore, the drug had no unexpected toxic effects in this population.

NIVOLUMAB COMBINED WITH IPILIMUMAB Preclinical and preliminary clinical evidence suggests a synergy between nivolumab and ipilimumab. While PD-1 and CTLA-4 are both co-inhibitory molecules, evidence suggests that they use distinct mechanisms to limit T cell activation. The preliminary clinical evidence has demonstrated a higher frequency of patients with substantial tumour burden reduction for the combination of nivolumab and ipilimumab. Improved overall survival associated with substantial tumour burden reduction has been noted with immunotherapies. For instance, improved overall survival has been noted in metastatic melanoma patients obtaining a complete response to IL-2. If this observation is also applicable to treatment with nivolumab combined with ipilimumab then there could also be the potential for large improvements in overall survival compared to ipilimumab.

STUDY DESIGN Initially, this study will recruit patients with melanoma brain metastases who have received no prior local treatment for their intracranial metastatic disease and who are asymptomatic (Cohort 1) and in parallel, patients who have been previously treated for their brain metastases, have symptoms, or have concurrent leptomeningeal disease (Cohort 2).

Combination therapy with nivolumab and ipilimumab is currently under investigation in a phase III trial in previously untreated metastatic melanoma (NCT01844505). That study however excludes those patients with active brain metastases. This current study will therefore enrol a third group of patients with brain metastases who have received no prior treatment for their metastatic disease and who are asymptomatic to receive combined nivolumab and ipilimumab (Cohort 3).

Recruitment to cohort 3 will commence once the first 6 patients from cohort 1 have received at least 3 doses of study treatment (equivalent to 6 weeks) and have a satisfactory adverse event record. A satisfactory record is deemed where ≤ 2 patients experience neurological CTCAE grade 3 or above nivolumab-related adverse events. Following this safety assessment, if no safety signals are detected, patients will continue to be recruited to this cohort to enable a complete patient group of 30. To minimise bias, treatment allocation to cohort 3 and the remainder of cohort 1 will be assigned by unequal randomisation ratio to achieve an overall balance of 30 patients receiving nivolumab 3 mg/kg (including the 6 patients from cohort 1) and 30 patients receiving the combination treatment. Randomisation to the remaining cohort 1 and all of cohort 3 will be stratified by participating site to minimise potential differences between patients who present to the different sites, or because of differences between the sites themselves.

CONTINUED TREATMENT IN SELECT CASES OF PROGRESSIVE DISEASE Accumulating clinical evidence indicates some patients treated with immune system stimulating agents may develop progression of disease (by conventional response criteria) before demonstrating clinical objective responses and/or stable disease. enhanced inflammation within tumours could lead to an increase in tumour size which would appear as enlarged index lesions and as newly visible small non-index lesions. Over time, both the malignant and inflammatory portions of the mass may then decrease leading to overt signs of clinical improvement. Alternatively, in some individuals, the kinetics of tumour growth may initially outpace anti-tumour immune activity. With sufficient time, the anti-tumour activity will dominate and become clinically apparent. Therefore, patients will be allowed to continue study treatment after initial investigator-assessed RECIST 1.1 defined progression if they are assessed to be deriving clinical benefit and tolerating study drug.

The application of traditional RECIST criteria in patients treated with immunotherapy may lead to premature discontinuation of treatment in a patient who will eventually respond to treatment or have prolonged disease stabilization. Disease progression may occur in extracranial lesions whilst patients may continue to have disease stabilisation or response of their intracranial melanoma disease, and vice versa.

INTRACRANIAL RESPONSE USING FET-PET FET PET (18F-fluoro-ethyl-tyrosine [FET]) will be used as a separate and independent modality to MRI brain scans to assess metabolic response of the brain metastases to immune therapy. FET PET has been extensively evaluated in humans. FET-PET has shown to be valuable in the management of brain tumours. It is not known what the effect of lymphocytic infiltration which occurs with immune therapy is on MRI brain, so it is important to have a separate modality to predict benefit and assess response. FET-PET demonstrates the change in cell proliferation (change in SUV [standardised uptake value] from baseline).

FET-PET will be used to assess to determine response to study treatment. The FET PET findings will be compared with conventional imaging. Only cohorts 1 and 3 will be assessed with this modality.

BLOOD AND TISSUE BIOMARKERS Blood will be collected to examine serum chemokines, cytokines, inflammatory markers, lymphocyte and T cell subsets and myeloid derived suppressor cells (MDSC) to assess correlation with disease response or progression. In patients with sufficient archival melanoma tissue from metastatic sites, a baseline tumour PD-L1 level, immune markers and genetics of response and resistance will also be measured.

If available, tumour tissue following progression of disease will also be tested for immune and genetic markers. Early work has indicated these biomarkers may be predictive of responders to study treatment. In a phase 1 study of 90 patients receiving nivolumab at different dose levels, high pre-treatment NY-ESO-1 and MART-1-specific CD8+ T cells were associated with progression of disease. At week 12, increased peripheral-blood T regulatory cells and decreased antigen-specific T cells were associated with progression. PD-L1 tumour staining was associated with responses to nivolumab, but negative staining did not rule out a response.

Description of adverse events by type, frequency and severity using National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE) version 4.03.

Number of patients who withdraw from the study due to intolerable adverse reactions.

Cohort 1 and 3

Inclusion Criteria:

1. ≥18 years of age.
2. Written informed consent
3. AJCC Stage IV (any T, any N, M1c) histologically confirmed melanoma or unknown primary melanoma. Patients must have at least 1 radiological definitive brain metastasis that is ≥ 5mm and ≤40mm measurable per RECIST version 1.1 guidelines.
4. In patients with prior BRAF inhibitor treatment, intracranial disease progression must be demonstrated (RECIST >20% or new measurable brain metastases) compared with nadir of intracranial response during BRAF inhibitor treatment, and confirmed with a second MRI brain scan after drug washout period (+/- 3 days).
5. No prior localised treatment for brain metastases (eg. surgery or radiotherapy).
6. Neurologically asymptomatic from brain metastases.
7. Eastern Cooperative Oncology Group (ECOG) Performance Status of 0-2, and life expectancy > 30 days.
8. Able to undergo MRI with Gadolinium contrast agent.
9. Adequate haematological, hepatic and renal organ function.
10. Women of childbearing potential: negative serum pregnancy test and effective contraception from 14 days prior to study treatment until 23 weeks after the last dose.
11. Men with female partner of childbearing potential to use effective contraception from 14 days prior to study treatment until 31 weeks after the last dose.

    Exclusion Criteria:

12. Any melanoma brain metastasis >40mm.
13. Ocular melanoma.
14. Prior treatment with an anti-PD-1 or anti-PD-L1 , anti-PD-L2, anti-CD137, or anti-CTLA-4 antibody, or any other antibody or drug specifically targeting T-cell co-stimulation or checkpoint pathways.
15. Patients with active, known or suspected autoimmune disease. Patients with vitiligo, type I diabetes mellitus, residual hypothyroidism due to autoimmune condition only requiring hormone replacement, psoriasis not requiring systemic treatment, or conditions not expected to recur in the absence of an external trigger are permitted to enroll.
16. Current systemic treatment with corticosteroids, except prednisone at nonimmunosuppressive doses of ≤ 10 mg/day (or equivalent). Past treatment for non-neurological symptoms allowed, if ceased 2 weeks prior to starting study treatment. Inhaled or intranasal corticosteroids (with minimal systemic absorption) may be continued if patient on a stable dose. Non-absorbed intraarticular steroid injections will be permitted.
17. Any investigational drug or other systemic drug therapy for melanoma within 28 days or 5 half-lives from baseline.
18. Known to be HIV positive, or a positive test for hepatitis B and C .
19. Another malignancy or concurrent malignancy unless disease-free for 3 years.
20. Serious or unstable pre-existing medical conditions or other conditions that could interfere with the patient's safety, consent, or compliance.
21. Pregnant or breastfeeding females.

Cohort 2 - per Cohorts 1 & 3, except patients must have at least one of the following:

1. Failed prior local therapy for brain metastases (including surgery, stereotactic radiotherapy or whole brain radiotherapy) where disease has progressed per RECIST (>20% increase in SOD or new measurable brain metastases),

   and/or;

2. Have current neurological symptoms related to brain metastases. IF they have received prior local therapy for brain metastases, the disease must have progressed per RECIST (>20% increase in SOD or new measurable brain metastases),

   and/or;

3. Have leptomeningeal disease concurrently with measurable brain metastases. IF they have had failed prior local therapy for brain metastases, this must have progressed per RECIST (>20% increase in SOD or new measurable brain metastases).