a) Patients were to have no standard effective therapeutic options for their melanoma. b) Patients in the dose-escalation phase were to have their Cmax modelled so that it could be ensured that the dose selected for the next cohort had a predicted Cmax that did not exceed 1 nM. Dose-escalation was to be stopped if Cmax reached 1 nM. c)Patients who had completed the dose-escalation phase could continue treatment at a higher dose than originally administered, provided the dose had been shown to be tolerable and effective in subsequent dose cohorts. d) Only 3 study centres were to participate in the dose-expansion phase (Arm 1). e) Patients could potentially experience severe systemic adverse reaction to IMCgp100 administration such as generalised cytokine storm; guidance on management of such events was included. f) Patients were to be closely monitored for the duration of infusion and for 2 hours after.

a) Patients could be enrolled if they had previous malignancy that, in the opinion of the investigator, was cured. b) Localised radiotherapy could be used to treat a tumour flare. c) The number of study centres that could participate in the dose-escalation phase was increased to 5, as recruitment was slower than anticipated. d) gp100 status was to be collected and reported, if available. e) Pregnancy was to be reported as an AE during the study and abortion, stillbirth or malformation/disease in the baby were to be reported as SAEs.

a) The phrase ‘clinically significant’ was added to the DLT definition, to allow the investigator discretion in cases of short-lived changes in clinical chemistry or haematology. In addition, transient lymphopenia (Grade 3 and 4) and transient non-life threatening cutaneous toxicity were excluded from the DLT definition; transient Grade 3/4 lymphopenia was no longer considered a reason for discontinuation of study treatment. b) Patients could be included in the study if their lymphocyte count was ≥0.5x109/L; this lowering of the permitted lymphocyte count reflected the results of data collected and assessed by the study team, which indicated that patients with low but stable lymphocyte count tolerated IMCgp100 without apparent consequences. c) Patients receiving anti-coagulant treatment could be included in the study, as there was no reason to suspect that IMCgp100 interferes with anticoagulant treatment or vice versa. d) The SMC could choose to expand a dose cohort to further define toxicity before deciding to dose escalate. e) Biopsies were not to be taken from patients whose medical history indicated a risk of uncontrollable bleeding.

The number of study centres was increased to 6 for the dose-escalation phase of Arm 1 and 8 for the dose-expansion phase; additional study centres could be set up in the USA and Australia.

Sites in the USA only: a) The phrase ‘clinically significant’ was removed from the DLT definition, as stipulated by the FDA. b) Follow-up visual, auditory and neurologic assessments were to be available for consideration by the team before escalation could proceed. c) Study treatment was to be discontinued for patients who experienced DLT during the first 30 days of the dose-escalation phase, and dose reduction was mandated for patients who experience Grade 3/4 toxicity that if observed during the DLT period would have met the definition of a DLT. d) Adverse events of Grade ≥3 nausea, diarrhoea, or vomiting that persisted beyond 72 hours despite optimal medical therapy were to be considered DLT events. e) A patient continuing treatment in the dose-escalation phase at a dose subsequently identified as being higher than the MTD was to have study treatment discontinued or could have treatment continued at a lower dose level. f) Patients with a change in LVEF of ≥20% from baseline were to be discontinued from the study.

a) The amount of HSA required as a blocking agent for IMCgp100 in saline infusion bags was increased from 125 mcg/kg to 125 to 250 mcg/kg, as a higher amount is needed for USP than for BP saline bags to ensure complete drug delivery. b) Patients were to have QTc, calculated using Bazett’s or a locally-preferred formula, of >500 ms.

a) Patients were to be included in the study if there was an appropriate window between alternative therapeutic options but excluded if early treatment with vemurafenib was an option. The intention was to include patients who had either slow-growing or stable disease that was asymptomatic, where a window existed between therapeutic options. b) Patients with brain metastases were only to be excluded if they were unstable, required steroid treatment, or had been irradiated within the previous 28 days. In addition, MRI scan of the head was to be conducted at screening (rather than CT scan) to provide a better baseline assessment of brain metastases. c) Removed the exclusion criterion specifying that patients with high disease volume were to be excluded from the study; it was deemed that sufficient patients had been treated in the study to date with no signs or symptoms of tumour lysis syndrome. d) Patients could be enrolled if their alkaline phosphatase level was >2.0 x ULN (the related inclusion criterion was removed). e) Biopsy of normal skin was introduced as an optional procedure, to allow investigation of a rash seen in treated patients; in addition, the timing of tumour biopsy was changed to facilitate investigation of tumour flare. f) Patients who required systemic steroid treatment for any reason other than treatment of IMCgp100 related AEs could be replaced. Steroid treatment abrogates IMCgp100 potency, and patients treated with steroids for prolonged periods were considered non-evaluable for efficacy.

a)The MTD for the IMCgp100 weekly-dosing regimen was 600 ng/kg; methodology relating to the dose-escalation phase of Arm 1 were removed from the protocol as the MTD had been identified. b)Infusion times should be shortened, as previous IMCgp100 infusions (>120 infusions) were tolerated well. c)The 30-day dosing break between treatment cycles was removed, and weekly treatment cycles were to comprise 8 doses administered over 56 days. Disease assessment was to be conducted every 8 weeks during treatment, to reflect the changed treatment cycle duration. d)Periodic brain scans were implemented for patients with brain metastases. e)The frequency of laboratory safety parameter sampling and ECG, echocardiogram, ophthalmological, auditory and neurological assessments was reduced after Cycle 1, as data from the dose-escalation phase indicated no significant concerns. f)At Cycle 1 a detailed pharmacokinetic profile was to be established for the first dose with peak and trough levels being assessed for subsequent doses. At subsequent cycles, pharmacokinetic parameters were only to be assessed at the Day 50 dosing. g)PBMC and serum samples were to be tested for evidence of “epitope-spreading”, to assess the ability of IMCgp100 treatment to initiate an adaptive immune response towards tumour antigens. h)Addition of 2 new study sites to recruit patients into the dose-expansion phase of the study. i)A minimum of 6 patients were to provide tumour and skin biopsy samples, and the timing of biopsies was revised to reflect the time of greatest inflammation and T cell infiltration. j)Overnight hospitalisation was required after the first IMCgp100 dose, with careful consideration being given to whether hospitalisation was required for the first dose administered after a treatment break. k)The remit of the SMC was extended, to ensure continued monitoring into the dose-expansion phase of Arm 1 and allow for modification of dosing regimens if appropriate. l)Allowed patient

a) A daily IMCgp100 dosing regimen was to be tested in a separate treatment arm. Patients were to receive 4 consecutive daily doses of IMCgp100, starting at 10 mcg and escalating to 20, 30, 40 and possibly 50 mcg per day based on observed tolerability, with a 2-week break between each 4 day dosing period. b) Dosing was to change from a per-kg-body-weight basis to a flat dose. For weekly-dosing patients the dose changed from 600 ng/kg to 50 mcg. c) The number of laboratory safety samples be reduced for Arm 1, Cycle 1 to limit the volume of blood being taken. d) The total number of patients to be enrolled was increased to 80, the number of study centres was increased to 10, and the duration of the study was extended to August 2015 to allow for treatment of patients in Arm 2. e) For patients in the dose-expansion phase of Arm 1, dose administrations were to be at least 3 days apart. f) Treatment allocation to Arm 1 or 2 would not be randomised but would depend on the resource availability at each site.

a) CT/MRI scans be taken at the end of each treatment cycle (every 6 weeks rather than every 12 weeks as planned previously) for the first 6 months of IMCgp100 dosing in Arm 2. This was to allow closer monitoring of disease status and make efficacy data more comparable to those in Arm 1. b) PFS, OS, and investigation of circulating tumour cells were specified as exploratory endpoints. c) For patients in Arm 2 infusion duration was to start at 30 minutes and could subsequently be reduced to 15 minutes if tolerable. d) The DLT assessment window for Arm 2 was 15 days, for logistical reasons. e) Use of low-dose steroids to compensate for a deficiency in natural levels could be considered on an individual-patient basis on discussion with the sponsor.

a) At least 5 patients in Arm 2 provide a CSF sample and contemporaneous blood sample at the end of dosing on Day 4 or Day 25, to compare IMCgp100 pharmacokinetic parameters in blood and CSF. b) irRECIST be used to assess tumour burden, in addition to RECIST. c) The stringency of the inclusion criteria for haemoglobin levels and creatinine clearance be reduced to reflect clinical experience. d) The time window between treating the first and additional patients in each cohort of the Arm 2 dose-escalation phase be reduced from 5 days to 4 days (after the last dose). e) Grade 3 laboratory values that were not clinically indicated would not trigger expedited communication and an SMC meeting. f) Prophylactic anti-coagulation therapy could be used by patients with or at risk of pulmonary embolism or deep vein thrombosis. g) The number of patients to be enrolled was increased to 100 and the number of study centres was increased to 15. h) Blood sampling for RNA analysis be implemented for patients in Arm 2. i) Recording of BRAF and NRAS status was implemented.

a) The number of patients enrolled into each expansion cohort be increased to 40, bringing the total number of patients to 140. b) Central independent review of CT scans would be implemented for patients with disease response.

a) Newly-treated patients would receive 40 mcg IMCgp100 for the first 2 doses administered in Arm 1 and the first 4 doses administered in Arm 2. Thereafter patients were dosed at the RP2D of 50 mcg. This change was implemented in response to an urgent safety measure. b) Patients receiving chronic corticosteroid treatment (longer than 8 weeks duration) for management of pre-existing AEs or patients with a history of chronic corticosteroid treatment of longer than 8 weeks' duration for AEs within 6 months of screening were to be excluded from the study. This change was implemented in response to an urgent safety measure. c) Patients with a history of adrenal insufficiency, maintained on stable replacement dose corticosteroid were eligible for the study, unless there was a history of adrenal crisis. Patients with a history of adrenal insufficiency receiving replacement dose corticosteroid were to receive prophylactic stress-dose corticosteroid prior to dosing for the first 4 IMCgp100 doses. This change was implemented in response to an urgent safety measure. d) The RP2D for Arms 1 and 2 was 50 mcg IMCgp100.

a) Procedures would be put into place to minimise the risk of severe hypotension Patients in Arm 1 were to receive lower initial IMCgp100 doses of 20 mcg on Cycle 1 Day 1 and 30 mcg on Cycle 1 Day 8, followed by 50 mcg at the third weekly dose and thereafter. b) Patients in Arm 1 required overnight hospitalisation beyond Cycle 1 Day 1 for the dose increase on Cycle 1 Day 8 and Cycle 1 Day 15. A requirement for in-patient monitoring at Cycle 1 Day 22 was to be determined based on the occurrence of hypotension requiring medical intervention at previous cycles. c) Patients experiencing a break or delay in treatment of >2 weeks who had previously experienced Grade 3 or Grade 4 hypotension following IMCgp100 dosing were to be hospitalised for their first dose following the break or delay.