ESMO 2014: Results of a Phase III Randomised Study of Nivolumab in Patients with Advanced Melanoma After Prior Anti-CTLA4 Therapy

PD-1 checkpoint inhibitor, nivolumab vs investigator’s choice chemotherapy in the second-line treatment of patients with advanced melanoma

In patients with metastatic melanoma who progressed on or after anti-CTLA-4 therapy and treatment with BRAF inhibitors in the case of BRAF mutation positive disease, nivolumab was well tolerated and showed a higher overall response rate (ORR) when compared with investigator’s choice chemotherapy. Durable tumour regression was observed in the majority of responders to nivolumab. The results of a phase III randomised, open-label study were presented by Dr Jeffrey Weber of Lee Moffitt Cancer Center & Research Institute, Tampa, USA at the Presidential Symposium 2 during ESMOCongress 2014 in Madrid, Spain.

Effective therapies are needed for patients with melanoma who progress on or after anti-CTLA-4 therapy and a BRAF inhibitor.

Nivolumab is a fully human IgG4 monoclonal antibody that inhibits the PD-1 immune checkpoint protein. In early studies, single-agent nivolumab demonstrated meaningful clinical activity and a manageable safety profile in pretreated patients with advanced melanoma with promising overall survival (OS) rates of 63%, 48%, and 41% observed at 1-, 2-, and 3-years, respectively.

In this phase III open-label trial, patients with advanced melanoma who progressed on or after anti-CTLA-4 therapy and a BRAF inhibitor in case of BRAF V600 mutation positive disease were randomised 2:1 to receive

  • nivolumab 3 mg/kg i.v. every two weeks (268 treated patients)
  • or investigator’s choice chemotherapy (dacarbazine 1000 mg/m2 every 3 weeks, or carboplatin AUC 6 plus paclitaxel 175 mg/m2 every 3 weeks)

(102 treated patients) until progression or unacceptable toxicity. Patients  receiving nivolumab may be treated beyond initial progression if considered by the investigator to be experiencing clinical benefit.

Patients were stratified by PD-1 ligand expression status (PD-L1 positive vs PD-L1 negative/indeterminate; PD-L1 positive status was defined as ≥ 5% tumour cell surface staining cut-off by immunohistochemistry); BRAF status (BRAF wild-type vs BRAF V600 mutant); and best overall response to prior anti-CTLA-4 therapy (clinical benefit defined as best overall response that included complete response/partial response/stable disease) vs no clinical benefit (progressive disease).

Exclusion criteria were active brain metastases; prior therapy with anti-PD-1, anti-PD-L1 or anti-PD-L2 antibodies; grade 4 toxicity or use of infliximab to manage adverse events from prior ipilimumab treatment and ocular melanoma.

Co-primary endpoints were ORR by independent radiology review committee and OS. Secondary objectives included to compare progression-free survival (PFS) of nivolumab to investigator’s choice chemotherapy at the time of OS analysis and to evaluate PD-L1 expression as a predictive biomarker for ORR and OS. However, OS analysis had not taken place at the time of interim ORR analysis.

Response according to RECIST 1.1 criteria was assessed 9 weeks after randomisation, followed by 6-week assessments for the first 12 months and then by assessments every 12 weeks.

Response to Nivolumab vs Chemotherapy

Response to Nivolumab vs Chemotherapy. © Jeffrey Weber

The ORR was assessed as planned in the first 120 patients treated with nivolumab and 47 patients who received investigator’s choice chemotherapy with follow-up of at least 6 months.

Baseline age, sex and metastasis stage were balanced between the arms. However, there were slightly more patients with history of brain metastasis and elevated LDH in nivolumab arm.

Median time on therapy was 5.3 months in the nivolumab  arm and 2 months in investigator’s choice chemotherapy arm. Disease progression was the most common reason for discontinuation in the nivolumab (43%) and investigator’s choice chemotherapy arms (61%).

Confirmed ORR by independent radiology review committee in nivolumab and patients treated with investigator’s choice chemotherapy was 32% and 11%, with a median time to response of 2.1 months (range: 1.6, 7.4) and 3.5 months (range: 2.1, 6.1), respectively.

Median duration of response for nivolumab was not reached (range: 1.4+, 10+ months) with 36 (95%) patients being still in response. Median duration of response in patients treated with investigator’s choice chemotherapy was 3.6 months (range: 1.3+, 3.5) with 4 (80%) patients still in response.

Consistently higher clinical activity was observed for nivolumab vs investigator’s choice chemotherapy regardless of pre-treatment PD-L1 expression status, BRAF mutation status and prior anti-CTLA-4 benefit.

Reduction of ≥50% in target lesion burden occurred in 82% (31/38) of nivolumab responders and 60% (3/5) of responders in the investigator’s choice chemotherapy arm. Among nivolumab-treated patients, an additional 10 (8%) patients had immune-related response patterns observed (≥30% reduction in target lesion tumour burden).

Grade 3-4 drug-related adverse events were seen in 9% and 31% of patients treated with nivolumab and investigator’s choice chemotherapy, respectively. Discontinuations due to drug-related adverse events of any grade occurred in 2% and 8% of treated patients, respectively. No drug-related grade 3–4 adverse events were reported in ≥2% of nivolumab treated patients. All grade 3–4 drug-related adverse events belonging to the select adverse event categories resolved. Corticosteroids were the most common immunosuppressive medication used.

There were no deaths related to study drug toxicity. One patient in the nivolumab group experienced grade 5 hypoxia, possibly pneumonitis, in the setting of lymphangitic spread and possible pneumonia; this patient’s cause of death was classified by the investigator as ‘other’ rather than ‘study drug toxicity’.


The study authors concluded that in patients with advanced melanoma who have progressed despite anti-CTLA-4 therapy and BRAF inhibitors if BRAF mutated, nivolumab monotherapy demonstrated superior efficacy to investigator’s choice chemotherapy. The majority of nivolumab treatment-related adverse events were of a low grade and manageable using recommended treatment algorithms. Co-primary endpoint (OS) data is pending at the time of presentation.

Dr Ignacio Melero, who discussed the study results, said that immunotherapy of cancer is no longer a quixotic goal. In his talk, Dr Melero highlighted the recent approvals of nivolumab in Japan and pembrolizumab in USA for the treatment of advanced melanoma. He further said that chemotherapy might be eliminated as an option in BRAF wild-type disease but the clinical trials are important because the best is yet to come. This is particularly true for combination therapy. In terms of biomarker analysis from tumour biopsy, antigenicity, immunogenicity, immune escape, and T-cell infiltrates should be considered. PD-L1 as single parameter is not good enough, according to Dr Melero. The scenario might be going towards a multiparameter scores (including tumor PD-L1 status…quantitatively graded). The highest number of mutations in solid tumours is observed in melanoma and non-small-cell lung cancer with lower frequencies in other tumour type and in that regard, melanoma remains at the top of the iceberg.


Abstract LBA3_PR - A phase 3 randomized, open-label study of nivolumab (anti-PD-1; BMS-936558, ONO-4538) versus investigator’s choice chemotherapy (ICC) in patients with  advanced melanoma after prior anti-CTLA4 therapy

The study was supported by Bristol-Myers Squibb, Ono Pharmaceutical Company, Ltd. and Dako (for collaborative development of the automated PD-L1 immunohistochemistry assay).