Nivolumab plus Azacitidine Shows Encouraging Activity in R/R AML or as Frontline Therapy in Elderly Patients with AML

February 2018, Vol 9, No 1 | Payers’ Perspectives In Oncology: ASH 2017 Highlights - ASH 2017

Available preclinical and clinical evidence suggests that inhibition of PD-1/PD-L1 pathways increases antileukemic responses in acute myeloid leukemia (AML). Moreover, azacitidine treatment results in upregulation of PD-1 signaling, which is associated with azacitidine resistance. Based on this rationale, the current study evaluated the safety and efficacy of combination nivolumab plus azacitidine treatment in 2 patient cohorts: those with relapsed or refractory (R/R) AML with poor-risk features, and in elderly patients with untreated AML.

In this single-institution, nonrandomized, phase 1b/2 trial, eligible patients were enrolled in 3 independent patient cohorts: cohort 1, comprised of patients with R/R AML who had adequate performance status (Eastern Cooperative Oncology Group ≤2) and organ function; cohort 2, comprised of AML patients ≥65 years and not suitable for induction therapy who received frontline nivolumab treatment; and cohort 3, comprised of patients in first and second salvage therapy for AML. All 3 cohorts received nivolumab (3 mg/kg on days 1 and 14) plus azacitidine (75 mg/m2 on days 1-7) every 4 to 5 weeks indefinitely, with cohort 3 receiving additional ipilimumab. The primary objective was best International Working Group response at any time on study; secondary objectives included progression-free survival, overall survival (OS), safety, and correlative studies.1

In the dose-escalation phase (n = 6), the dose-limiting toxicity was grade 3 pneumonia, with 75 mg/m2 established as the recommended phase 2 dose. A total of 70 eligible patients with R/R AML have been treated in cohort 1 and serve as the intent-to-treat population of the current analysis; 14 patients have been enrolled in cohort 2 (12 evaluable), and 12 in cohort 3 (not included in this analysis).

The median age of patients in cohort 1 was 70 years (range, 22-90 years); 44% had secondary AML, 34% had poor-risk cytogenetics, and they had received a median of 2 salvage therapies (range, 1-7). Common mutations harbored by evaluable patients (n = 28) included TP53 (23%), DNMT3A (17%), ASXL1 (16%), and TET2 (16%).

At a median follow-up of 13.3 months, 16 (23%) patients achieved complete remission (CR), CR with insufficient recovery of counts (CRi) or partial response (PR); 7 (10%) showed hematologic improvement (HI), 17 (24%) had ≥50% bone marrow blast reduction, and 5 (7%) had stable disease >6 months. The response rate was similar in patients ≤65 years versus >65 years of age, and was higher in patients with diploid cytogenetics, in patients without prior therapy with a hypomethylating agent, and in patients with ASXL1 mutations. The median OS was 17.1 months in CR/CRi patients, 11.9 months in HI patients, and 5.6 months in nonresponders. The median OS associated with nivolumab plus azacitidine was found to be significantly higher than with hypomethylating agent–based salvage protocols from MD Anderson Cancer Center in similar settings (10.6 vs 3.9 months; P <.001).

Grade 3/4 immune-related adverse events (irAEs) were observed in 8 (12%) patients, and grade 2 irAEs were seen in 8 (12%) patients; median time to onset was 6 weeks (range, 4 days-14 weeks). Common grade 2 to 4 irAEs included pneumonitis, colitis, nephritis, skin rash, and hypophysitis.

At data cutoff, cohort 2 had enrolled 14 patients with frontline AML aged ≥65 years. Median age was 73 years (range, 68-82 years), 3 patients had secondary AML, and 7 patients had diploid cytogenetics. Of the 12 patients evaluable for response, 8 patients achieved CR/CRi/PR with insufficient recovery of counts (median follow-up was 5.9 months).

Based on these results, it was concluded that combination therapy with nivolumab plus azacitidine yielded encouraging response rates, with improved OS compared with historical data in patients with R/R AML and poor-risk features.


  1. Daver N, et al. 2017 ASH. Abstract 1345.