Accrual Barriers and Detection of Early Toxicity Signal in Older Less-Fit Patients Treated with Azacitidine and Nivolumab for Newly Diagnosed Acute Myeloid Leukemia (AML) or High-Risk Myelodysplastic Syndrome (MDS) in the SWOG 1612 Platform Randomized Phase II/III Clinical Trial

Background: Less-fit patients with acute myeloid leukemia (AML) or high-risk myelodysplasia (MDS) age 60 years and older constitute the majority of patients with AML/MDS but are not well represented in clinical trials. DNA-methyltransferase inhibitor (HMA) monotherapy (e.g. azacitidine or decitabine) is usual. Overall response rates (ORR) are low; improvement in overall survival relative to supportive care alone is modest, highlighting the critical need for efficient identification of effective novel therapies. Blocking programed cell death protein-1 (PD-1) signaling with nivolumab may increase the sensitivity of AML cells to azacitidine and improve outcomes and is the focus of one arm of the S1612 trial. Signaling through PD-1 contributes to tumor immune evasion and growth in AML [Chen, Cancer Biol Ther 2008]. Increased expression of PD-1 (~40% of AML), is associated with poor HMA response [Ørskov, Oncotarget 2015]. A single center azacitidine/nivolumab non-randomized phase II study in relapsed/refractory AML reported ORR of 33% (23/70), including 22% complete remission/complete remission with incomplete hematologic recovery [Daver, Cancer Discovery 2018]. About 25% of the patients developed grade 2-4 immune toxicities; nivolumab immune-related adverse events led to treatment discontinuation in nearly 1 in 7 patients.

Study Design and Methods:The S1612 trial [NCT03092674] is a platform randomized phase II/III clinical trial with a common azacitidine control arm (CA) and two currently active experimental arms (EA). Therapy is intended for community setting: azacitidine/nivolumab; and azacitidine/midostaurin. The innovative design utilizes a phase II go/no-go decision comparing each EA with the CA independently when there are 100 pts/arm and 104 deaths on the EA and CA combined. EAs will proceed to phase III if the null hypothesis (HR=1) is rejected in favor of the EA (15% one-sided alpha). When the two currently active EAs complete phase II accrual, a third EA (decitabine/cytarabine) will open. The CA stays open the entire length of the trial and only concurrently randomized patients will be compared across arms.

If an EA proceeds to phase III, 200 additional patients (300 total patients) will be accrued. Phase III analysis occurs 1.5 years after accrual or at 414 deaths (in the respective EA and CA), whichever comes first. The phase III tests the null hypothesis with 4.5% two-sided alpha and 83% power for each EA to detect an improvement in median OS from 10.4 to 15.6 months. Eligible pts are age ≥60 years, newly diagnosed AML with ≥ 20% blasts or myelodysplastic syndrome with excess blasts-2 (MDS-EB-2); deemed by the treating physician unfit for standard cytotoxic chemotherapy; and no prior HMA permitted.

Trial in Progress Issues:Between December 2017 and October 2018, 113 patients were screened and 78 randomized to study treatment (median/range age: 75/61-86 years; median/range performance status 1/0-3). Two concerns challenged this trial: 1) required administration of 7-day azacitidine at the enrolling sites created a burden for this population and 2) an early excess grade 5 toxicity signal in the azacitidine/nivolumab arm compared with the control arm. Without a control comparison, this safety signal likely would have been missed. Strategies to address these concerns were developed by the study team. Discussion with the sponsor and US Food and Drug Administration (FDA) focused on allowance of standard-of-care protocol-directed azacitidine administration in the patient's primary care doctor's office rather than at the oncology center. Because of the toxicity concern, the trial was placed on partial clinical hold for further evaluation and possible nivolumab-arm eligibility changes, along with new surveillance and pre-emptive action including prompt steroid initiation for suspected immune-related toxicities. The S1612 trial highlights special concerns when enrolling vulnerable populations onto leukemia clinical trials, and the importance of collaborative strategies including with the FDA to preserve clinical trial integrity for patient benefit. It also demonstrates the efficiency this novel platform design has for therapeutic investigation and, importantly, very early identification of serious toxicity. Updates on accrual and resolution and of these issues will be presented.