Dr Shallis on Data for Pembrolizumab Plus Venetoclax/Azacitidine in AML

“[This was] a negative trial, which is a disappointment, but we learn a lot from these trials. This is a trial that is enriched for a number of correlative analyses being done through the CIMAC collaboration with a number of outstanding laboratory and translational colleagues. Most of these are still underway, [and] we hope that the data generated from this trial will inform the next generation of trials.”

Rory Shallis, MD, assistant professor of medicine (hematology), Yale School of Medicine, Yale Cancer Center, discusses findings from the phase 2 BLAST AML-2 trial (NCT0428478) evaluating the addition of pembrolizumab (Keytruda) to venetoclax (Venclexta) and azacitidine in patients with acute myeloid leukemia (AML). Findings were presented by Shallis and colleagues at the 2024 ASH Annual Meeting.

The investigator-initiated trial included 60 patients who were randomly assigned 1:1 to receive either venetoclax plus azacitidine alone (n = 31), or pembrolizumab plus venetoclax and (n = 29). In both arms, 29 patients received treatment. The primary end point was the rate of measurable residual disease (MRD)–negative complete remission (CR) or CR with incomplete hematologic recovery (CRi) within the first 6 treatment cycles.

Shallis indicated that no significant efficacy differences were seen between the two arms; the MRD-negative CR/CRi rate was 44.8% in both groups (P = 1.000), and the CR rates were both 24.1% (P = .999). Similarly, best composite response rates—comprised of CR plus CRi and morphologic leukemia-free survival) were 89.7% in both the control arm and the triplet arm (P = .528). Median overall survival (OS) was numerically shorter in the triplet arm at 8.87 months compared with 17.94 months in the control arm, although Shallis notes that this difference was not statistically significant (P = .137).

Safety data revealed that rates of treatment-related adverse effects (TRAEs) were comparable across arms, with any-grade TRAEs occurring in 89.7% of patients in the control arm vs 93.1% of patients in the triplet arm. However, the triplet arm had a higher incidence of serious AEs (274 events vs 192 events) and a higher proportion of patients who discontinued treatment due to TRAEs (6 vs 1). There were no new safety signals observed with pembrolizumab.

The study was discontinued early following a recommendation from the data safety monitoring board due to futility. Exploratory correlative analyses conducted through the CIMAC collaboration will aim to identify biomarkers of response and guide the design of future immunotherapy trials in AML.

Shallis concludes that although the addition of pembrolizumab to azacitidine and venetoclax did not improve outcomes, the findings underscore the importance of conducting randomized trials in AML. Ongoing translational studies are expected to inform the next generation of combination regimens in this challenging patient population.