A Multi-Center Phase Ib Trial of the Histone Deactylase Inhibitor (HDACi) Entinostat in Combination with Anti-PD1 Antibody Pembrolizumab in Patients with Refractory/Relapsed Myelodysplastic Syndromes (RR-MDS) or Oligoblastic Acute Myeloid Leukemia (RR-AML) after Hypomethylating Agent (HMA) Failure

Introduction: Pts with MDS and HMA failure have a poor prognosis and need novel therapies. Success with anti-PD1 therapy in myeloid malignancies has been limited and resistance mechanisms are poorly understood. Preclinical data suggest that myeloid-derived suppressor cells (MDSCs) contribute to MDS progression (Chen et al. JCI 2013). and may mediate resistance to anti-PD1 therapy (Kim et al. PNAS 2014). As HDACi decrease MDSCs, synergistic anti-tumor efficacy with anti-PD-1 has been hypothesized leading to this phase Ib study (NCT02936752) of the HDACi entinostat and anti-PD1 antibody pembrolizumab in pts with MDS or oligoblastic AML and HMA failure.

Methods: We conducted an investigator-initiated, NCI-Cancer Therapy Evaluation Program (NCI-CTEP)-sponsored, multicenter, dose escalation and expansion phase Ib trial. Adult (≥18 years) pts with MDS or oligoblastic AML (bone marrow [BM] blasts 20-30%) and HMA failure were enrolled. Pts had to have adequate end organ function and ECOG performance status ≤2. The dose-escalation part used 2 dose levels (DL1: entinostat 8mg PO on days 1 and 8 of each 21-day cycle; DL2: entinostat 8mg PO on days 1, 8, and 15 of each 21-day cycle) with pembrolizumab 200mg IV on day 1 of each 21-day cycle starting with cycle 2. In cycle 1, pts received only entinostat to determine its impact on MDSCs separately from pembrolizumab. Toxicities were tabulated and graded according to CTCAE-5. Dose-limiting toxicities (DLT) were defined as any of the following if deemed possibly related to trial therapy: Grade (G) 3/4 non-hematologic toxicity, any ≥G3 immune related adverse events (IrAEs) or any IrAE resulting in pembrolizumab interruption or discontinuation during the first 2 cycles of combined therapy. Hematologic AEs were considered a DLT only if they were unrelated to MDS and lasted for >42 days. The primary endpoint was safety and determination of the maximum tolerated dose. The secondary endpoint was the overall response rate defined as a composite of complete remission, partial remission, and hematologic improvement (HI) per the 2006 IWG response criteria in MDS and change in MDSCs and PD1/PDL-1 in BM.

Results: Between 06/26/2017 and 02/03/2021, we enrolled 28 pts (25 MDS and 3 AML). Median age was 74 years (range [R], 57-86), all pts had received prior HMA with 5 pts (18%) having had ≥3 prior lines of therapy. Median IPSS-R at screening for MDS pts was 5.5 (R: 2-10). Baseline demographic and clinical characteristics are shown in Table 1. During dose escalation, 7 pts were enrolled in DL1 (1 pt not evaluable due to early disease progression during DLT period), and 6 pts were treated on DL2. There was 1 DLT on DL1 (G5 pneumonia/bronchoalveolar hemorrhage (BAH), while 2 pts experienced DLTs at DL2 (G3 pharyngeal mucositis and G3 anorexia). There were no hematologic DLTs during dose escalation. Thus, DL1 was expanded, and another 15 pts were enrolled at DL1.

Across the entire study, pts received a median of 3 cycles of therapy (R: 1-12 cycles). The most frequent cause of treatment discontinuation was disease progression in 9 pts. Four pts died during the study, 2 due to therapy-related complications (1 from pneumonia/BAH, 1 from multi-organ failure which occurred after the DLT period in DL2). A summary of the AEs is presented in Table 2. Hematologic AEs were common. The most common non-hematologic ≥G3 AEs independent of treatment attribution were infection (32.1%), hypoxia/respiratory failure (10.7%) and dyspnea (10.7%). Potential IrAEs of any grade were reported in 5 pts including G3 pneumonitis in 2 pts. Serious AEs were reported in 14 pts; 8 of which were potentially attributed to treatment.

Among the 19 pts with the first BM efficacy assessment after 3 months, no pt achieved a protocol-defined response. However, 2 pts had mCR (BM blasts decreased from a baseline of 7% and 10% to 2% and 4%, respectively) without HI. Twelve pts had stable disease, while 5 had progressive disease. With a median follow-up of 5.9 months (R: 0.5 – 59.1 months), median OS was 6.9 months (95% CI: 4.8 – 13.6 months). Correlative studies are being conducted.

Conclusion: In this multi-center phase 1b investigator-initiated trial in MDS and oligoblastic AML pts after HMA failure, entinostat + pembrolizumab at DL1 was reasonably tolerated but had limited therapeutic efficacy. Ongoing correlative work will explore the disconnect between the preclinical and clinical data regarding potential synergy of entinostat + pembrolizumab.