Phase 1b/2a Study of AZD4573 (CDK9i) and Acalabrutinib in Patients with Relapsed/Refractory Diffuse Large B-Cell Lymphoma (r/r DLBCL): Results from Dose-Escalation

Background: Inhibition of Cyclin-dependent kinase 9 (CDK9), a transcriptional regulator, leads to short-term reduction of anti-apoptotic and oncogenic proteins. AZD4573, a potent and highly selective inhibitor of CDK9, triggers downregulation of MCL-1, BFL-1 and MYC, rapidly inducing apoptosis in human hematological cancer cell lines. The safety of AZD4573 as monotherapy was previously established in a Phase 1 trial in hematological malignancies (NCT03263637). Inhibition of Bruton tyrosine kinase with acalabrutinib increases expression of pro-apoptotic proteins such as Bim. Combining AZD4573 with acalabrutinib in DLBCL in in vitro models accelerates induction of cleaved Caspase-3, suggesting a synergistic mechanism that pushes malignant cells closer to their apoptotic threshold. Here, we report the dose-escalation of a multicenter, open-label, Phase 1b/2a study to assess safety and tolerability of AZD4573 with acalabrutinib in patients with r/r DLBCL (NCT04630756).

Methods: Patients with r/r DLBCL with ≥2 prior lines of therapy (stem cell transplant or CAR-T allowed) were eligible. AZD4573 was given intravenously weekly and acalabrutinib 100 mg was given orally twice daily. In cycle 1, there was a weekly intrapatient dose ramp up of AZD4573 with a target dose of 9 mg in cohort 1 (3 mg, 6 mg, 9 mg) and 12 mg in cohort 2 (6 mg, 9 mg, 12 mg). Treatment was given until progressive disease/unacceptable toxicity. Responses were assessed by Lugano 2014 criteria and adverse events (AEs) were graded using CTCAE v5.0. The primary objective was to assess safety and determine the recommended phase 2 dose; pharmacokinetics and antitumor efficacy were secondary and exploratory endpoints, respectively.

Results: As of August 2, 2022, 21 patients have been dosed. At data cutoff (DCO) on May 17, 2022, 13 patients were evaluable for safety (cohort 1, n=6; cohort 2, n=7) and 11 were evaluable for response (cohort 1, n=4; cohort 2, n=7). Median age was 61 years, 69.2% of patients were male, and median number of prior lines of treatment was 4 (range 2–8). The safety set included 8 DLBCL not otherwise specified (NOS), 1 high-grade B-cell lymphoma, 1 primary mediastinal B-cell lymphoma, and 3 patients with other subtypes, all T-cell rich. Cell of origin, determined locally, was reported in 10 patients (6 germinal center B-cell-like [GCB]; 4 non-GCB). At DCO, median duration of AZD4573 treatment was 4.0 weeks (range, 1–53) in cohort 1 and 12.6 weeks (range, 2–23) in cohort 2. AEs considered possibly related to either treatment occurred in 76.9% of patients and were grade ≥3 in 69.2% (Table 1). The most common AEs possibly related to AZD4573 were neutropenia (69.2%), thrombocytopenia (38.5%), alanine aminotransferase increased (38.5%) and aspartate aminotransferase increased (38.5%). Liver function test increases were short-lived with spontaneous resolution before subsequent weekly infusion. These findings were mainly due to down-modulation of hepatic transporter proteins and reduced clearance rather than hepatocellular impairment (based on clinical, in-vitro and in-silico assessments). Clinical tumor lysis syndrome (TLS) related to AZD4573 was not reported in cohort 1 and occurred in 1 of 7 patients in cohort 2 (TLS and transient 1.5xULN creatinine increase). No treatment-emergent AEs led to death. One patient in cohort 1 discontinued treatment due to grade 2 fatigue related to AZD4573. No dose limiting toxicities (DLTs) were observed in either cohort. The objective response rate (ORR) was 63.6% and the complete response (CR) rate was 36.4%. Of 4 response-evaluable patients in cohort 1, 1 had a CR and 1 had a partial response (PR), for an ORR of 50.0%. Of 7 evaluable patients in cohort 2, 3 had CRs and 2 had PRs, for an ORR of 71.4% (Table 2). Notably, 3/4 patients with prior CAR-T failure responded (2 CRs and 1 PR). Responses were seen in both GCB (1 CR/2 PRs in 5 evaluable patients) and non-GCB (2 CRs in 3 evaluable patients) subtypes.

Conclusions: In this heavily pretreated population including CAR-T failures, the combination of AZD4573 and acalabrutinib resulted in high response rates (ORR 63.6%, CR rate 36.4%). No DLTs were detected, the combination was well tolerated with a manageable safety profile, and no new safety signals were identified. The study is currently enrolling in the Phase 2a expansion phase and updated data will be presented at the meeting.