A Phase II Study of Vibecotamab, a CD3-CD123 Bispecific T-Cell Engaging Antibody, for MDS or CMML after Hypomethylating Failure and in MRD-Positive AML

Abstract

Background Leukemic stem cells have high expression of CD123 compared to normal hematopoietic stem cells and is therefore a therapeutic target in multiple leukemias including acute myeloid leukemia (AML), myelodysplastic syndrome (MDS), and chronic myelomonocytic leukemia (CMML). Vibecotamab (formally XmAb14045) is a CD3-CD123 bispecific engaging antibody that has shown clinical activity in relapsed/refractory AML, particularly in low-blast disease. We therefore sought to evaluate vibecotamab in other low-blast states, including MDS or CMML after hypomethylating agent failure and MRD-positive AML. Methods In this two-arm, open-label, phase II study, adults with either MDS (IPSS-R intermediate or higher risk) or CMML (CMML-1 or CMML-2) after failure of hypomethylating agents or AML in first or second morphologic remission with detectable MRD at a level of ≥0.1% by flow cytometry were eligible. CD123 expression ≥20% on aberrant myeloid blasts was required for enrollment. Vibecotamab was given IV in a ramp-up dose schedule on days 1 (0.43µg/kg), 3 (0.75µg/kg), 5 (1.1µg/kg), and 8 (1.7µg/kg) in cycle 1, followed by weekly doses of vibecotamab at a dose of 1.7µg/kg. Patients (pts) received up to 4 cycles of vibecotamab in 28-day cycles. The primary endpoint of the MDS/CMML cohort was response rate (CR + mCR + PR + HI + clinical benefit) within 4 cycles. The primary endpoint of the AML MRD cohort was the MRD negativity rate within 4 cycles. Results Between May 2022 and July 2023, 23 pts were treated (11 MDS/CMML, 12 AML MRD). Baseline characteristics are shown in Table 1. In the MDS/CMML cohort, 7 pts (63%) had received two or more prior lines of therapy, 8 pts (73%) had prior venetoclax exposure, and 2 pts (18%) had prior hematopoietic stem cell transplant (HSCT). Six MDS pts (66%) were IPSS-R high or very high risk. In the AML cohort, 6 pts (50%) had received two or more lines of prior therapy, 11 pts (92%) had prior venetoclax exposure, and 5 pts (42%) had prior HSCT. Ten pts (83%) were ELN 2022 adverse risk. The baseline CD123 expression was 72% (range, 43%-99%) in the MDS/CMML cohort and 89% (range, 55%-99%) in the AML MRD cohort. The baseline MRD by flow cytometry in the AML MRD cohort was 1.1% (range, 0.1%-3.9%). Responses are shown in Table 2. In the MDS/CMML cohort, 7 pts responded (64%), with 6 pts (56%) achieving marrow complete remission (mCR) and 1 pt (9%) achieving hematologic improvement (HI) per International Working Group (IWG) 2006 criteria. Among the 9 MDS pts, 4 (44%) achieved mCR + HI (2 HI-N, 1 HI-P, and 1 HI-P + HI-N), and 1 (11%) achieved HI (HI-N + HI-P). Per revised IWG 2023 MDS response criteria, 5 of the 9 MDS pts (56%) achieved complete remission with limited count recovery (CR L). Two of 4 MDS pts (50%) with TP53 mutations achieved CR L. Both CMML pts achieved mCR, with one pt also achieving HI-N. Among 9 pts with baseline bone marrow blasts =>5% at trial enrollment, 6 (67%) achieved a mCR, with or without HI. Best response occurred after the first cycle in all pts. CD123 expression was not associated with likelihood of response. Of the 7 responders, 5 are in ongoing response (range 0.3-6.9 months), one died in CR L from non-hematologic complications (heart failure), and one relapsed 5 months after achieving CR L. Of the 12 pts in the AML MRD cohort, 3 (25%) achieved MRD negativity, all of which occurred after 1 cycle of vibecotamab. Among the 3 responders, all were ELN adverse risk and had prior venetoclax exposure, 2 had prior HSCT, and 1 had inv(3). The median MRD and CD123 expression in responders was 0.2% (range 0.1%-0.2%) and 98% (range 95%-99%) vs 1.8% (range 0.5%-3.9%) and 87% (range 66%-96%) in non-responders, respectively. At last follow-up, all 3 responders are still in MRD-negative remission (range 3.2-12.8 months). Vibecotamab was well-tolerated with no pts requiring dose reductions or being taken off study due to adverse events. Ten pts (44%) experienced grade 2 infusion reactions and 1 pt (4%) experienced a grade 3 infusion reaction. Myelosuppression was minimal, consistent with previous studies of vibecotamab. Conclusion Vibecotamab was safe and active in low-blast, high-risk myeloid diseases, with a response rate of 64% in MDS/CMML after HMA failure and 25% in MRD-positive AML. The clinical activity of vibecotamab, including in pts with prior venetoclax exposure and/or HSCT, and its lack of clinically significant myelosuppression provide rationale to combine it with other agents in AML, MDS, and CMML.