ADM28 in CD123-positive R/R AML
| Image credit: © Свет Лана – stock.adobe.com
AFM28 led to disease control in most patients with CD123-positive, relapsed/refractory acute myeloid leukemia (AML) across 5 dose levels and demonstrated a manageable safety profile that manifested most often in infusion-related reactions (IRRs), according to preliminary data from the phase 1 dose-escalation AFM28-101 trial (NCT05817058) presented during the 51st Annual EBMT Meeting.1
Regarding efficacy, 1 complete response (CR) occurred at dose level 5; this lasted for 6 cycles. Four of 10 evaluable patients at dose level 6 achieved a CR or CR with incomplete hematologic recovery as the best response to date, reflecting a composite CR rate of 40%.
“In the ongoing trial, AFM28 monotherapy demonstrated early signs of clinical efficacy and a well-managed safety profile up to a weekly flat dose of 300 mg intravenously in patients with relapsed/refractory AML. The efficacy was observed independent of the adverse baseline mutational status,” presenting study author Lydia Wunderle, MD, medical director of research and development at Affimed, and coauthors, wrote in the poster.
CD123 is universally and highly expressed in AML on leukemic blasts and stem and progenitor cells. Overexpression is linked to tumor growth and poor prognosis. AFM28 is a high-affinity bispecific, tetravalent antibody-based construct that redirects natural killer (NK) cells to leukemic cells by binding to CD16A on NK cells and CD123 on AML cells. Through this mechanism, AFM28 eliminates CD123-expressing AML blasts and leukemic stem and progenitor cells. Moreover, AFM28-redirected NK cell–mediated antibody-dependent cellular cytotoxicity is active even at low CD123 expression levels and is not affected by CD64-mediated tumor escape, rendering it a more effective option than conventional CD123-targeting fragment crystallizable–enhanced antibodies.
The open-label, first-in-human, multicenter, phase 1 study enrolled adult patients with primary refractory or multiple-relapsed, CD123-positive AML following no more than 3 prior regimens. Patients who were eligible for transplant at the time of study entry were denied enrollment.
Dose-escalation/validation was performed using a Bayesian logistic regression model.
AFM28 was administered intravenously in flat weekly dosing in 28-day cycles. Six dose levels were evaluated: 25 mg (n = 2; dose level 1); 50 mg (n = 2; dose level 2); 100 mg (n = 2; dose level 3); 200 mg (n = 4/6; dose level 4); 250 mg (n = 6; dose level 5); and 300 mg (n = 11; dose level 6).
The primary end points were the maximum tolerated dose, and/or one or more recommended phase 2 doses; and the incidence of dose-limiting toxicities (DLTs). Secondary end points included safety, pharmacokinetics/pharmacodynamics, immunogenicity, and efficacy.
Of all patients in the safety set (n = 29) the mean age was 67 years (range, 28-82) and most patients were female (52%) with an ECOG performance status of 1 (52%). The mean bone marrow blast count at screening was 22%. Represented mutational subgroups at screening included those with alterations in IDH2 (14%), TP53 (17%), RUNX1 (31%), other (ASXL1/SRSF2; 35%), and unknown (21%). The median number of prior therapies was 2 (range, 1-3) and 58% of patients were refractory to their last line of therapy. Prior therapies included venetoclax (Venclexta; 69%), a hypomethylating agent–containing regimen (76%), 1 to 2 anthracycline-containing regimens (55%), and prior stem cell transplant (21%).
“The majority of patients have a high risk for relapse and refractoriness due to age, mutational status, and insufficient response to prior treatments,” the authors wrote.
Any adverse effects (AEs) included infusion-related reaction (any grade, 45%), febrile neutropenia (any grade, 28%; grade ≥3, 28%), hyperglycemia (any grade, 28%; grade ≥3, 3%), pneumonia (any grade, 24%; grade ≥3, 24%), anemia (any grade, 21%; grade ≥3, 17%), and asthenia (any grade, 21%; grade ≥3, 3%).
Any-grade AEs related to AFM28 occurred in 62% of patients (grade ≥3, 17%). The most frequent of these events was grade 1 or 2 IRR (grade 1/2, 45%).
Two cases of cytokine release syndrome (CRS) occurred, deemed grade 1 at dose level 6 and grade 3 at dose level 2 per ASTCT guidelines. “All CRS and IRR events were fast resolving without the use of tocilizumab [Actemra]. No neurotoxicity events occurred,” the study authors said.
One DLT occurred at 200 mg in a patient with possibly related fatal pneumonitis. No other pneumonitis events, DLTs, or unexpected safety signals occurred.
With respect to pharmacokinetic analyses, investigators reported non-linear kinetics caused by target-mediated drug disposition, which the authors explained was “likely due to binding to CD16 and CD123 receptors.” Relevant AFM28 Ctrough accumulation was also seen at doses of 250 mg or higher. At the 300-mg dose, steady-state appeared to have been reached after 5 to 6 weeks.
Assessment of pharmacodynamics revealed that the correlation between receptor occupancy (RO) and exposure illustrated early saturation for CD123 reaching a plateau at lower dose levels. Additionally, CD16 saturation on NK cells showed a dose-dependent increase in RO with each dose level, and the highest RO was reached at 250 mg and 300 mg without reaching a plateau. “RO levels measured for the higher doses of 250 mg and 300 mg AFM28 may be associated with the efficacy observed in these cohorts,” the authors noted.
“Pharmacokinetic and pharmacodynamic results, including RO dynamics, are in accordance with the preliminary clinical efficacy observed so far and together with the safety profile support further dose escalation of AFM28. The promising monotherapy activity and tolerability warrant further development of AFM28 as monotherapy and in combination with standard of care or novel therapies,” the authors concluded.
Disclosures: No disclosures were listed.
Reference
Recher C, Arnan M, de Botton S, et al. Safety and efficacy of AFM28, a CD123-targeting innate cell engager in patients with relapsed/refractory CD123-positive acute myeloid leukemia: interim results from a phase 1 study. Presented at: 51st Annual EBMT Meeting; March 30-April 2, 2025; Florence, Italy. Abstract B223.