DARTs point the way forward in AML

In this issue of Blood, Uy et al report the results of a multicenter phase 1/2 study of a CD3ε × CD123 bispecific drug in patients with relapsed/refractory acute myeloid leukemia (AML).¹  The drug, flotetuzumab, belongs to a class of dual affinity retargeting proteins (DARTs) that bring together effector cells (for example, T cells) with target cells (for example, CD123-expressing AML blasts). The concept is similar to that of blinatumomab, a CD3ε × CD19 bispecific T-cell engager now in routine clinical use in B-cell acute lymphoid leukemia. The target molecule, CD123, is not unique to AML blasts and is also expressed on some normal hematopoietic cell populations.²  Nonetheless, CD123 is among the leading potential cell surface targets for immunotherapy in AML with 24 active studies listed on www.clinicaltrials.gov (accessed 7 September 2020).

The dose-finding cohort of the trial enrolled 42 patients and led to a “recommended phase 2 dose” of 500 ng/kg per day by continuous infusion after a stepwise dose escalation. Forty-six patients were then enrolled to the dose expansion cohort. Almost all patients treated at the recommended phase 2 dose experienced infusion-related reactions/cytokine release syndrome, although these were generally mild and easily managed with dose interruptions and/or tocilizumab. Neurologic adverse events were rare. Because most patients were cytopenic at baseline, it is difficult to disentangle disease-associated from treatment-associated cytopenias and hence, in my view, the true hematopoietic toxicity of the drug remains unclear. At the recommended phase 2 dose, 12/50 (24%) patients experienced a response, with 9/50 (18%) attaining a complete response with or without complete hematologic recovery. The median duration of response in patients classified as achieving a complete response with or without complete hematologic recovery was 6.9 months.

The authors then performed 2 interesting exploratory subgroup analyses: In the first, patients in primary induction failure (PIF) or early relapse (ER) occurring within 6 months after initial treatment were found to be more likely to respond to flotetuzumab (response here defined as >50% blast reduction) compared with those patients who were enrolled with late AML relapse or after failure of hypomethylating agents (43% vs 14%, respectively). In the second exploratory analysis, the overall response rate was 5/9 (56%) in patients with 2 lines of prior treatment vs 0/7 (0%) in those with more than 5 lines of prior treatment. These 2 analyses are of course interrelated because patients with PIF are likely to have had fewer lines of treatment, making it premature to conclude that the drug truly is more active in patients with PIF/ER solely on biological grounds. The conclusion may well be that patients with more advanced disease (who have failed more lines of therapy) are more difficult to treat, which will not come as a surprise to any practicing hemato-oncologist.

Nonetheless, the authors then sought biomarkers of response by building upon immune gene signature scores identified in their recent publication.³  Complete response after flotetuzumab positively correlated with high immune infiltration scores, and these were also higher in patients with PIF/ER compared with patients who entered the trial with late AML relapses. Superficially, high immune infiltration was related to expression of antigen processing machinery and inflammatory chemokine genes, although careful examination of the differentially expressed genes will likely yield further hypothesis-generating insights regarding the mechanism of action of flotetuzumab in particular and/or susceptibility of AML to immunotherapy in general.

This report is interesting and important because it represents one of the first complete datasets of immunotherapy in AML (excluding allogeneic hematopoietic cell transplantation). Risk categorization in AML links molecular and cytogenetic aberrations to prognosis, largely via resistance by AML blasts to “traditional” chemotherapeutic agents.⁴  In turn, the presence of residual disease at allogeneic hematopoietic cell transplantation is highly predictive of relapse.⁵  The hope is that T-cell–based immunotherapies that do not rely on the mechanism of action of chemotherapy may be agnostic to the traditional adverse cytogenetic or molecular risk factors. There may be a faint signal from this dataset that this is the case, and it will be important to expand upon and validate these observations in future, larger studies.

Years of experience with the CD3ε × CD19 bispecific T-cell engager blinatumomab have shown us that the quantity and quality of remissions are higher and the side effects lower when used in patients with measurable residual disease rather than in active disease.^6,7  Having now demonstrated the clinical activity of flotetuzumab, one wonders whether the same concept will hold true in AML.