Diverse Histopathologic and Molecular Responses of Acute Myeloid Leukemia to the FLT3 Inhibitor Quizartinib (AC220)

Abstract 885

Although high clinical response rates to the FLT3 inhibitor quizartinib (AC220) among subjects with relapsed/refractory AML suggest a potential clinical breakthrough, there remains concern that few subjects achieved CR or PR by strict IWG response criteria. Prior reports describe terminal neutrophil differentiation of AML blasts during clinical response to AC220 (Sexuaer, et al. ASH 2011, #943), suggesting that traditional chemotherapy metrics may not prove relevant to measure FLT3 inhibitors' clinical activity. Here we characterize responses to AC220 from a recent Phase II clinical trial, based on histomorphologic, molecular, and cytogenetic analysis and independent of that study's modified IWG metrics. We describe at least two distinct types of morphologic responses to AC220 among the 21 subjects treated at our site on this single-arm trial. All subjects had AML that relapsed or was refractory to prior chemotherapy. Subjects received single agent AC220 daily at 90–200 mg and were evaluated by bone marrow aspiration and biopsy on treatment days 15 and 29 and then monthly until blasts were cleared from blood and marrow. 15 of 21 enrolled subjects were FLT3-ITD+; 6 were FLT3 wild type (WT). 14/16 subjects with circulating or extramedullary blasts at study entry showed complete morphologic elimination of these populations. 18/20 subjects with marrow blasts >20% at study entry reduced marrow blasts to <20% and 13/20 to <5% at maximal response. Comparable blast reductions occurred in FLT3-ITD+ and FLT3-WT subjects. Importantly, concurrent with blast reduction, sustained dysmyelopoiesis occurred in nearly all subjects and robust trilineage hematopoietic reconstitution in marrow and blood was rare (2/21). Two patterns of response were seen, typically by day 29: 5/15 FLT3-ITD+ and 2/6 FLT3-WT subjects demonstrated a “cytotoxic” response, characterized by a profoundly hypocellular marrow with reduction or elimination of the leukemic clone by PCR or cytogenetic/FISH analysis. In this setting, one subject demonstrated myelofibrosis and transient proliferation of dysplastic megakaryocytes. In contrast, 7/15 FLT3-ITD+ subjects demonstrated a differentiation response with a hypercellular marrow, near absence of leukemic blasts, left shifted granulopoiesis, and marked dysplasia among differentiating cells. In subjects with differentiation response, PCR and/or FISH studies on purified circulating granulocytes or differentiating marrow showed persistence of a molecular marker of leukemia, confirming that neutrophil recovery represented leukemic differentiation and not recovering normal hematopoiesis. 6/7 subjects with a differentiation response had normal cytogenetics at enrollment, while non-responders and those with cytotoxic response generally had abnormal karyotype, many with complex aberrations. Clinically, responding subjects showed persistent cytopenias over several months of therapy, suggesting that residual bone marrow damage after the leukemic insult was not reversed by AC220 or that the potent FLT3 and KIT inhibitory activity of AC220 at the doses tested in this study impairs recovery of normal hematopoiesis. In conclusion, rather than re-growth of normal hematopoiesis, abnormal hematopoiesis is commonly seen during treatment with AC220 and is associated with variable elimination of the FLT3-ITD+ clone and differentiation of persisting neoplastic cells. It is currently not clear if these effects represent unmasking of an underlying MDS phenotype, abnormal hematopoiesis secondary to drug effects, or other mechanisms. Overall, these finding complicate analysis of treatment response to high potency FLT3 inhibitors. Despite morphologically and genetically abnormal hematopoiesis, subjects generally remained clinically stable for months, with modest transfusion requirements and few infections; 7/21 received allogeneic transplantation following AC220 response. We propose that IWG criteria may fail to appropriately estimate the clinical benefits of FLT3 inhibitors. Reversion to an MDS-like state may provide clinical benefit if sustained and/or supports a non-toxic bridge to transplant. Randomized studies will be necessary to confirm this observation. In addition, our data demonstrate that diverse biologic responses to FLT3 inhibitor therapy are seen although the biologic basis of these differences requires further investigation.