Real-World Treatment Patterns and Effectiveness of Midostaurin Versus Quizartinib in FLT3-ITD Mutated Acute Myeloid Leukemia Undergoing Intensive Induction

Introduction: Mutations in FMS-like tyrosine kinase 3 (FLT3) are one of the most common genetic alterations in acute myeloid leukemia (AML), with internal tandem duplication (ITD) conferring increased risk of relapse and mortality historically. Midostaurin and quizartinib are FDA-approved tyrosine kinase inhibitors for newly diagnosed FLT3-ITD mutated AML based upon superior survival of each drug compared to placebo when added to standard cytarabine and anthracycline (7+3) induction and ongoing intensive post-remission therapy. Given the lack of prospective randomized comparisons of these FLT3 inhibitors, we evaluated patient characteristics, treatment patterns and associated outcomes of real-world AML patients treated with 7+3 and either midostaurin or quizartinib. Methods: This was a retrospective multicenter cohort study of 5 academic hospital systems in the US. Newly diagnosed AML patients who received either midostaurin or quizartinib in the frontline were included. Patients were excluded if they did not have a FLT3-ITD mutation, did not receive a 7+3 backbone, or if treatment doses were unknown due to induction initiation at a separate institution. Patients who received these TKIs as part of a clinical trial were also excluded. Descriptive analyses were conducted for baseline demographic and disease characteristics, treatment dosing and durations, safety events, and response rates. Composite complete remission (CRc) included complete remission (CR) and CR with incomplete count recovery (CRh/CRi). As a secondary analysis, we compared patients who received 90 mg/m2 vs 60 mg/m2 of daunorubicin (DNR). Median overall survival (OS) was assessed with Kaplan-Meier analysis, and differences between groups with log-rank test. Results: 110 patients received 7+3+midostaurin and 20 received 7+3+quizartinib. The median age was 58 years (interquartile range [IQR]: 47-65), 50% were male, 89% were white, and 6% had secondary AML. Most patients (80%) had AML with normal cytogenetics, and the most common co-occurring mutations with FLT3 were NPM1 (62%) and DNMT3A (45%). Baseline demographics and AML characteristics were similar between the 2 groups. Most patients (95%) were treated with on-label TKI dosing. Compared to those who received quizartinib, patients who received midostaurin were more likely to have dosing changes, interruptions, or early discontinuations (26% vs 5%, p=0.04). Rates of CRc were significantly higher with quizartinib (85% vs 73%, p=0.004). No patients in the quizartinib cohort required re-induction compared to 17.3% in the midostaurin group (p=0.04). Between the midostaurin and quizartinib cohorts, rates of febrile neutropenia (93% vs 85%), documented infections (42% vs 55%), parenteral nutrition (21% vs 35%), intensive care unit transfer (14% vs 10%), and granulocyte colony-stimulating factor use (12% vs 15%) were comparable (p>0.05 for all). The median follow-up (FU) was 48 months for midostaurin vs 4.1 months for quizartinib with 44 deaths in the former and 2 in the latter cohort. Of the patients who died, 22 had relapsed AML. Median OS was not reached in both. In the midostaurin cohort, 32% of patients relapsed, with median time to relapse of 8.3 months. No patients have relapsed with quizartinib during limited FU. Early (30-day) mortality was 2% for midostaurin and 5% for quizartinib (p=0.4). Our secondary analysis of DNR 90 mg/m2 (n= 33) vs 60 mg/m2 (n=63) found no significant differences in rates of CRc, relapse, transplant, and safety events. The average FU time was 43 months (39 months in DNR 90 mg/m2 vs 43 months in DNR 60 mg/m2); median OS was 27.1 months for DNR 90 mg/m2 and not reached for DNR 60 mg/m2 (Hazard ratio [HR]: 0.99; 95% confidence interval=0.51-1.92; p=0.77). Conclusion: We provide the first real-world evidence on comparative effectiveness of midostaurin and quizartinib in newly diagnosed AML patients. Patients treated with 7+3+quizartinib were more likely to achieve CRc, less likely to experience TKI dosing interruptions, and less likely to require re-induction. Limitations in our analyses include its retrospective nature, and lack of systematic MRD collection to potentially overcome short FU. Our analyses also suggest similar outcomes between DNR 90 mg/m2 vs 60 mg/m2. We will continue to evaluate safety, remission duration, event-free survival, and OS with longer FU.