Kinetics of Complete Remission (CR) and CR Duration and Its Impact on Overall Survival (OS) and Event-Free Survival (EFS) in Quantum-First

Abstract

Background Results from the phase 3 QuANTUM-First study (NCT02668653) showed that in patients (pts) with newly diagnosed (nd) acute myeloid leukemia (AML) and positive for FMS-like tyrosine kinase 3-internal tandem duplication ( FLT3-ITD+), the addition of the oral, highly potent, and selective, type 2 FLT3 inhibitor quizartinib (Quiz) to standard chemotherapy ± allogeneic hematopoietic cell transplantation (allo-HCT), followed by continuation monotherapy for up to 3 years, decreased the relative risk of death by 22.4% vs placebo (PBO) (PMID: 37116523). We assessed 1) the impact of CR duration on OS and EFS and 2) the kinetics of CR achievement over time after induction. Methods QuANTUM-First enrolled pts aged 18-75 years with FLT3-ITD+ nd AML, randomized to receive standard induction chemotherapy with either Quiz (40 mg/d) or PBO, and stratified by region (Europe, North America, or Asia/Australia/South America), age (<60 years, ≥60 years), and white blood cell count (WBC; <40×10 9/L, ≥40×10 9/L) at diagnosis. Pts who achieved CR or CR with incomplete hematologic recovery (CRi) received up to 4 cycles of high-dose cytarabine combined with Quiz (40 mg/d) or PBO and/or allo-HCT followed by up to 3 years (36 cycles) of continuation therapy with Quiz (30-60 mg/d) or PBO. OS was the primary endpoint, EFS and rates of CR and composite complete remission (CRc) were secondary endpoints; CR duration was an exploratory endpoint. The impact of CR duration on OS was analyzed by extended Cox regression model stratified by region, age, and WBC count at diagnosis and included treatment group and CR duration status (time-dependent covariate). CR duration status was 0 for pts who did not achieve CR, 1 for CR pts throughout the duration of sustained CR, and 0 for CR pts after they relapsed. Results Of the 539 randomized pts enrolled in QuANTUM-First (Quiz, n=268; PBO, n=271), there were 297 CR pts (Quiz, n=147; PBO, n=150) based on independent review committee assessment after 1 to 2 courses of induction. Among 533 treated pts, 20% had 2 induction cycles. Baseline characteristics in CR pts were well balanced between arms: median age in years (range), 56.0 (23-75) vs 55.5 (20-74); females, 55.8% vs 60.0%; ECOG ≥1, 64.6% vs 63.3%; mutated NPM1 / CEBPA,66.7% vs 64.7% / 25.9% vs 23.3%; variant allele frequency (VAF) >25%, 68.7% vs 66.7%; and WBC count at diagnosis ≥40×10 9/L, 55.8% vs 49.3%. Median OS in CR pts was not reached (95% CI 48.6 months-NE) in the Quiz arm vs 35.4 (95% CI 15.1-NE) months in the PBO arm (HR 0.572, 95% CI 0.400-0.816; Figure 1). Although CR rates were similar between arms (55% in both), median CR duration was longer with Quiz vs PBO (38.6 months, 95% CI 21.9-NE vs 12.4 months, 95% CI 8.8-22.7; HR 0.621, 95% CI 0.451-0.857). Based on the extended Cox model, CR duration status covariate was strongly predictive for OS (HR 0.156, 95% CI 0.113-0.216; nominal P<0.0001; Figure 2). Importantly, the HR for Quiz vs PBO effect on OS in the model with the time-dependent covariate CR duration status was higher (HR 0.851, 95% CI 0.672-1.076) vs the HR of the primary OS analysis (HR 0.776, 95% CI 0.615-0.979), implying that a substantial fraction of OS benefit was mediated through its effects on achieving a durable CR (Figure 2). EFS favored Quiz over PBO when induction treatment failure (ITF) was defined as not achieving CRc or CR, by the end of induction up to day 56 from the last induction cycle, with HR 0.729 (95% CI 0.592-0.897; nominal P=0.0031) and 0.818 (95% CI 0.669-0.999; nominal P=0.032), respectively. EFS was similar between arms when ITF was defined, according to FDA guidance, as not achieving CR by day 42 from the start of the last induction cycle (HR 0.916, 95% CI 0.754-1.114; P=0.24; primary EFS analysis). Between day 42 and the end of the induction phase, there were 51 pts who achieved CR (Quiz, n=33; PBO, n=18), and they were considered as ITF with EFS event on day 1 in the primary analysis of EFS. Among these 51 pts, 9 (Quiz, n=5; PBO, n=4) had CRi by day 42 and became CR after day 42 in the induction phase. Conclusions These analyses demonstrated that in QuANTUM-First: 1) a substantial fraction of the estimated effect of Quiz on OS was mediated through its effect on achieving a durable CR and 2) more pts in the Quiz arm vs the PBO arm achieved CR after day 42 of the last induction cycle, suggesting that there was a delay in CR achievement with Quiz. Longer EFS was observed in Quiz over PBO based on the EFS definition including CRs between day 42 and end of induction.