Clinical Benefit of Maintenance Treatment in the RATIFY Trial
The CALGB 10603/RATIFY trial, a global, randomized, placebo-controlled phase 3 study, demonstrated that midostaurin, an oral, multitargeted, small-molecule FLT3 inhibitor, when added to induction and consolidation chemotherapy followed by maintenance, led to a statistically significant overall survival benefit compared with standard chemotherapy in newly diagnosed patients with activating FLT3 mutations. The objectives of the current unplanned post-hoc subset analysis of the RATIFY trial were to evaluate the contribution of midostaurin maintenance treatment to overall outcomes and disease-free survival (DFS) during maintenance treatment of the patient subset that achieved complete remission (CR) and started maintenance.
In the RATIFY study, 717 patients were randomized to receive either midostaurin (50 mg orally twice daily on days 8-22) or placebo in combination with induction therapy, which consisted of daunorubicin 60 mg/m2 intravenously (IV) on days 1 through 3 and cytarabine 200 mg/m2 on days 1 through 7, continuous IV infusion, plus midostaurin or placebo. Only patients achieving CR were allowed to receive consolidation therapy, consisting of 4 cycles of cytarabine 3 g/m2 every 12 hours on days 1, 3, and 5, plus midostaurin or placebo (50 mg orally twice daily on days 8-22) followed by twelve 28-day cycles of maintenance therapy with midostaurin or placebo (50 mg orally twice daily).
Landmark DFS analyses, defined as time from start of maintenance to the first of death or relapse, where patients alive and disease-free were censored at the time they completed the planned maintenance or discontinued study drug early, were performed. Landmark DFS analyses were also performed to understand the long-term impact of midostaurin versus placebo on the subset of patients who achieved CR and completed all protocol treatment (ie, time from end of all planned maintenance to the first of death or relapse).
At median follow-up of 59 months from enrollment, all patients were off active treatment. Overall, 403 (56%) patients achieved a CR by day 60 (CR60), with no significant differences between arms (midostaurin, 59%; placebo, 54%); an additional 101 patients achieved CR after 60 days (31 entered maintenance) but were excluded from this analysis. Of the 403 CR60 patients, 174 began maintenance therapy still in first remission (CR1) and remained on their originally assigned double-blind treatment arm, and were not re-randomized. There were no significant differences between the midostaurin and placebo maintenance arms in terms of pretreatment characteristics; however, compared with the patients who did not start maintenance therapy (n = 543), patients on the maintenance cohort were slightly older, significantly fewer were female, and they had more favorable FLT3 mutational status and cytogenetics.
Median duration of exposure was 336 days on both arms. There was 8% discontinuation due to adverse events in the midostaurin arm versus 6% in the placebo arm. During 12 cycles of maintenance therapy, 30 patients relapsed, and 1 death was reported in the midostaurin arm (30%) versus a 32% relapse rate in the placebo arm (22 patients). Landmark analysis did not show DFS differences between the 2 arms during the maintenance phase (hazard ratio [HR], 0.83; P = .49). At the end of the maintenance phase, 60 (57%) patients remained on the midostaurin arm versus 44 (64%) patients on the placebo arm. Postmaintenance, 16 DFS events (all relapses) were reported on the midostaurin arm versus 9 DFS events on the placebo arm (7 relapses, 2 deaths). Using a landmark analysis of DFS for the 104 patients who completed all planned maintenance from the last dose of study drug, there was no difference in DFS between the 2 arms (HR, 1.4; 95% confidence interval [CI], 0.63-3.3; P = .38). DFS at 1 year from the end of maintenance was 75% (95% CI, 62%-84%) for midostaurin and 91% (95% CI, 77%-96%) for placebo.
The authors concluded that the clinical benefit of maintenance therapy could not be determined based on this unplanned post-hoc subset analysis of the CALGB 10603/RATIFY trial, and that randomization was warranted to address this question.
Larson RA, et al. 2017 ASH. Abstract 145.