Maintenance therapy with the FMS-like tyrosine kinase 3 inhibitor gilteritinib in patients with FMS-like tyrosine kinase 3-internal tandem duplication acute myeloid leukemia: A phase 2 study

Abstract

Background: The GOSSAMER phase 2 study assessed the FMS-like tyrosine kinase 3 (FLT3) inhibitor gilteritinib as maintenance therapy in patients with FLT3-internal tandem duplication (FLT3-ITD) acute myeloid leukemia (AML) in first complete remission without previous hematopoietic stem cell transplantation (HSCT).

Methods: Patients had to be within 2 months of their last consolidation cycle and have completed the recommended number of cycles per local practice. FLT3 inhibitors were allowed only during induction and/or consolidation. The primary end point was relapse-free survival (RFS). Secondary end points included overall survival (OS), event-free survival, and measurable residual disease (MRD).

Results: In total, 98 patients were randomized (gilteritinib, n = 63; placebo, n = 35). RFS was not significantly different between the arms (hazard ratio, 0.74; 95% confidence interval, 0.41-1.34; p = .16). RFS rates for the gilteritinib and placebo arms were 68.5% and 55.3% at 1 year, 51.8% and 44.9% at 2 years, and 41.2% and 40.8% at 3 years, respectively. OS was not significantly different between the arms but may have been affected by subsequent AML therapies after discontinuation. In patients who received subsequent therapy (gilteritinib, 46.8%; placebo, 60.0%), a higher percentage of placebo-treated (57.1%) versus gilteritinib-treated patients (27.6%) underwent HSCT. At the end of treatment, 96.4% of gilteritinib-treated and 85.7% of placebo-treated patients had undetectable MRD. Relapsed placebo-treated (86.7%) versus gilteritinib-treated patients (34.8%) had a greater FLT3 mutational burden. No new significant safety concerns were noted.

Conclusions: The primary end point was not achieved; however, an observed trend toward potential benefit was noted in patients with FLT3-ITD AML who had not undergone prior HSCT.

Keywords: FLT3 inhibitor; FMS‐like tyrosine kinase 3–internal tandem duplication acute myeloid leukemia (FLT3‐ITD AML); gilteritinib; maintenance therapy; measurable residual disease.

FIGURE 1

Patient disposition. *The patient’s MRD status was not available in time for randomization; therefore, the patient was rescreened and given a new patient number. ^†In total across both treatment arms, the end‐of‐treatment category of “other” included physician decision (n = 3); the patient became eligible for transplant and proceeded to transplant (n = 3); the patient was not eligible for the study (n = 1); MRD was detected (n = 1); and the patient was randomized in error (n = 1). ^‡After treatment discontinuation, patients had a 30‐day follow‐up visit for safety as per the study protocol. ^§After the 30‐day safety follow‐up visit, patients entered the long‐term follow‐up period for collection of subsequent acute myeloid leukemia treatment, remission status, and survival information as per the study protocol. Follow‐up continued until the final database lock. MRD indicates measurable residual disease.

FIGURE 2

Kaplan–Meier estimate of RFS. The primary analysis of RFS was performed at a one‐sided 0.075 significance level to test the null hypothesis that RFS in the gilteritinib arm is worse than or equal to RFS in the placebo arm, versus the alternative hypothesis that RFS in the gilteritinib arm is better. The p value was calculated from a stratified one‐sided log‐rank test. Stratification factors were age, geographic region, the presence of measurable residual disease at screening, and use of FLT3‐inhibiting agents per interactive response technology. HRs were based on a Cox proportional hazards model. Assuming proportional hazards, an HR of <1 indicated a reduction in hazard in favor of the gilteritinib arm. CI indicates confidence interval; FLT3, FMS‐like tyrosine kinase 3; HR, hazard ratio; NE, not estimable; RFS, relapse‐free survival.

FIGURE 3

Kaplan–Meier estimate of OS. OS was analyzed in the full analysis set with a stratified log‐rank test. With the sequential multiple test procedure to control for overall type I error at the one‐sided 0.075 significance level, formal significance testing of OS was conducted only if the RFS comparison was statistically significant. Otherwise, OS analysis was considered exploratory. Stratification factors were age, geographic region, the presence of measurable residual disease at screening, and use of FLT3‐inhibiting agents per interactive response technology. HRs were based on a Cox proportional hazards model. Assuming proportional hazards, an HR of <1 indicated a reduction in hazard in favor of the gilteritinib arm. CI indicates confidence interval; FLT3, FMS‐like tyrosine kinase 3; HR, hazard ratio; NE, not estimable; OS, overall survival; RFS, relapse‐free survival.

FIGURE 4

Kaplan–Meier estimate of EFS. EFS was analyzed with a stratified one‐sided log‐rank test with a one‐sided 0.075 significance level. Stratification factors were age, geographic region, the presence of measurable residual disease at screening, and use of FLT3‐inhibiting agents per interactive response technology. HRs were based on a Cox proportional hazards model. Assuming proportional hazards, an HR of <1 indicated a reduction in hazard in favor of the gilteritinib arm. CI indicates confidence interval; EFS, event‐free survival; FLT3, FMS‐like tyrosine kinase 3; HR, hazard ratio.

FIGURE 5

Median change from baseline in quantitative MRD as measured by a log₁₀‐transformed overall FLT3/ITD mutation ratio. Error bars represent min and max values. *MRD was measured as a log₁₀‐transformed overall FLT3/ITD mutation ratio (limit of detection, 10⁻⁶). Change from baseline by ratio was defined as postbaseline value/baseline value. Median (min, max) baseline values were −6.0 (−6.0, −1.3) in the gilteritinib arm and −5.8 (−6.0, −1.7) in the placebo arm. ^†Two‐sided p values from analysis of covariance including treatment, age group, geographic region, and use of FLT3‐inhibiting agents per interactive response technology as fixed factors and baseline score as the covariate. EOT indicates end of treatment; FLT3, FMS‐like tyrosine kinase 3; ITD, internal tandem duplication; max, maximum; min, minimum; MRD, measurable residual disease.

FIGURE 6

Kaplan–Meier estimate of RFS by MRD status at baseline. MRD was measured as a log₁₀‐transformed overall FLT3/ITD mutation ratio (limit of detection, 10⁻⁶). MRD was considered detectable if the log₁₀‐tranformed FLT3/ITD mutation ratio was >‐6; a mutation ratio of ≤‐6 meant MRD was undetectable. *In each subgroup, the HR was estimated with an unstratified Cox proportional hazards model. Assuming proportional hazards, an HR of <1 indicated a reduction in hazard in favor of the gilteritinib arm. ^†Based on the Wald test. ^‡For each subgroup, the interaction p value is from a Cox regression model, which included treatment, covariate, and treatment–covariate interaction terms. CI indicates confidence interval; FLT3, FMS‐like tyrosine kinase 3; HR, hazard ratio; ITD, internal tandem duplication; MRD, measurable residual disease; RFS, relapse‐free survival.