Cut-point selection using maximally log-rank statistics. (A) Optimal cut-point for the AR by maximally selected log-rank statistics in intensively treated FLT3-ITD positive AMLs. Patients who proceeded to allogeneic HSCT were excluded. Maximally selected log-rank statistics performed for the continuum of the AR to test for a potential cut-point separating 2 groups with different survival distributions. The AR is shown on the x-axis and the corresponding standardized log-rank statistic on the y-axis. The estimated cutoff point was 0.51, with an M statistic of 3.15 and a corresponding corrected P value of .038. The vertical dashed line represents the optimal cut-point for AR evident on maximally selected log-rank statistics and the corresponding M statistics. (B) Impact of AR on OS in first CR according to preselected optimal cut-point in intensively treated FLT3-ITD–positive patients excluding allogeneic HSCT. (C) Impact of AR on relapse-free survival according to preselected optimal cut-point in intensively treated FLT3-ITD–positive patients excluding allogeneic HSCT.