OS in CEBPA⁺ versus CEBPA-wt patients according to the presence of a normal karyotype (CN-AML) or FLT3-ITD. In the presence of the CEBPA mutation, estimated 5-year OS was 64% (95% CI, 36%-82%) in the 26 patients with CN-AML and no FLT3-ITD (solid black curve), 30% (95% CI, 7%-58%) in the 10 patients with CN-AML and FLT3-ITD (solid orange curve), and 32% (95% CI, 11%-56%) in the 15 patients with non-CBF abnormal karyotype (solid blue curve), with a hazard ratio of 0.30 (95% CI, 0.12-0.72) in the former compared with the 2 latter patient subsets (P = .004). In these 3 patient subsets, estimated 5-year DFS was 62% (95% CI, 38%-79%), 43% (95% CI, 10%-73%), and 27% (95% CI, 7%-54%), respectively (P = .028 for the former vs 2 latter subsets comparison), whereas estimated 5-year RFS was 62% (95% CI, 38%-79%), 43% (95% CI, 10%-73%), and 34% (95% CI, 8%-63%), respectively (P = .07 for the former vs 2 latter subsets). In the context of de novo CN-AML without FLT3-ITD, estimated 5-year OS in the 221 patients without CEBPA mutation (dashed black curve) was 37% (95% CI, 30%-45%), which significantly differed from the solid black curve (P = .035). In the context of de novo CN-AML with FLT3-ITD, estimated 5-year OS in the 59 patients without CEBPA mutation (dashed orange curve) was 28% (95% CI, 17%-41%), not significantly different from the solid orange curve (P = .85). In the context of de novo AML with non-CBF abnormal karyotype, estimated 5-year OS in the 211 patients without CEBPA mutation (dashed blue curve) was 26% (95% CI, 20%-33%), not significantly different from the solid blue curve (P = .80). Patients with favorable CBF-AML as well as those with secondary AML were not included in these estimations.