Purpose: Secondary and therapy-related acute myeloid leukemia (sAML and tAML, respectively) remain therapeutic challenges. Still, it is unclear whether their inferior outcome compared with de novo acute myeloid leukemia (AML) varies as a result of previous hematologic disease or can be explained by differences in karyotype and/or age.
Patients and methods: In a Danish national population-based study of 3,055 unselected patients with AML diagnosed from 2000 to 2013, we compared the frequencies and characteristics of tAML, myelodysplastic syndrome (MDS) -sAML, and non-MDS-sAML (chronic myelomonocytic leukemia and myeloproliferative neoplasia) versus de novo AML. Limited to intensive therapy patients, we compared chance of complete remission by logistic regression analysis and used a pseudo-value approach to compare relative risk (RR) of death at 90 days, 1 year, and 3 years, overall and stratified by age and karyotype. Results were given crude and adjusted with 95% CIs.
Results: Overall, frequencies of sAML and tAML were 19.8% and 6.6%, respectively. sAML, but not tAML, was associated with low likelihood of receiving intensive treatment. Among intensive therapy patients (n = 1,567), antecedent myeloid disorder or prior cytotoxic exposure was associated with decreased complete remission rates and inferior survival (3-year adjusted RR for MDS-sAML, non-MDS-sAML, and tAML: RR, 1.14; 95% CI, 1.02 to 1.32; RR, 1.27; 95% CI, 1.16 to 1.34; and RR, 1.16; 95% CI, 1.03 to 1.32, respectively) compared with de novo AML. Among patients ≥ 60 years old and patients with adverse karyotype, previous MDS or tAML did not impact overall outcomes, whereas non-MDS-sAML was associated with inferior survival across age and cytogenetic risk groups (adverse risk cytogenetics: 1-year adjusted RR, 1.47; 95% CI, 1.23 to 1.76; patients ≥ 60 years old: 1-year adjusted RR, 1.31; 95% CI, 1.06 to 1.61).
Conclusion: Our results support that de novo AML, sAML, and tAML are biologically and prognostically distinct subtypes of AML. Patients with non-MDS-sAML have dismal outcomes, independent of age and cytogenetics. Previous myeloid disorder, age, and cytogenetics are crucial determinants of outcomes and should be integrated in treatment recommendations for these patients.
© 2015 by American Society of Clinical Oncology.