Characterization of secondary and therapy-related acute myeloid leukemia

Abstract

Secondary acute myeloid leukemia (s-AML) refers to patients with either therapy-related AML (t-AML), that is, AML after treatment with chemo- and/or radiation for a prior disease, or AML progressing from an antecedent hematologic disorder (AHD-AML), typically a myelodysplastic syndrome (MDS) or a myeloproliferative neoplasm. Patients with s-AML present with higher rates of adverse cytogenetic aberrations and higher frequencies of adverse mutations and, subsequently, respond worse to therapy, and have poorer outcome compared to de novo AML. To identify clinical and molecular factors that may improve prognostication is therefore of importance to guide clinical decision-making.

The general aim of this thesis was to broaden the real-world knowledge of s-AML, regarding disease properties and outcome, using population-based registries, and additionally to investigate the importance of mutations in the transformation from MDS to AML. The research papers presented herein cover mutational screening in MDS and s-AML (study I), general characteristics and outcome of s-AML (study II), the role of allogeneic hematopoietic cell transplantation (HCT) in patients with s-AML (study III), and the epidemiology and treatment outcome specifically for t-AML (study IV).In study I, high-throughput methods were used to find mutations in 22 genes in 100 MDS and 92 AML patients. The AML cohort consisted of t-AML, AHD-AML and AML with MDS-like cytogenetics. In AML, mutations were most commonly seen in oncogenes and cell signaling genes, and in MDS in splicing factor genes and epigenetic regulators. A key finding was the overrepresentation of mutated U2AF1 in cases with MDS progressing to AML. Furthermore, in addition to established risk scores, mutational status improved prognostication.

In study I, high-throughput methods were used to find mutations in 22 genes in 100 MDS and 92 AML patients. The AML cohort consisted of t-AML, AHD-AML and AML with MDS-like cytogenetics. In AML, mutations were most commonly seen in oncogenes and cell signaling genes, and in MDS in splicing factor genes and epigenetic regulators. A key finding was the overrepresentation of mutated U2AF1 in cases with MDS progressing to AML. Furthermore, in addition to established risk scores, mutational status improved prognostication. In study II we used the Swedish AML registry (SAMLR) to characterize s-AML in a population-based setting. Of the 3263 AML patients included, 19% were AHD-AML and 8% t-AML. Differences between the subtypes were found in age, gender distribution and cytogenetic risk. Compared to de novo AML, complete remission rates were lower in s-AML, but early death rates were similar. In multivariable analysis, both t-AML and AHD-AML emerged as independent prognostic factors, with a more pronounced negative impact in younger age groups. HCT is a potentially curable consolidation treatment in eligible patients. In study III, data on 3337 intensively treated patients in SAMLR were combined with data from the Swedish Cancer Registry (SCR) and the Swedish transplantations centers to investigate the role of HCT in s-AML. HCT in first remission was superior to consolidation treatment with chemotherapy only. Long-term survivors with s-AML were rare without HCT. In study IV we studied 686 patients with t-AML in detail using SAMLR, SCR and the Swedish Rheumatology Quality Register. We found an increasing incidence of t-AML over time, and an increasing proportion with t-AML of AML in total. Survival was overall dismal, but comparable to de novo AML in patients with favorable cytogenetic risk and in patients with mutated NPM1 in combination with absence of FLT3-ITD.

In conclusion, secondary AML is a highly heterogeneous disease with a particularly poor outcome. However, the clinical and genetic differences within the disease enable risk stratification of patients and may thus aid in treatment recommendations.