Treatment-Related AML Is An Independent Adverse Prognostic Factor for Relapse-Free and Overall Survival. An Analysis of 2,868 Adult Patients with Newly Diagnosed AML Enrolled On Seven AMLSG Treatment Trials.

Therapy-related AML (t-AML) is a recognized clinical syndrome occurring as a complication following cytotoxic and/or radiation therapy. The etiology and specific factors that predispose to t-AML largely remain unknown. Survival of t-AML patients has been poor compared with that of patients with de novo AML. However, there is a paucity of studies evaluating the impact of t-AML as a risk factor, in particular in the context of other clinical and biological prognostic markers.

To study the clinical impact of t-AML in a large cohort of patients with newly diagnosed AML, in the context of clinical characteristics as well as cytogenetics and mutational status of the NPM1 and FLT3 genes.

The study included 3,139 adult patients (median age, 53.5 years; range, 16-85 years) with newly diagnosed AML entered on 7 protocols of the German-Austrian AML Study Group (AMLSG) between 1993 and 2008. In all protocols intensive induction and consolidation therapy was used. Information on type of AML, karyotype and molecular marker status of NPM1 and FLT3 (internal tandem duplication [ITD] and tyrosine kinase domain mutation [TKD]) was availabel in 2,858 of 3,139 (91%) and 2,126 of 3,139 (68%) patients, respectively. Since this report focuses on the comparison of t-AML versus de novo AML, patients with secondary AML following myelodysplastic syndrome (s-AML) and those lacking information on type of AML were excluded (n=151, n=120, respectively).

Two hundred of the 2,868 patients (7%) were classified as t-AML. In more than two thirds of t-AML cases, a solid cancer was the primary malignancy with breast cancer being the most common (55%), followed by gastrointestinal (7.5%), prostate (7%) and testicular cancer (7%). In 27% of the cases, a hematologic neoplasm was the primary malignancy with non-Hodgkin (43%) and Hodgkin lymphoma (35%) being the most common. Three patients received cytotoxic treatment for autoimmune diseases. The median latency period between diagnosis of primary malignancy and occurrence of t-AML was 4 years (range, 4 months to 44 years). Patients with t-AML were significantly older compared to patients with de novo AML (58 versus 53 years; p<0.0001), and they had significantly lower median white blood counts (WBC) (7.4 vs 12.5 ×10⁹/L; p=0.002). The frequencies of cytogenetic risk groups in t-AML versus de novo AML were as follows: favorable [t(8;21), t(15;17), inv(16) or t(16;16)] (16% vs. 16%), intermediate [all cytogenetic abnormalities not classified as favorable or adverse] (46% vs. 66%), and adverse [t(v;11q23), inv(3) or t(3;3), t(9;22), -5 or 5q-, -7 or 7q-, abn(17p), complex karyotype] (38% vs. 18%, respectively). Response to induction therapy was not significantly different between t-AML (64%) and de novo AML (69%) in uni- (p=0.18) and multivariable (p=0.58) analysis. In contrast, for the clinical endpoints relapse-free (RFS) and overall survival (OS), t-AML was an adverse prognostic factor in univariable (p<0.00001, p<0.00001, respectively) and multivariable analyses (HR, 1.69, p=0.001; HR, 1.3, p=0.004, respectively). The negative prognostic impact of t-AML on RFS was due to both a higher cumulative incidence of relapse (p=0.01) and death in complete remission (CR) (p=0.0001). In cytogenetic subgroup analyses, t-AML was an unfavorable factor for OS in particular in patients with inv(16) or t(16;16) (p=0.008), whereas this was not the case in t(8;21) and t(15;17). In cytogentically normal AML, OS was significantly inferior in t-AML patients (p=0.009), and this negative impact was due to a significant inferior OS in the molecularly defined subgroups NPM1^mut/FLT3-ITD^neg (p=0.05) and the triple negative group (p=0.004), whereas there was no difference in the subgroup FLT3-ITD and CEBPA^mut. In as treated analyses in patients younger than 61 years with t-AML, allogeneic hematopoietic stem cell transplantation (HSCT) in first CR had a beneficial effect in cytogenetic intermediate- and adverse-risk (OS, p=0.008), but not favorable-risk patients.

In this large cohort of adult patients with newly diagnosed AML, t-AML was an independent adverse prognostic factor for RFS and OS. The negative impact on RFS was not only due to an increased relapse rate but also to a higher rate of deaths in CR, possibly reflecting cumulative toxicity of primary and secondary cancer therapy. Allogeneic HSCT appears to have a beneficial impact in younger adults with t-AML.

No relevant conflicts of interest to declare.