Added prognostic value of secondary AML-like gene mutations in ELN intermediate-risk older AML: ALFA-1200 study results

Abstract

In this study, we aimed to refine prognostication of older with acute myeloid leukemia (AML) after intensive chemotherapy. Five hundred and nine patients aged 60 years or older (median age, 68 years) were prospectively enrolled in the intensive Acute Leukemia French Association (ALFA)-1200 trial between 2012 and 2016, and 471 patient samples were submitted to multigene analysis. Mutations in any of 8 genes frequently altered in myelodysplastic syndromes (MDS), including ASXL1, SRSF2, STAG2, BCOR, U2AF1, EZH2, SF3B1, and ZRSR2, defined a secondary AML (sAML)-like disease, as reported. Of the samples analyzed, 48% included sAML-like gene mutations. These mutations were associated with a shorter event-free survival, both overall (hazard ratio, 1.46; 95% confidence interval, 1.19-1.79; P < .001) and within the European LeukemiaNet (ELN)-2017 intermediate-risk subgroup (hazard ratio, 1.52; 95% confidence interval, 1.01-2.28; P = .044), which excludes ASXL1-mutated cases by definition. We therefore included patients with intermediate-risk AML carrying sAML-like mutations in a single high-risk patients group together with adverse-risk patients with AML, whereas other intermediate-risk patients were included in a standard-risk group together with favorable-risk patients (high-risk/standard-risk patient ratio, 1.00). Using this 2-class risk assessment, we observed that transplantation prolonged overall survival from remission in patients with high-risk AML only, not in patients with standard-risk AML. Routine analysis of sAML-like gene mutations may thus improve the definition of high-risk older patients with AML, and better identify the half of older patients who clearly derive survival benefit from allogeneic transplantation in first remission. This trial was registered at www.clinicaltrials.gov as #NCT01966497.

Conflict of interest statement

Conflict-of-interest disclosure: The authors declare no competing financial interests.

© 2020 by The American Society of Hematology.

Figures

Graphical abstract

Figure 1.

Incidences of sAML-like gene mutations. (A) sAML-like gene mutation in the whole population of 471 patients analyzed. (B) sAML-like gene mutation in the subset of 129 patients with ELN-2017 intermediate-risk AML who were analyzed. Of note, none of these patients had AML with ASXL1 and/or RUNX1 gene mutation, which are 2 criteria for adverse-risk AML in this classification. In these patients, the incidence of sAML-like mutations was not higher in patients with chromosomal abnormalities compared with those with a normal karyotype (41% vs 36%; P = .59). sAML is clinically defined as patients with prior myelodysplastic syndromes or chronic myelomonocytic leukemia (N = 74 patients). (C) Incidences of NPM1, FLT3, IDH1/2, and DNMT3A gene mutations in ELN-2017 intermediate-risk AML analyzed, according to the presence (n = 49) or absence (n = 80) of sAML-like gene mutations. As shown, FLT3-ITD mutations were more frequently observed in patients without sAML-like mutations, whereas IDH2 mutations were more frequently observed in those with sAML-like mutations. Of note, among these IDH2 gene mutations, the incidence of the poor-prognosis IDH2 R172 mutation was 6/49 vs 9/80 (P = .54).

Figure 2.

Prognostic effect of sAML-like gene mutations. (A) EFS according to the 3 ELN-2017 risk subgroups (P < .001). At 2 years, EFS was estimated at 42.6% (95% CI, 34.3%-50.8%), 31.8% (95% CI, 24.7%-39.2%), and 15.2% (95% CI, 10.6%-20.6%) in the favorable-, intermediate-, and adverse-risk subgroups, respectively. At 4 years, EFS was estimated at 34.8% (95% CI, 26.7%-43.1%), 21.8% (95% CI, 15.2%-29.2%), and 10.3% (95% CI, 6.0%-16.1%) in the favorable-, intermediate-, and adverse-risk, respectively. (B) EFS according to sAML-like mutations in the ELN-2017 intermediate-risk subgroup. The presence of sAML-like gene mutations significantly influenced EFS in the ELN-2017 intermediate-risk subgroup (HR, 1.52; 95% CI, 1.01-2.28; P = .044). (C) EFS according to the newly defined high-risk/standard-risk groups. EFS was significantly reduced in the high-risk group (HR, 2.03; 95% CI, 1.65-2.48; P < .001). At 2 years, EFS was estimated at 39.8% (95% CI, 33.7-45.9) in the standard-risk group compared with 16.4% (95% CI, 12.0%-21.3%) in the high-risk group. At 4 years, EFS was estimated at 29.6% (95% CI, 23.6%-35.8%) in the standard-risk group compared with 12.3% (95% CI, 8.2%-17.2%) in the high-risk group. (D) OS according to the 3 ELN-2017 risk subgroups (P < .001). At 2 years, OS was estimated at 60.3% (95% CI, 51.6%-68.0%), 50.1% (95% CI, 42.0%-57.6%), and 32.8% (95% CI, 26.4%-39.4%) in the favorable-, intermediate-, and adverse-risk subgroups, respectively. At 4 years, OS was estimated at 49.8% (95% CI, 41.0%-58.0%), 32.7% (95% CI, 24.9%-40.7%), and 16.2% (95% CI, 10.7%-22.6%) in the favorable-, intermediate-, and adverse-risk subgroups, respectively. (E) OS according to sAML-like mutations in the ELN-2017 intermediate-risk subgroup. The presence of sAML-like gene mutations did not significantly influence OS in the ELN-2017 intermediate-risk subgroup (HR, 1.31; 95% CI, 0.85-2.03; P = .22). (F) OS according to the newly defined high-risk/standard-risk groups. OS was significantly reduced in the high-risk group (HR, 1.86; 95% CI, 1.49-2.31; P < .001). At 2 years, OS was estimated at 57.6% (95% CI, 51.2%-63.5%) in the standard-risk group compared with 34.2% (95% CI, 28.3%-40.2%) in the high-risk group. At 4 years, OS was estimated at 42.9% (95% CI, 36.3%-49.3%) in the standard-risk group compared with 19.0% (95% CI, 13.8%-24.7%) in the high-risk group.