Validation and refinement of the revised 2017 European LeukemiaNet genetic risk stratification of acute myeloid leukemia

The revised 2017 European LeukemiaNet (ELN) recommendations for genetic risk stratification of acute myeloid leukemia have been widely adopted, but have not yet been validated in large cohorts of AML patients. We studied 1116 newly diagnosed AML patients (age range, 18–86 years) who had received induction chemotherapy. Among 771 patients not selected by genetics, the ELN-2017 classification re-assigned 26.5% of patients into a more favorable or, more commonly, a more adverse-risk group compared with the ELN-2010 recommendations. Forty percent of the cohort, and 51% of patients ≥60 years, were classified as adverse-risk by ELN-2017. In 599 patients <60 years, estimated 5-year overall survival (OS) was 64% for ELN-2017 favorable, 42% for intermediate-risk and 20% for adverse-risk patients. Among 517 patients aged ≥60 years, corresponding 5-year OS rates were 37, 16, and 6%. Patients with biallelic CEBPA mutations or inv(16) had particularly favorable outcomes, while patients with mutated TP53 and a complex karyotype had especially poor prognosis. DNMT3A mutations associated with inferior OS within each ELN-2017 risk group. Our results validate the prognostic significance of the revised ELN-2017 risk classification in AML patients receiving induction chemotherapy across a broad age range. Further refinement of the ELN-2017 risk classification is possible.


AMLCG-2008 trial
In the AMLCG-2008 trial 3 (NCT01382147; recruitment period, 2009 to 2012), patients <60 years and fit patients up to the age of 70 ('younger' patients) were randomized to receive either double induction chemotherapy with TAD-9 and HAM (21 days apart), or dose-dense induction therapy according to the sHAM regimen (cytarabine 3 g/m² [1 g/m² in patients ≥60 years] twice daily on days 1,2,8 and 9; and mitoxantrone 10 mg/m² on days 3,4,10 and 11). AlloSCT from an HLA-matched related or unrelated donor was the recommended postremission therapy for all younger patients achieving CR except those with favorable genetic features (defined as favorable cytogenetics or cytogenetically normal patients with mutated NPM1 and no FLT3-ITD) and good response to induction chemotherapy (<10% blasts in a bone marrow aspirate obtained on d16 after start of induction therapy). For younger patients without a donor, those unable or unwilling to undergo allotransplantation, and those with a favorable risk profile, postremission therapy consisted of one cycle of TAD-9 for consolidation, followed by 3 years of cytarabine-based maintenance therapy.
Less fit patients aged ≥60 years, and all patients aged ≥70 years, were randomized to receive induction therapy according to the HAM regimen (cytarabine, 1g/m² per dose) followed by a second HAM induction cycle on day 21 only if a bone marrow aspirate on day 16 showed ≥5% blasts, or to dose-dense induction with sHAM (cytarabine, 1g/m² per dose). Postremission therapy in this group consisted of one cycle of TAD-9 for consolidation, followed by 3 years of maintenance therapy.

Definition of clinical end points
Clinical endpoints were defined, in accordance with generally accepted criteria, 4,5 as follows: Complete remission (CR) required a bone marrow (BM) aspirate with cellularity greater than 20% and maturation of all cell lines, less than 5% blasts and no Auer rods; and in the peripheral blood, an absolute neutrophil count of ≥1,500/µL, platelet count of ≥100,000/µL, and no leukemic blasts; and no evidence of extramedullary leukemia. Relapse was defined by the

Ambiguities in ELN-2017 risk group assignment
We identified a small number of patients (approximately 1%) in which the ELN-2017 recommendations for genetic risk stratification did not allow an unambiguous risk group assignment. Specifically, four patients had mutated NPM1 in the context of complex karyotypes, and were classified as adverse risk in agreement with recently published data. 6 Of these four patients, three achieved a CR. Two had alloSCT in first remission and were alive and relapse-free 3.4 and 8.2 years from diagnosis, the third patient received an alloSCT for early relapse and died 9 months after initial presentation. Subsequent therapy for the patient not achieving CR is unknown, but he was alive 2 years after initial diagnosis. One patient had a FLT3-ITD with a high allelic ratio and biallelic CEBPA mutations, and another patient had a FLT3-ITD with a high allelic ratio and an inversion inv(16). Both were assigned to the favorablerisk category, and were alive in CR1 approximately 14.5 years and 8.5 years after AML diagnosis, respectively. Four patients had mutated TP53 in the presence of a balanced chromosomal translocation [inv(16), n=1 and t(9;11), n=3] and were assigned to the favorableand intermediate-risk categories, respectively. The patient with mutated TP53 and inv (16) received an allogeneic transplant in CR1 and was alive in remission 4 years after his initial diagnosis. Among the three patients with mutated TP53 and t(9;11), one received alloSCT in CR1 but relapsed and died 2.5 years from AML diagnosis, while two did not achieve CR and died within the first two months from diagnosis.

Validation of the proposed refinement of the ELN-2017 classification in an independent patient cohort
To validate our proposed refinement of the ELN-2017 risk classification, we analyzed published genetic and survival data of patients treated on three consecutive clinical trials of the German AML-SG study group (n=1540). 7 In brief, the AML-HD98A trial included younger patients (18-60 years) who received induction chemotherapy with idarubicin, cytarabine and etoposide (ICE). Allogenic transplantation was offered to patients with adverse cytogenetic risk, intermediate-risk patients received allogeneic transplantation or intensive postremission chemotherapy, and low-risk patients received chemotherapy only. The AML-SG 07-04 trial included younger patients who were randomized to induction with either ICE or ICE plus alltrans retinoic acid (ATRA). In the AML-HD98B trial, patients aged ≥60 years were randomized to receive induction with either ICE or ICE plus ATRA.
For this cohort, information on FLT3-ITD-to-wild type allelic ratio is not publicly available.
Patients that could not be reliably classified according to the ELN-2017 criteria, mostly due to missing data on FLT3 allelic ratio, as well as patients with acute promyelocytic leukemia, were excluded, leaving 1192 patients for the validation cohort (median age, 51 years; range, 18-84 years; 83% aged <60 years). Of note, the lack of data on FLT3-ITD allelic ratio does not interfere with classifying patients into the proposed "very favorable" and "very adverse" subgroups, allowing us to use this cohort for the purpose of validating our refinement of the ELN-2017 classification.

Outcomes of patients within genetic subsets of the ELN-2017 categories
We examined outcomes of specific genetic subsets with the ELN-2017 risk categories (Supplemental Table 4). Off note, since several of the resulting subgroups are relatively small and since these analyses were not adjusted for multiple testing, these results should be considered exploratory until further validation.
Within the ELN-2017 favorable risk group (Supplemental Figure 9A

Potential further refinement of the ELN-2017 risk categories through inclusion of additional gene mutations
When the gene mutations listed in Table 1 were individually included in the multivariate analyses shown in Figure 4, only DNMT3A mutations significantly associated with inferior RFS as well as OS, with an approximately 1.3-fold risk increase for either outcome (Supplemental Figure 19). Within each ELN-2017 risk category, mutated DNMT3A identified a subgroup with significantly inferior OS compared to DNMT3A wild-type patients (Supplemental Figure 20). # Two of these patients also had a FLT3-ITD with high allelic ratio, and one also had mutated TP53. † Six of these patients also had mutated ASXL1 and/or RUNX1, and one also had mutated TP53 ‡ Four of these patients also had mutated RUNX1