Prognosis of acute myeloid leukemia (AML) is mainly determined by presenting cyto- and molecular genetics, as systematized in the recent European LeukemiaNet (ELN) risk classification. Moreover, mutational profiling is currently used to assign post-remission therapy, especially in the category of intermediate-risk cytogenetics (IR-AML), given the observed benefit of allogeneic hematopoietic stem-cell transplantation (alloHSCT) in first complete response in patients harboring high-risk molecular features such as FLT3-ITD and the lack of benefit in patients with a favorable genotype (i.e., mutated NPM1 and CEBPA biallelic mutation). Nonetheless, the prognosis for patients with an IR-AML lacking mutations of these three genes (“triple negative” AML) is uncertain, and optimal post-remission strategy for these bona fide intermediate-risk patients is mostly unknown.

In order to elucidate if alloHSCT could improve the prognosis of this molecular subgroup, we analyzed the outcome in a large series of patients with “de novo”, IR-AML (70% harboring a normal karyotype), diagnosed between 2000 and 2013, lacking mutations of NPM1, FLT3-ITD, and CEBPA (only patients with biallelic mutations were excluded), who had achieved a first complete remission after 1 or 2 courses of standard induction therapy. These 630 patients (median age, 52 years, range, 18-72; 58% male) have been included in the database of three different cohorts, the cooperative AML groups AMLSG (n=239) and Spanish CETLAM (n=146), and the HSCT registry of the ALWP-EBMT (n=245). The prognostic impact of alloHSCT in CR1 was analyzed as a time-dependent variable in univariable (Mantel-Byar method, Simon-Makuch plots) and multivariable (Cox with time dependent-variable) analyses.

After a median follow-up of 37 months, 3-year survival (OS), leukemia free-survival (LFS) and relapse incidence (RI) in the pooled two subgroups of AMLSG and CETLAM cooperative groups, regardless of post-remission therapy, was 53% [95% CI: 47-58], 36% [95% CI: 31-41], and 51% [95% CI=45-56], respectively (57±3%, 37.5±3%, and 44±4% in patients up to 60 year-old, respectively), without significant differences between both cohorts. In order to investigate the potential benefit of alloHSCT, we analyzed the entire group, including also patients from the ALWP-EBMT cohort. Remarkably, the outcome of patients from national cooperative groups AMLSG-CETLAM compared to registry ALWP-EBMT who received an alloHSCT in CR1 was not statistically different (3-year OS, LFS, and RI: 57 vs. 68%, p=0.13; 61 vs. 73%, p=0.08; and 22 vs. 19%, p=0.67). Patients who received an alloHSCT in CR1 (n=396) had an improved outcome compared to other post-remission options (high-dose cytarabine based chemotherapy, n=189; autoHSCT, n=45) with a higher OS (3-yr: 65 vs. 49%, p=7x10-5) and LFS (3-yr: 60 vs. 26%, p<10-13), and lower RI (3-yr: 23 vs. 72%, p<10-16). Moreover, post-remission option (alloHSCT in CR1 vs. other), as time-dependent variable, was the strongest prognostic factor in the multivariate analysis in terms of OS (HR=0.61, 0.46-0.8; p=0.0003), LFS (HR=0.49, 0.37-0.65; p<0.0001), and RI (HR=0.24, 0.17-0.34, p<0.001). Other independent prognostic factors identified were age for OS (HR=1.24, 1.11-1.40 -by 10 years-; p=0.0004) and LFS (HR: 1.21, 1.08-1.34; p=0.0007), presenting WBC at diagnosis for OS (HR=1.29, 1-1.67 –over median-; p=0.05), and female gender for OS (HR=0.74, 0.57-0.97; p=0.03). Interestingly, the outcome after alloHSCT did not differ depending on donor type (2-yr RI, LFS, and OS: 14 vs. 19%, p=0.97; 67 vs. 68%, p=0.42; 73 vs. 72%, p=0.95 after alloHSCT with matched sibling, n=200, or an unrelated donor, n=183, respectively).

In conclusion, IR-AML patients with a triple negative genotype constitute a subgroup with a high risk of relapse after postremission therapy with high-dose cytarabine based chemotherapy or autoHSCT. This large study, involving patients from two cooperative groups and the EBMT registry, indicates that an alloHSCT in first CR is associated with a marked relapse reduction and survival benefit in the two cooperative group cohorts with prospective treatment data as well as in the whole cohort including the EBMT registry data.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.