Prognostic Value of Integrated Cytogenetic and Mutational Risk Classification in Acute Myeloid Leukemia

Get Permission

Jay P. Patel, BS, and colleagues from Memorial Sloan-Kettering Cancer Center in New York recently performed mutational analysis of 18 genes in a subgroup of newly diagnosed acute myeloid leukemia (AML) patients who had been randomized to receive cytarabine plus high-dose or standard-dose daunorubicin induction therapy in the phase III ECOG E1900 trial. As reported in The New England Journal of Medicine, they identified a number of gene alterations and combinations of mutant and wild-type genes that were predictive of overall survival and incorporated these findings into a prognostic schema along with cytogenetic risk classification.1 Mutational analysis further identified alterations that were predictive of improved outcome with high-dose daunorubicin. These findings raise important issues regarding the use of more-extensive genetic profiling in AML patients prior to starting therapy.

Study Design

In this study, mutational analysis of 18 genes recently identified as having recurrent somatic mutations in AML was performed in 398 patients aged ≤ 60 years from the ECOG E1900 population (test cohort). Overall, at least one somatic alteration in these genes was identified in 97.3% of patients.

Univariate analysis showed that FLT3 internal tandem duplication (FLTD-ITD) mutations (P = .001), MLL partial tandem duplication (MLL-PTD) mutations (P = .009), PHF6 mutations (P = .006), and ASXL1 mutations (P = .05) were associated with reduced overall survival, whereas CEBPA mutations (P = .05) and the core-binding-factor alterations t(8;21) and inv(16)/t(16;16) (P < .001) were associated with improved overall survival. IDH2 mutations were associated with improved overall survival in the entire cohort of patients (3-year overall survival of 66%, P = .01), but the beneficial effect of these mutations was confined to patients with the IDH2 R140Q mutation. KIT mutations were associated with reduced overall survival among patients with the t(8;21) core-binding-factor alteration (P = .006) but not among those with the inv(16)/t(16;16) alteration.

On multivariate analysis, with adjustment for age, white blood cell (WBC) count, transplantation status, and cytogenetic characteristics, all the associations in the univariate analyses remained significant, except for the findings for MLL-PTD, PHF6, and ASXL1 mutations.

Mutational Analysis Finds Large Risk Differences

In patients with intermediate-risk AML on the basis of cytogenetic analysis, FLTD-ITD mutations were associated with reduced overall survival (P = .008), and multivariate analysis showed that FLTD-ITD mutations were the primary predictor of outcome. It was also found that patients with NPM1 mutations and IDH1 or IDH2 mutations had significantly improved 3-year overall survival (89% vs 31%, P < .001).

Patients with wild-type FLTD-ITD could be categorized into three risk groups with marked differences in 3-year overall survival (adjusted P < .001): those with NPM1 and IDH1 or IDH2 mutations—89% (favorable risk); those with wild-type TET2, ASXL1, PHF6, and MLL-PTD without NPM1 or IDH2 mutations—46.2% (intermediate risk); and those with TET2, ASXL1, PHF6, and MLL-PTD mutations—6.3% (high risk).

After analyzing differences in overall survival according to other mutations in patients with mutant FLT3-ITD, it was found that patients could be categorized into three risk groups according to 3-year overall survival: those with trisomy 8 or TET2, DNMT3A, or MLL-PTD mutations—14.5% (high risk); those with wild-type CEBPA, TET2, DNMT3A, and MLL-PTD—35.2% (P < .001 vs high-risk group); and those with CEBPA mutations--42% (P < .001 vs high-risk group).

Overall, the mutational analysis allowed patients at cytogenetically intermediate risk to be distinguished into three risk profiles—a favorable mutational risk profile, with a 3-year overall survival of 85%; intermediate mutational risk profile, with a 3-year overall survival of 42%; and unfavorable mutational risk profile, with a 3-year overall survival of 13%.

New Prognostic Schema Changes Risk Distribution

These findings permitted the investigators to develop a prognostic schema that integrated the findings from their mutational analysis with cytogenetic classification to identify three risk profiles: favorable risk profile, with median survival not reached and 3-year overall survival of 64%; intermediate risk profile, with median survival of 25.4 months and 3-year overall survival of 42%; and adverse risk profile, with median survival of 10.1 months and 3-year overall survival of 12%. This schema predicted outcome independently of age, WBC, induction dose, and transplantation status on multivariate analysis (adjusted P < .001) and was found to be accurate irrespective of type of post-remission therapy (autologous or allogeneic transplantation or consolidation chemotherapy).

The prognostic schema was shown to predict outcome when analyzed in an independent validation cohort of 104 patients from the ECOG E1900 trial (adjusted P < .001). The predictive value of the schema was independent of risk with regard to treatment-related death (death within 30 days after starting treatment) or lack of response (complete remission) to induction therapy and in the combined test and validation cohorts.

On the basis of cytogenetic classification alone, 19% of patients were categorized as having favorable risk (3-year overall survival of 58%), 63% as having intermediate risk (3-year overall survival of 36%), and 18% as having unfavorable risk (3-year overall survival of 11%). The integration of mutational risk into risk classification resulted in an increase in proportion of favorable risk patients to 26%, a decrease in proportion of intermediate risk patients to 35%, and an increase in proportion of unfavorable risk patients to 39% (with the respective 3-year overall survival rates noted above) compared with the cytogenetic classification alone.

Mutations Linked to Improved Survival

The ECOG E1900 trial showed that induction therapy including high-dose daunorubicin improved outcome compared with standard-dose daunorubicin.2 In their mutational analysis, Patel and colleagues found that high-dose daunorubicin was associated with improved survival in patients with mutant DNMTA (P = .04), but not in those with wild-type DNMTA.1 Univariate analysis also showed that high-dose daunorubicin was associated with improved survival in patients with MLL translocations (P = .01, but P = .06 with adjustment for multiple testing) and NPM1 mutations (P = .01, but P = .10 with adjustment for multiple testing).

Overall, high-dose daunorubicin was associated with a marked improvement in overall survival (P = .001) in patients with DNMT3A or NPM1 mutations or MLL translocations, but not in those without these alterations. The finding was independent of age, WBC count, and status with regard to transplantation, treatment-related death, or response to chemotherapy (P = .008 for patients with mutations, P = NS for those with wild-type genes). Overall, high-dose daunorubicin was associated with greater 3-year overall survival than standard-dose daunorubicin (44% vs 25%) in patients with these alterations. For all other genotypes, 3-year overall survival was 35% in patients receiving high-dose daunorubicin and 39% in those receiving standard-dose daunorubicin.

The current evaluation for risk in AML includes assessment for FLT3, NPM1, and CEBPA alterations. In providing context for their findings, the investigators noted, “[Our] data show that mutational analysis of a larger set of genetic alterations than that currently used in the clinic setting could be used to retrospectively classify patients with AML into more precise subgroups with favorable-risk, intermediate-risk, and unfavorable-risk profiles, with marked differences in the overall outcome. This approach could be used to identify an additional subgroup of patients who would have a mutationally defined favorable outcome with induction and consolidation therapy alone and a subgroup of patients with mutationally defined unfavorable risk who would potentially be candidates for allogeneic stem-cell transplantation or participation in a clinical trial.”

Clinical Impact

How might the findings of these investigators affect current practice? In an accompanying editorial, Lucy A. Godley, MD, PhD, of the University of Chicago stated “If we think about extending the findings of Patel and colleagues to clinical practice, we would need to know the genetic profile of patients with AML within the first few days of presentation, in order to tailor an induction regimen to the patient.”3 She noted that rapid clinical assessment is currently being used in a Cancer and Leukemia Group B study, in which patients with core-binding-factor leukemias are being identified within 48 hours of presentation in order to test the effectiveness of adding dasatinib [Sprycel] to induction and consolidation therapy.

Dr. Godley further stated, “A critical question is whether the data presented by Patel and colleagues are sufficient to justify the expansion of the number of genetic mutations being examined in patients with AML at presentation. [Their] findings challenge the field to address at what point data are compelling enough to change routine practice.” ■

Disclosure: The study authors reported no potential conflicts of interest. Dr. Godley reported an institutional contract with Celgene to provide mass spectrometry services for mouse samples.


1. Patel JP, Gönen M, Figueroa ME, et al: Prognostic relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med 366:1079-1089, 2012.

2. Fernandez HF, Sun Z, Yao X, et al: Anthracycline dose intensification in acute myeloid leukemia. N Engl J Med 361:1249-1259, 2009.

3. Godley LA: Profiles in leukemia—editorial. N Engl J Med 366:1152-1153, 2012.