The Outcome in AML Is Predicted by Cytogenetics, NPM1/FLT3 Mutation, Age and Sex, but Not by Treatment Variables.

In order to test current risk factors in a prospective multicenter setting we evaluated the AMLCG 99 trial. Patients were randomly assigned to induction by TAD-HAM (HAM with araC 3 for age <60y and 1 for age ≥60y g/m² × 6), or HAM-HAM, and also to TAD consolidation and maintenance or (age <60y) myeloablative chemotherapy and autologous SCT. Patients with histocompatible family donors preferentially underwent allogeneic SCT. Since any randomization was done up-front, informations from completely unselected patients were available. 2547 patients of 16–85 (median 61) y entered the trial. 1858 pts had de-novo and 689 pts secondary AML. The CR rate was 61%, 54% in older (60) and 69% in younger patients. The overall survival (OS) at 4 years was 27%, 15% in older and 41% in younger patients. The relapse risk (RR) was 65%, 80% in older and 50% in younger patients and the relapse-free survival (RFS) was 30%, 14% and 44%, respectively. In the entire patients complete outcome (CR, OS, RR, RFS) was predicted by favorable and unfavorable karyotype. Among patients with any abnormal karyotype complete outcome was predicted by unfavorable karyotype in the older and favorable karyotype in the younger age group. In both age groups with normal karyotype outcome for the complete parameters was predicted by the NPM1+/FLT3- ITD- mutation status. As a new finding in patients of <60 years with normal karyotype female sex turned out being an independent predictive factor for longer OS (HR 1.45;95%CI 1.04–2.03), longer RFS (HR1.64;95%CI 1.10–2.44), and lower RR (HR 0.59;95%CI 0.38–0.92). Female sex was the only predictive factor besides the NPM1/FLT3 mutation status in this group. The OS at 4 years in patients of <60y with normal karyotype is 52% in women and 40% in men (log-rank P=0.047), the RR is 37% and 52% (P=0.016), and the RFS is 55% in women and 42% in men (P=0.025). Furthermore, the favorable NPM1+/FLT3- mutation status was more frequent in women than in men (35% vs 26%; P=0.0075). Remarkably, the NPM1+/FLT3- mutation status was equally predictive in patients of ≥60y as in those of <60y with HR for OS of 2.51 (95% CI 1.75–3.61) and 3.27 (95%CI 2.11–5.05) and HR for RR of 0.33 (95% C 0.21–0.51) and 0.29 (95%CI 0.17–0.49). The difference in the OS at 4 years between patients with NPM1+/FLT3- mutation and those with other NPM1/FLT3 combinations was 42% vs 18% (P=<0.001) in the older, and 69% vs 34% (P<0.001) in the younger patients. The related differences in RR were 58% vs 84% (P<0.001) in the older, and 22% vs 59% (P<0.001) in the younger patients. Among the NPM1/FLT3 mutations the favorable +/− constellation accounted for 27% of older, and 35% of younger patients (P=0.0197). Besides karyotypes and mutations, also age, de-novo AML, blast clearance, LDH and WBC partly predicted outcomes. In contrast no prognostic impact was found by multi-and univariate analyses of treatment alternatives.

Conclusion: As from a large multicenter prospective trial the outcome in AML is mainly determined by cytogenetics, NPM1/FLT3 mutation, age and sex, but not by the assigned treatment variables.