Multivariate Analysis of Prognostic Impact of Flow Cytometrically Determined Minimal Residual Disease in Patients with Acute Myeloid Leukemia.

Treatment of acute myeloid leukemia (AML) is adapted according to individual risk profiles. The assessment of levels of minimal residual disease (MRD) is feasible in the vast majority of patients with AML and may significantly improve prognostication. We applied multiparameter flow cytometry to highly sensitively quantify MRD in patients with AML and analyzed its prognostic impact in a multivariate analysis. A total of 356 patients receiving standardized intensive antileukemic treatment was analyzed at different checkpoints: CP1 (up to day 28, n=156), CP2 (day 28–60, n=122), CP3 (day 61–120, n=195), CP4 (day 121–365, n=172), and CP5 (after day 365, n=73). Leukemic cells were identified by their individually defined leukemia-associated aberrant immunophenotypes (LAIPs). MRD levels were calculated as the logarithmic difference (LD) between LAIP-positive cells at diagnosis and LAIP-positive cells at follow-up assessment. The median LD amounted to 2.18 (range, −0.14 to 4.20) at CP1, 2.31 (−0.03 to 4.17) at CP2, 2.49 (0.11 to 4.17) at CP3, 2.58 (−0.28 to 4.28) at CP4, and 2.87 (0.46 to 4.02) at CP5. A higher LD (continuous variable) was related to a better event-free survival (EFS; CP1, p=0.0002; CP2, p=0.00001; CP3, p=0.0002; CP4, p<0.00001; CP5, p=0.00007) and to a better overall survival (OS; CP1, p=0.004; CP2, p=0.001; CP3, p=0.021; CP4, p=0.00006). Other parameters related to EFS and OS were age and cytogenetics in the present series. The prognostic impact of MRD levels on outcome was idependent of cytogenetics and age for EFS (CP2 to CP5) and OS (CP2 and CP4). MRD was the most important prognostic parameter at CP4 and CP5. Separation of patients into two groups, respectively, by the median LD resulted in significant differences in EFS at all CPs and in OS at CP1 to CP4. The largest difference was observed at CP4: median EFS, 57.1 vs. 13.7 months, p<0.00001; 3-year-OS, 95% vs. 65%, p=0.0003. Determination of MRD levels by multiparameter flow cytometry provides a highly powerful and independent prognostic parameter which is applicable to the total of an AML population. It should be evaluated as a stratification parameter in clinical trials.