Objective: To evaluate the impact of the percentage of residual blasts in bone marrow at the end of induction chemotherapy (T1) or during myelosuppression phase (T2) on prognosis of de novo acute myeloid leukemia (AML) (non M(3)) in 105 cases. To refine AML risk-stratification by combining the percentage of residual blast cells (T1 or/and T2) with cytogenetic data based the South West Oncology Group (SWOG) criteria.
Methods: The data of 105 de novo AML (non M(3)) patients hospitalized between January 1st 1999 and February 1st 2008 were retrospectively reviewed. Results were analyzed with SPSS15.0 software.
Results: (1) Patients were divided into two subgroups by a cutoff of 5% residual bone marrow blasts at T1 or T2 time point. Patients with percentage of residual bone marrow blast cells < 5% had better complete remission (CR) rate, relapse-free survival (RFS) and overall survival (OS) than the patients with percentage > or = 5% at T1 or T2. The percentage of residual bone marrow blast cells at T1 was correlated with that at T2. (2) The prognosis of patients with intermediate karyotypes with percentage < 5% at T1 or T2 was similar to that of the patients with favorable karyotypes. The patients with intermediate karyotypes and percentage of residual bone marrow blasts > or = 5% at T1 or T2 are defined as a subgroup with prognosis similar to that of patients with unfavorable karyotypes. (3) COX regression analysis showed that the percentage of residual bone marrow blasts at T1 or T2 is an independent prognostic factor of AML. The percentage of residual bone marrow blasts at T1 may be more helpful in prognostication than that at T2.
Conclusion: AML patients with percentage of residual bone marrow blasts < 5% after induction chemotherapy (T1 or T2) have better CR rate, RFS, OS than the patients with percentage > or = 5% at the same time point. Combination of cytogenetics and percentage of residual bone marrow blasts at T1 or T2 is helpful to divide patients with intermediate karyotypes into two subgroups with different prognosis. Thus, a better decision of treatment strategy can be designed.