Probability of Eventual CR After Course 1 of Induction Therapy for Newly-Diagnosed AML As a Function of Platelet and Neutrophil Recovery.

After initial induction chemotherapy for acute myeloid leukemia (AML), it is commonplace that reinduction or intensified therapy is not indicated if the bone marrow has <5% blasts, even despite persistently low neutrophil (ANC) and/or platelet counts. This practice suggests that complete remission (CR), i.e. ANC >1000/μl and platelet count >100,000/μl per standard criteria (Cheson BD, et al. J Clin Oncol. 1990;8(5):813-9), might still occur and that the lack of blood count recovery may not bear prognostic significance. However, the time to CR after the first induction has been shown to be inversely related to subsequent duration of disease-free survival (DFS) and survival (OS), independent of age, treatment, and cytogenetics (Estey EH, et al. Blood. 2000;95(1):72-7). Additionally, the level of ANC and platelet recovery at time of CR is prognostic, with significantly better DFS among patients with higher counts (Yanada M, et al. Leuk Res. 2008;32(10):1505-9). Newly-diagnosed AML patients often present with below normal neutrophil and platelet counts, suggesting that persistence of such cytopenias after induction may be a clinical indicator of minimal residual disease (MRD) in the marrow. We therefore examined whether blood count recovery affected the probability of subsequent CR in patients with <5% bone marrow blasts.

We included 85 patients who, by day 21 or thereafter of induction therapy for newly-diagnosed AML, had not met blood count criteria for CR despite a bone marrow in the prior week with <5% blasts by morphology. Patients were classified by type of induction therapy based on cytarabine dosing. G-CSF was not systematically administered. Marrows were planned for day 21 after chemotherapy and/or weekly thereafter to assess for disease status and evidence of marrow recovery. Because patients were often managed as outpatients, counts and marrows were not uniformly available and thus “day 28” included days 21–28, “day 35” included days 29–35, etc. If a patient had more than one marrow evaluation after day 21, we included only the first one.

Overall cohort CR rate was 64%. Eventual CR rate was significantly affected by platelet count, with 44% eventual CR for patients with platelets <30,000, 66% CR for platelets 30,000–100,000, and 95% CR for platelets >100,000. The effect of ANC recovery on eventual CR was less dramatic, with an OR 0.4 (0.2–1.0, p=0.049), for ANC <0.1 vs. >0.1 in the univariate analysis. By day 28, patients with either ANC or platelet recovery were significantly more likely to obtain CR than patients with neither count recovery (89% vs. 51%), OR 8.05 (2.2–30, p=0.002). In the multivariate analysis, (a) lack of platelet recovery to >30,000 was associated with significantly lower incidence of CR, OR 0.26 (0.1–0.8, p=0.02), and was independent of cytogenetic risk, antecedent hematologic disorder, and induction regimen, and (b) there was a suggested association between earlier count recovery and CR (>28 days vs. day 21–28), OR 0.31 (0.1–1.0, p=0.051).

Persistence of low peripheral blood counts, despite the presence of <5% bone marrow blasts, is predictive of low eventual CR rates after induction chemotherapy. These results suggest that initiation of further and possibly different therapy, rather than continued observation, should be investigated in this setting.

Becker:Sanofi: Research Funding.