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

After the initial induction chemotherapy for AML, physicians are often reluctant to offer further treatment if the bone marrow has <5% blasts, even in the setting of persistently low neutrophil (ANC) and/or platelet (plt) counts. This reluctance would seem to presume that complete remission (CR) might still occur in these patients without further treatment. However, it has been shown that there is an inverse relation between (A) time to achievement of CR after course 1 of induction therapy and subsequent survival¹  and (B) level of ANC and plt count at CR and subsequent relapse-free survival² , with (A) and (B) independent of each other and of cytogenetics and antecedent hematologic disorder. Because patients often present with low ANC and plt counts, these data suggested that their persistence was a clinical indicator of minimal residual disease (MRD) in the marrow, and prompted us to examine whether slow ANC and/or plt recovery despite a marrow with <5% blasts affected the probability of subsequent CR.

We included patients who, at various times after course 1 of induction chemotherapy, had not achieved both ANC >500 and plt >50,000, despite a bone marrow in the prior week with <5% blasts by morphology and flow cytometry. We originally selected these times to be day 28, day 35, day 42, and day 49; however because patients were often treated as outpatients, counts at these specific days were not uniformly available, and thus “day 28” included days 21–28, etc. A variety of “intensive” induction regimens were used. We examined whether incomplete count recovery, defined as (a) neither ANC >500 nor plt >50,000 or (b) either ANC >500 or plt >50,000 but not both, affected eventual CR rates as a function of days from induction chemotherapy.

Seventy-five patients were eligible for analysis, including 3 patients with blood and marrow assessments done at 2 time points (ANC <500 and plt <50,000 at days 28/42, 35/42, and 35/42). Median patient age was 53 years (range 18–78), with an antecedent hematologic disorder seen in 28 patients (37%). By SWOG cytogenetic risk categories, 12 patients (16%) had favorable, 35 (47%) intermediate, 21 (28%) unfavorable, and 7 (9%) unknown risk. Eventual course 1 CR rates for the specific patient groups are indicated below, with only the first time point included for each of the above 3 patients with multiple evaluations.

At day 28, patients with ANC <500 and plt<50,000 were much less likely to obtain CR than patients with either ANC >500 or plt>50,000 (31% vs. 85%, p=0.001). As time from start of treatment increased without rise in ANC to >500 and plt to >50,000, the probability of subsequent CR decreased (p=0.03) despite a marrow with < 5% blasts by both morphologic and flow assessment.

Persistence of ANC <500 and/or plt <50,000, despite the presence of <5% bone marrow blasts, is predictive of low eventual CR rates after induction chemotherapy, with the likelihood of obtaining CR decreasing with greater time from start of induction chemotherapy. We are particularly interested in seeing whether blood count recovery data adds information to such pretreatment predictors of CR as cytogenetics, age, and presence of antecedent hematologic disorder, and in examining the relation between sluggish count recovery and induction regimen, marrow cellularity, and degree of MRD by flow. However, our data suggests that optimism regarding the possibility of CR without further therapy (e.g. hematopoietic cell transplant) should wane in the setting of slow ANC and/or platelet recovery after course 1 of induction chemotherapy.

No relevant conflicts of interest to declare.