Kinetics of bone marrow blasts during induction and achievement of complete remission in acute myeloid leukemia

In acute myeloid leukemia (AML), bone marrow is typically examined 14 days after beginning initial induction therapy. If significant residual blasts remain, the National Comprehensive Cancer Network (NCCN) Guidelines for AML recommend re-treatment.1 Here we examine whether bone marrow findings on day 21 might modulate the day 14 findings and thus influence the decision to begin a second course.

Our database comprised those 586 adults who had both day 14 and day 21 bone marrows (±2 days) after receiving, from 1995 to 2004, cytarabine (≥1g/m per day) -containing therapy for newly diagnosed AML. Their median age was 60 years. Six percent had inv(16)/t(16;16) or t(8;21), 33% had abnormalities of chromosomes 5 and/or 7 or complex karyotype, and 61% had other findings. In addition to high-dose cytarabine, induction therapy contained idarubicin without fludarabine or topotecan in 18%, fludarabine with or without idarubicin in 35%, and topotecan without idarubicin in 32%, and other therapies n 15%.2 This study was approved by the M.D Anderson Institutional Review Board, and patients were treated in accordance with the Declaration of Helsinki.

Bone marrow status was defined as “too few cells to count (TFTC)” if the total cell count was less than 100 after reviewing 4 slides. Complete remission (CR) was defined by a bone marrow with <5% blasts, a neutrophil count ≥1×10/L and a platelet count ≥100×10/L. Groups were compared using Fisher’s exact test for categorical variables, and the Wilcoxon rank-sum test for numerical variables.

Three-hundred and seventy-five patients (64%) achieved CR on course 1. A second course with a similar regimen was given to 64 patients, and resulted in CR in an additional 19 patients. As expected, the probability of CR on course 1 decreased as the percentage of blasts in either the day 14 or the day 21 marrow increased, such that only a minority of patients who had ≥20% blasts on either date entered CR on this course (Table 1). Of most interest are the probabilities of course 1 CR according to the combination of the day 14 marrow with the day 21 marrow (Table 2). In particular, 37 of the 72 patients (51%) with 20–59% blasts on day 14 had <20% blasts on day 21, and 23 of the 37 (62%) entered CR on course 1 without further therapy. In contrast, patients with 20–59% blasts on day 14 whose day 21 marrow did not improve had a course 1 CR rate of only 8/26 (31%); TFTC on day 21 after 20–59% blasts on day 14 was not advantageous (CR rate 2/7). Thirty out of 37 (81%) patients with ≥60% blasts on day 14 continued to have ≥20% blasts on day 21 and the course 1 CR rate in these patients was only 7%, and was only 16% for all patients with ≥60% blasts on day 14 regardless of the day 21 findings. Patients with TFTC or <20% blasts in the day 14 marrow but whose day 21 marrows had ≥20% blasts seemed less likely to achieve CR, but such cases accounted for only 12% and 11% of each group. In all groups described above, failure to enter CR typically reflected resistance to therapy rather than death before response could be evaluated (eg. before day 35).

Table 1.Probability of complete remission according to bone marrow findings on days 14 and 21.

Previous studies have shown a significant correlation between the percentage of blasts in early bone marrow and the subsequent probability of CR.3–6 However, most such reports analyzed patients who received double induction (DI) therapy3,5,6 in which a second course was given regardless of the percentage of bone marrow blasts, so that CR was evaluated after the second course. In patients not given DI, however, a decision must be made whether to start a second course or to wait for recovery; hence evaluation of early bone marrow is more important. The NCCN guidelines recommend that the decision be guided by bone marrow 7–10 days after completion of induction therapy.1 However, clinical experience suggests that patients with a high blast percentage in early bone marrow can enter CR without further therapy, prompting us to investigate whether the day 21 marrow adds useful information to the day 14 marrow. Our results show that the day 21 marrow does not materially alter the decision to begin course 2 as based on the day 14 marrow in patients with TFTC, <20% blasts, or ≥60% blasts in the day 14 marrow. However, approximately half of patients with 20–59% blasts in the day 14 marrow will have <20% blasts on day 21 and in these patients the probability of CR without administering a second course seems sufficiently high (62%) to warrant, in at least some patients, delaying a second course until the day 21 marrow can be examined.

Table 2.Probability of complete remission according to combination of bone marrow findings on days 14 and 21.

This conclusion is subject to several criticisms. First, our patients uniformly received cytarabine at doses considerably above those used in 3+7. Accordingly, the proportion of patients given 3+7 who have 20–59% blasts on day 14 but who have <20% blasts on day 21 might be less than the 51% noted here. Second, not all our patients who were alive on day 14 or day 21 had a marrow examined then. Third, we did not examine marrow cellularity, feeling that this might be particularly susceptible to inter-observer variability. Finally, delaying until day 21 in patients with 20–59% blasts on day 14 might not affect CR rate, but might shorten CR duration. However, this risk has to be weighed against the competing risk of giving a second induction course, particularly to older patients. Specifically, while there will be myelosuppression on either a second induction course or a first post-remission course, the risk of infection at any given neutrophil count is less when a patient is in CR than when not, and duration of neutropenia is often less in patients in CR. Hence in older patients physicians might prefer to wait until CR to re-treat. Our data make the option of delay more plausible and suggest the need to revisit the NCCN recommendations in patients given 3+7.