Adult patients with acute myeloid leukemia who achieve complete remission after 1 or 2 cycles of induction have a similar prognosis

A report on 1980 patients registered to 6 studies conducted by the Eastern Cooperative Oncology Group

Authors


  • See editorial on pages 4896–901, this issue.

Abstract

BACKGROUND:

Patients with newly diagnosed acute myeloid leukemia (AML) often have residual leukemia in the bone marrow 10 to 14 days after the start of induction therapy. Some cooperative groups administer a second cycle of similar induction therapy on Day 14 if there is residual leukemia. It is a common perception that the presence of residual leukemia at that point predicts a worse prognosis irrespective of the therapy received. The objective of this study was to determine whether patients who required a second cycle of induction (given on or about Day 14) to achieve complete remission (CR) had a worse prognosis than patients who achieved CR after only 1 cycle, because a worse prognosis may alter postremission therapy.

METHODS:

Patients who were enrolled on 6 consecutive studies for AML that were conducted by the Eastern Cooperative Oncology Group (ECOG) between 1983 to 1993 received induction therapy. If residual leukemia was present in the bone marrow on the Day 14 after the start of induction, then patients were to receive a second cycle of identical induction therapy. All patients who achieved CR after 1 or 2 cycles received the identical postremission therapy.

RESULTS:

In each of the 6 ECOG studies, the long-term outcome was similar for patients who required 1 or 2 cycles of induction therapy to achieve CR, and their outcome was independent of other prognostic variables, such as age or karyotype.

CONCLUSIONS:

The presence of residual leukemia in bone marrow 10 to 14 days after induction therapy did not predict a worse prognosis if patients received second, similar cycle of induction therapy and achieved CR. Cancer 2010. © 2010 American Cancer Society.

Prognostic factors for acute myeloid leukemia (AML) at diagnosis include cytogenetics.1, 2 the fms-related tyrosine kinase 3 gene (FLT3),3, 4 the CCATT/enhancer binding protein alpha gene (CEPBA),5, 6 the Wilms tumor gene (WT1),7 the myeloid-lymphoid or mixed lineage leukemia gene (MLL),8 the nucleophosmin gene (NMP1),9 age,10 multidrug resistance (MDR) status,11 and white blood cell (WBC) count at presentation.12 The time to achieve complete remission (CR)13 and the time to clearance of peripheral blast cells14, 15 have been reported as independent postinduction prognostic factors in patients with AML. Furthermore, some investigators have considered patients who do not achieve CR with 1 cycle as refractory or as primary induction failure.16, 17 In those studies, patients who failed to achieve CR after 1 cycle generally were not given additional therapy on Day 14 if the bone marrow demonstrated residual leukemia, and most (if not all) of such patients can be expected not to be in CR at the end of induction therapy. Residual leukemia present in bone marrow on Day 14 can be an early indicator of a highly resistant clone, but it also can represent a slower response to therapy. Although it is possible that some patients with residual leukemia will achieve a subsequent CR,18 most such patients will not be in CR at the end of induction. At the same time, some patients may enter CR if early therapeutic intervention, on or about Day 14, is administered. Furthermore, if, on or about Day 14, further therapy is given and the patient achieves CR, little is known about the long-term outcome. Thus, it remains to be determined whether it is appropriate to consider a patient with residual leukemia on Day 14 as refractory.

The standard practice of the Eastern Cooperative Oncology Group (ECOG) for the treatment of AML mandates that, if the bone marrow on Day 10 to Day 14 of induction demonstrates unequivocal residual leukemia in a bone marrow sample that is not hypocellular, then a repeat course of the identical induction therapy is given at that point. The planned postremission therapy is identical whether 1 or 2 cycles are required to achieve CR.

Because the postremission therapy is not altered by the presence of residual leukemia on approximately Day 14, it is possible to evaluate its true impact on long-term prognosis. For the current report, we evaluated patients who were treated on 6 consecutive ECOG studies, which were conducted between 1983 and 1993 and comprised approximately 2000 patients, to assess whether long-term survival was affected by the number of cycles given to achieve CR. This issue is important to address if patients who require a second cycle of induction therapy to achieve CR have a worse prognosis, because it may change their postremission strategy.

MATERIALS AND METHODS

Between 1983 and 1993, 1980 patients were registered on 6 consecutive ECOG studies (E3483, PC486, E1490, E2491/INT0129, E3489, and E3993). All studies were approved by the institutional review boards, and all patients provided written informed consent. These studies were for patients ages 18 to 55 years (E3483, PC486, and E3489), ages 56 to 70 years (E1490), aged >56 years (E3993), and with no age limit (E2491/INT0129) (Fig. 1). Of all 1980 patients, 1272 achieved CR (64%) and were available for analysis. In each study, patients received standard induction therapy, which consisted of daunorubicin 60 mg/m2 (E3483, PC486, and E1490) or 45 mg/m2 (E2491/INT0129); or idarubicin 12 mg/m2 (E3489); or a randomization between daunorubicin 45 mg/m2, idarubicin 12 mg/m2, or mitoxantrone 12 mg/m2 (E3993)—all for 3 days—together with cytarabine 100 mg/m2 to 200 mg/m2 as a continuous infusion for 7 days (Fig. 1). It should be noted that in the North American Intergroup Acute Promyelocytic Leukemia (APL) study (E2491/INT0129), patients who received all-trans-retinoic acid (ATRA) at induction were excluded from the current analysis. In each study, patients were to receive a second cycle of identical induction therapy if the Day 10 to Day 14 bone marrow demonstrated unequivocal residual leukemia. Identical study-specific postremission therapy, including consolidation therapy and/or allogeneic or autologous bone marrow transplantation, was received by all patients who achieved CR irrespective of whether it was achieved after 1 cycle or after 2 cycles.

Figure 1.

(A) Study E3483 for patients with newly diagnosed acute myeloid leukemia aged <65 years was the only prospective study in young adults that included an observation arm postremission. IV indicates intravenous; PO, orally; CR, complete remission, BMT, bone marrow transplantation; HD, high dose. (B) Study PC486 study was a pilot study for young adults of induction therapy to be followed by autologous transplantation with in vitro purging for hydroxyl-peroxy-cyclophosphamide. HSCT indicates hematopoietic stem cell transplantation. (C) E3489 was a major intergroup study that was designed to evaluate 3 prospective postremission therapies given at approximately the same time point. (D) E1490 evaluated the role of granulocyte–macrophage-stimulating factor (GM-CSF) when given after bone marrow aplasia was identified on Day 10. (E) E2491/INT 0129 was a North American intergroup study that established the role of all-trans-retinoic acid (ATRA) in induction and as maintenance therapy. For the purpose of the current analysis, patients who received ATRA in induction were not included. (F) E3993 was a study in older patients that compared 3 induction regimens and randomized patients to receive either GM-CSF or placebo as priming therapy.

Study E3483

E3483 (Fig. 1A) was a phase 3 trial for patients aged 18 to 56 years that consisted of therapy with daunorubicin 60 mg/m2 for 3 days, cytarabine 200 mg/m2 for 7 days, and 6-thioguanine 100 mg/m2 for 5 days as induction therapy followed by postremission therapy, which consisted of allogeneic hematopoietic stem cell transplantation (HSCT) if a patient had a histocompatible (human leukemic antigen [HLA]) sibling and was aged <55 years, high-dose therapy with cytarabine and amsacrine, and maintenance therapy with low-dose cytarabine and 6-thioguanine; an observation arm also was included.19

Study PC486

PC486 (Fig. 1B) was a phase 2 study for patients aged 18 to 55 years using induction therapy with daunorubicin 60 mg/m2 for 3 days and cytarabine 200 mg/m2 for 7 days. Patients in CR underwent autologous HSCT without any prior intensification.20

Study E3489

E3489 (Fig. 1C) was a US intergroup study for patients aged 18 to 55 years who received induction therapy with idarubicin 12 mg/m2 for 3 days together with cytarabine 100 mg/m2 for 7 days. If they were in CR, then all patients received a second cycle of “attenuated” induction (“2 + 5”) consisting of idarubicin and cytarabine. The purpose of this added therapy at that point was to enable the various postremission modalities to be given at the same time, which included the period needed (in the year 1989) to search for a donor and refer a patient to a transplantation center. For postremission, patients were assigned to undergo allogeneic HSCT if they had an HLA-compatible sibling. Otherwise, they were randomized to undergo autologous HSCT versus a single course of high-dose cytarabine.21

Study E1490

E1490 (Fig. 1D) was a study for patients aged 55 to 70 years. For induction, they received daunorubicin 60 mg/m2 for 3 days and cytarabine 100 mg/m2 for 7 days. All patients were randomized to receive either granulocyte-macrophage–colony-stimulating factor (GM-CSF) or placebo at induction. For consolidation, patients received 1 course of high-dose cytarabine (attenuated for age) with the same randomized study drug (GM-CSF or placebo) that they had received during induction.22

Study E2491/INT0129

E2491/INT0129 (Fig. 1E) was an international intergroup study with no age limit in which induction therapy consisted of daunorubicin 45 mg/m2 for 3 days and cytarabine 100 mg/m2 for 7 days. This induction regimen was compared with ATRA given as a single agent without concurrent chemotherapy. For consolidation, patients received 2 cycles of daunorubicin and cytarabine followed by a maintenance randomization to either ATRA therapy for 1 year or observation.23

Study E3993

E3993 (Fig. 1F) was a study for patients aged ≥56 years with no upper age limit. They were randomized at induction to receive daunorubicin 45 mg/m2 for 3 days, versus mitoxantrone 12 mg/m2 for 3 days, versus idarubicin 12 mg/m2 for 3 days. All patients received cytarabine 100 mg/m2 for 7 days. At induction, there also was a randomization to GM-CSF given as priming therapy versus placebo. For consolidation, patients received only 1 cycle of high-dose cytarabine (attenuated for age).24

Statistical Analysis

Two primary endpoints were explored in this analysis: disease-free survival (DFS) and overall survival (OS). DFS was defined as the time from documented CR until relapse or death from any cause. The Kaplan-Meier (KM) method was used to estimate DFS and OS, and the log-rank test was used to test the difference in KM curves between 1 cycle and 2 cycles of induction therapy. Furthermore, multivariate Cox regression analysis was used to investigate whether the number of cycles of induction therapy had prognostic significance in the presence of other known prognostic factors for AML, such as age at study entry, sex, race, baseline hemoglobin levels, WBC count, platelet count, and karyotype, although the latter was available only for studies E3489 and E3993. The proportional hazards assumption in the Cox model was assessed, and all 2-way interactions were examined. In addition, baseline characteristics were compared using a 2-sided Wilcoxon rank-sum test for continuous variables and the Fisher exact test for categorical variables.

All patients who achieved CR, which we defined according to the International Working Group,25 were included in this analysis. Thus, the study also included those patients who achieved CR but did not receive further postremission therapy.

RESULTS

Baseline Characteristics

Baseline characteristics are listed in Table 1. Apart from age in E3483 and white cell count in E2491, the number of induction cycles that led successfully to CR was not affected by age, WBC count, platelet count, or hemoglobin level at presentation.

Table 1. Baseline Characteristics
ECOG Study/CharacteristicNo. of Induction CyclesAll PatientsPa
1 Cycle2 Cycles
  • ECOG indicates Eastern Cooperative Oncology Group; WBC, white blood cell count; Plts, platelet count; Hb, hemoglobin.

  • a

    P values for comparison of continuous variables calculated using two-sided Wilcoxon rank-sum test.

E3483    
 No. of patients189128317 
 Median age [range], y44 [15-65]38 [15-65]41 [15-65].004
 No. of males (%)98 (52)65 (51)163 (52).82
 Median WBC [range], ×103/μL16.4 [0.3-270]12.5 [0.5-271.2]15.6 [0.3-271.2].49
 Median Plts [range], ×103/μL51 [7-434]57.5 [1-520]52.5 [1-520].17
 Median Hb [range], g/dL9.5 [0.5-13.9]9.3 [4-14.7]9.3 [0.5-14.7].65
PC486    
 No. of patients335184 
 Median age [range]32 [17-53]34 [16-54]34 [16-54].73
 No. of males (%)20 (63)25 (49)45 (54).26
 Median WBC [range], ×103/μL7.9 [0.5-147]8.5 [0.4-159.8]8.1 [0.4-159.8].76
 Median Plts [range], ×103/μL63 [11-400]68 [11-999]65.5 [11-999].17
 Median Hb [range], g/dL9.3 [6.4-12.8]9 [3.5-15.1]9.1 [3.5-15.1].68
E3489    
 No. of patients45375528 
 Median age [range], y38 [17-55]39 [16-54]38 [16-55].43
 No. of males (%)219 (48)30 (41)249 (47).26
 Median WBC [range], ×103/μL13.2 [0.2-360]11.2 [0.3-396]12.9 [0.2-396].17
 Median Plts [range], ×103/μL51 [3-820]44 [6-259]49 [3-820].22
 Median Hb [range], median9 [0.2-16.4]9 [4.7-13.5]9 [0.2-16.4].91
E1490    
 No. of patients551267 
 Median age [range], y63 [56-70]62.5 [56-69]63 [56-70].81
 No. of males (%)32 (58)3 (25)35 (52).06
 Median WBC [range], ×103/μL4.6 [0.8-201]3.4 [1-64.8]3.8 [0.8-201].48
 Median Plts [range], ×103/μL64 [9-273]58 [14-312]64 [9-312].97
 Median Hb [range], g/dL9.6 [4.2-14.9]9.9 [7.6-11.1]9.6 [4.2-14.9].95
E2491    
 No. of patients8739126 
 Median age [range], y40 [1-81]36 [7-68]39 [1-81].33
 No. of males (%)38 (44)14 (36)52 (41).44
 Median WBC [range], ×103/μL2.9 [0.2-139]1.5 [0.3-82.8]2.3 [0.2-139].03
 Median Plts [range], ×103/μL30.5 [2-223]25 [3-220]30 [2-223].51
 Median Hb [range], g/dL9.1 [4.5-14.3]8.8 [4.2-12.5]9.1 [4.2-14.3].91
E3993    
 No. of patients12822150 
 Median age [range], y66 [56-80]67 [58-78]66 [56-80]0.28
 No. of males (%)59 (46)9 (41)68 (45)0.82
 Median WBC [range], ×103/μL6.9 [0.6-200.2]2.5 [1.1-77]6.5 [0.6-200.2].12
 Median Plts [range], ×103/μL62.5 [7-351]52.5 [18-285]60 [7-351].80
 Median Hb [range], g/dL9 [4-14.4]9.6 [6.2-13.7]9.2 [4-14.4].18

Table 2 summarizes the CR rate in each study and the number of induction cycles required to achieve CR. In total, 1272 of 1980 patients (64%) achieved CR, including 74% who reached CR after 1 cycle and 26% who reached CR after 2 cycles. Table 3 provides a further breakdown of the CR rates and the number of cycles given in each study.

Table 2. Achievement of Complete Remission (CR) and the Number of Induction Cycles Required for CR
StudyNo. With CR/Total No. (%)1 Cycle, %2 Cycles, %
E3483317/485 (65)6040
PC48684/113 (74)3961
E3489528/748 (71)8614
E149067/111 (60)8218
E2491126/175 (72)6931
E3993150/348 (43)8515
All studies1272/1980 (64)7426
Table 3. Response in Each Study Depending on the Number of Induction Cycles
StudyNo. of Patients (%)Pa
1 Cycle2 CyclesAll
No CRCRNo CRCRNo CRCR
  • CR indicates complete remission.

  • a

    P values for the comparison of CR rates between 1 cycle and 2 cycles.

E348357 (23)189 (77)111 (46)128 (54)168 (35)317 (65)<.001
PC48613 (28)33 (72)16 (24)51 (76)29 (26)84 (74).66
E3489218 (33)453 (68)2 (3)75 (97)220 (29)528 (71)<.001
E149020 (27)55 (73)24 (67)12 (33)44 (40)67 (60)<.001
E249129 (25)87 (75)20 (34)39 (66)49 (28)126 (72).22
E3993141 (52)128 (48)57 (72)22 (28)198 (57)150 (43).002
All478 (34)945 (66)230 (41)327 (59)708 (36)1272 (64).001

In multivariate analysis, the important prognostic factors at diagnosis for achieving CR included age (E3483 and E3993), being a man (E3483 and E1490), WBC count (E3489), hemoglobin level (E3483), and cytogenetics (E3489 and E3993; data not shown). The rates of consolidation actually given to patients according to protocol were similar whether CR was achieved after 1 cycle or after 2 cycles.

Disease-Free and Overall Survival

The 5-year and 10-year DFS and OS rates for each study are listed in Table 4. The difference in long-term outcome was not affected significantly by the need for 1 cycles or 2 cycles to achieve CR in any study. In multivariate Cox regression model, these overall results demonstrating no difference between 1 cycle and 2 cycles were not affected by age, presenting WBC count, hemoglobin level, platelet count, or cytogenetics (data not shown) (Table 5). Karyotypes were routinely available only for studies E3489 and E3993. The distributions of the cytogenetic risk groups were similar between 1 cycle and 2 cycles for all patients or for patients who achieved CR only in those 2 studies (Table 6). In E2491, almost all patients had the abnormal 15;17 translocation. In the Cox model (Table 5), the hazard ratio was similar with and without karyotype data in these 2 studies (E3489 and E3993). In fact, the P value was even less significant with the inclusion of karyotype data in the model. Figure 2 graphically depicts the DFS in all 6 studies according to the number of induction cycles received.

Table 4. Five-Year and 10-Year Disease-Free and Overall Survival Rates According to the Number of Induction Cycles Needed to Achieve Complete Remission
ECOG StudyPercentage of Patients ± SE
OSPDFSP
1 Cycle2 Cycles1 Cycle2 Cycles
  1. SE indicates standard error; OS, overall survival; DFS, disease-free survival; ECOG, Eastern Cooperative Oncology Group; NA: not available.

E3483      
 5-Y29 ± 326 ± 4.5621 ± 324 ± 4.45
 10-Y22 ± 324 ± 4 17 ± 322 ± 4 
PC486      
 5-Y42 ± 937 ± 7.9339 ± 935 ± 7.87
 10-Y36 ± 835 ± 7 33 ± 833 ± 7 
E3489      
 5-Y46 ± 234 ± 6.2036 ± 228 ± 5.36
 10-Y41 ± 232 ± 5 34 ± 228 ± 5 
E1490      
 5-Y24 ± 60.1020 ± 50.26
 10-Y21 ± 60 18 ± 50 
E2491      
 5-Y57 ± 548 ± 8.4431 ± 538 ± 8.88
 10-Y51 ± 645 ± 8 29 ± 538 ± 8 
E3993      
 5-Y14 ± 314 ± 7.227 ± 25 ± 4.53
 10-YNANA NANA 
Table 5. Hazard Ratio of Induction Cycle 2 Compared With Induction Cycle 1 in a Multivariate Cox Model
ECOG StudyOSDFS
HRPHRP
  1. OS indicates overall survival; DFS, disease-free survival; HR, hazard ratio; ECOG, Eastern Cooperative Oncology Group.

E34831.200.171.05.71
PC4861.160.611.20.53
E34891.260.171.17.36
E14901.860.071.54.21
E24911.380.241.15.57
E39931.540.081.25.37
Table 6. Eastern Cooperative Oncology Group Studies E3993 and E3489: Cytogenetic Risk Group Breakdown by the Number of Induction Cycles
Cytogenetic Risk GroupE3993: No. of Patients (%)E3489: No. of Patients (%)
All Patients, n = 338CR Only, n = 145All Patients, n = 578CR Only, n = 411
1 Cycle, n = 3522 Cycles, n = 861 Cycle, n = 1232 Cycles, n = 221 Cycle, n = 5162 Cycles, n = 621 Cycle, n = 3512 Cycles, n = 60
  1. CR indicates complete remission.

Favorable2 (0.8)1 (1.2)2 (1.6)1 (4.6)106 (21)10 (16)88 (25)9 (15)
Intermediate98 (38.9)27 (31.4)57 (46)5 (23)236 (4631 (50)173 (49)31 (52)
Unfavorable59 (23.4)28 (32.6)23 (19)5 (23)152 (29)20 (32)78 (22)19 (32)
Indeterminate93 (36.9)30 (34.9)41 (33)11 (50)22 (4)1 (2)12 (3)1 (2)
P.36 .19 .6 .2 
Figure 2.

Disease-free survival is illustrated according to the number of induction cycles to achieve complete remission for Eastern Cooperative Oncology Group studies (A) E3483, (B) PC486, (C) E3489, (D) E1490, (E) E2491/INT0129, and (F) E3993.

DISCUSSION

Several studies have addressed the significance of Day 10 to Day 14 bone marrow status without regard to subsequent therapy that was given.26-30 The issue that we addressed in the current study is what happens to patients who have residual leukemia in bone marrow on Day 10 to Day 14, receive a second cycle of induction therapy at that point, and subsequently achieve CR. Do they have a worse prognosis at that point than patients who achieved CR after only 1 course of induction? The data reported from our study demonstrate that such patients do not have a worse long-term outcome than patients who achieve CR after 1 cycle; therefore, no modification of postremission therapy is warranted for these patients. It must be emphasized that the results from this study in no way suggest a recommendation for patients to receive a second cycle on Day 14. Rather, the study negates a common perception and practice that patients who achieve CR after a second cycle of induction given on Day 10 to Day 14 have a worse prognosis such that their planned postremission intervention needs to be altered.

The current analysis included patients who were registered on clinical trials since 1983. According to protocol, all patients who had residual leukemia on Day 10 to Day 14 were to receive a second course of therapy. Twenty-six percent of patients who eventually achieved CR required 2 courses of induction therapy, although the percentage of patients who received 2 courses varied between studies (Table 2). Above all else, this also reflects the feasibility and safety in young adults of administering a second course of induction therapy on or about Day 14. This does not suggest that all patients who received such therapy would not have achieved CR if a second cycle had not been given.18 Indeed, in some earlier studies of APL, residual leukemia on Day 14 may have been overdiagnosed.31 Furthermore, in the early studies, patients may have been diagnosed with residual leukemia based on a very hypocellular bone marrow. The rigorous criteria of having unequivocal residual leukemia in the bone marrow that is not hypocellular were instituted more commonly only in the protocols that were conducted in the early 1990s.32 Nonetheless, the rigorous requirement for the assessment of bone marrow that may not be hypocellular is not uniform, and, in most of the earlier analyses,33, 34 this was not required. The data in this report encompass 6 studies that were conducted at different time points, with possibly slightly different criteria for defining residual leukemia, but with an overall result that was similar and consistent in each study. Patients with APL who did not receive ATRA were included in this analysis, because they were a part of consecutive studies for patients with newly diagnosed AML or APL, and the results from the specific study (E2491/INT0129) are virtually superimposable on the results from the other studies and, excluding that study, do not alter the final conclusions. In addition, a few patients with APL were included in the earlier studies for AML (E3483, PC486, and E1490) and, thus, were not excluded.

The prognostic significance of the time to CR in this report (1 cycle or 2 cycles) was independent from other prognostic factors, such as age, presenting WBC count, hemoglobin level, platelet count, or karyotype. Although there is an overall perception that patients who have residual leukemia on approximately Day 14 have a worse prognosis, this perception should not affect the management of those patients who achieve CR if a second cycle is administered promptly on Day 14. Such patients had a long-term prognosis that was indistinguishable from the prognosis for patients who achieved CR after 1 cycle of induction therapy.

There is no doubt that the presence of residual leukemia in bone marrow on Day 14 portends a worse prognosis.13-17 In many studies, no further therapy is given on Day 14, and treatment decisions for postremission therapy are based entirely on whether or not a CR was achieved at the end of a single cycle. The actual presence of residual leukemia on Day 14 or earlier or the delayed clearance of blasts from the peripheral blood predicted a worse prognosis according to several publications.13-15, 18, 27-29, 33-35

Two major studies routinely administered 2 cycles of induction irrespective of whether CR was achieved after the first cycle. However, none administered this as early as Day 14. The Medical Research Council (MRC) in the United Kingdom evaluated more than 1700 patients on their AML10 trial.36 In a careful analysis, the 2 most predictive parameters for outcome were cytogenetics and response after 1 cycle of therapy. In that MRC trial, patients received 2 courses of induction therapy. However, the second course invariably was given at the end of completion of the first cycle whether or not a patient had achieved a CR. Patients never were treated as early as Day 14. The response was assessed only at the end of induction, which was about 4 weeks after the start of induction.33, 36 In the German AML Cooperative Group 1992 trial, patients received induction therapy with daunorubicin, cytarabine, and 6-thioguainine. A second course of therapy that consisted of high-dose cytarabine with mitoxantrone was administered on about Day 21. On Day 16, bone marrow was assessed for residual leukemia, but no therapy was given at Day 16. The authors reported that patients who had >10% blasts in bone marrow on Day 16 had only a 54% CR rate compared with patients who had <10% blasts in bone marrow on Day 16, who had an 84% CR rate. The adverse prognostic value of the Day 16 bone marrow was independent of all other variables, including cytogenetics.34 The impact of bone marrow cellularity on the assessment of the presence of bone marrow blasts was not discussed.

Several other studies assessed the time to CR or clearance of peripheral blasts. In a large retrospective analysis reported from the M. D. Anderson Cancer Center in 1996, 1101 patients with newly diagnosed AML or advanced myelodysplasia were examined. In that study, the authors observed a strong correlation between OS and the time to CR, which was an independent prognostic factor and was not affected by other prognostic variables such as karyotype.13 In that study, as in the previous studies, patients did not receive a second course of induction on Day 14 if there was evidence of residual leukemia at that point. An early report from over 2 decades ago assessed bone marrow on Day 6 for the degree of cellularity and residual blasts. In that study, 116 of 253 patients had >30% cellularity in the bone marrow biopsy or had >10% abnormal blast cells in the aspirate, and those patients received high-dose therapy on Days 8, 9, and 10 of induction. There was a nonstatistically significant trend toward an improved CR rate for the patients who received the added therapy on Day 8 (69%) versus those who did not (60%).28 Several smaller reports in recent years have described the relation between the rate of clearance of leukemic blasts from the peripheral blood and the likelihood of achieving a response.14, 15 In 1 report from Italy,14 patients who achieved CR had a significantly more rapid clearance of peripheral blood that could be determined from Day 2 to Day 6 of induction therapy. For example, on Day 6, patients who achieved CR had a 3.71 log reduction in their peripheral blasts compared with a 2.08 log reduction for those who did not achieve CR (P = .0001). Another report on 86 adult patients with previously untreated AML demonstrated once again the correlation between the overall recurrence-free survival and the rate of blast clearance and actually separated the patients into 3 prognostic groups. Patients at “good risk” were those who had blast clearance before Day 3, the “intermediate-risk” group included patients who had blast clearance on Days 4 or 5, and patients in the “poor-risk” group had peripheral blast clearance only after Day 6 (P < .001).15 Not surprisingly, another Italian group correlated the findings in bone marrow on Day 15 with cytogenetics and reported that >90% of patients who had the favorable karyotype had <10% bone marrow blasts at Day 14 compared with <20% bone marrow blasts among patients who had the unfavorable karyotype. It is clear that, as new prognostic markers are being used increasingly in patients with AML, such as FLT3 or NPM1, these analyses are likely to be refined further in the future.

The issue addressed in this report has practical implications. For example, in a large prospective study of patients with AML, residual leukemia in a patient who had a normal karyotype was an indication for referral to undergo a matched unrelated donor HSCT.37 The data from our study do not support assigning such high-risk status to patients based on the Day-14 bone marrow alone. Furthermore, future studies will need to determine whether the findings in our study also are applicable to patients who have adverse molecular subtypes, such as FLT3-positive/NPM1-negative disease.9

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary