Questions regarding frontline therapy of acute myeloid leukemia


  • See referenced original article on pages 5012-21, this issue.


The results from multiple investigations using different induction approaches suggest that better regimens are available and should be investigated further in the context of acute myeloid leukemia (AML) heterogeneity and emerging prognostic knowledge concerning cytogenetic and molecular abnormalities. Today, based on existing data, evaluating bone marrow on Days 10 to 14 for residual AML and reacting to the bone marrow findings may be a common practice, but is not validated by objective data.

In this issue of Cancer, a review of 6 studies by the Eastern Cooperative Oncology Group (ECOG) involving 1980 patients with acute myeloid leukemia (AML) by Rowe et al focuses on 1272 patients (64%) who achieved a complete response (CR) either after 1 course (n = 945) or after 2 courses (n = 327) of chemotherapy.1 The authors' objective was to define the prognosis of such patients, because the common notion is that patients with AML who require more than 1 course of induction chemotherapy have a worse outcome. They demonstrated that patients who received a second course of induction had an outcome similar to that of patients who required only 1 course to achieve CR. They emphasized that “the results of this study in no way suggest a recommendation for patients to receive a second cycle on Day 14,” but that their study negates the common perception that patients who require 2 courses of induction to achieve a CR (the second course delivered because of residual AML on a bone marrow analysis on Day 10-14 of Course 1) have a worse prognosis.

These results raise important questions related to frontline AML therapy: 1) What is the prognosis for unselected patients with AML who receive standard frontline AML therapy? 2) What is the significance of “residual leukemia” documented after 1 course of induction; and should such bone marrow analyses be performed early (Day 10-14) and be followed with a second course of induction if “residual leukemia” is noted? 3) Is daunorubicin at a dose of 45 to 60 mg/m2 daily for 3 days and cytarabine 100 to 200 mg/m2 daily for 7 days (“3 + 7 regimen”) still considered standard frontline AML therapy?

Prognosis of Patients With AML who Receive Frontline Therapy

The CR rate of 64% and the relatively low survival rates in the 6 ECOG studies probably are because the analyses included subsets of older patients (E1490, E3993, and E2491) (see Tables 3 and 4 in the report by Rowe et al).1 In recent reports of younger patients (aged ≤60 years) with AML who were eligible for cooperative group trials (reasonable performance status, no organ dysfunctions), the CR rate with “3 + 7 regimens” was 57%, the median survival was 16 months, and the estimated 5-year survival rate was 20% to 25%.2 The low 5-year survival rate is concerning, because previous reports by Cancer and Leukemia Group B (CALGB) and others indicated that the 5-year survival rate ranged between 35% and 50%.3-5 This may be attributed to the inclusion of patients who had antecedent hematologic disorders and multilineage dysplasia (patients with poor performance status, organ dysfunctions, secondary AML, and comorbid conditions that likely were excluded but were not specified in the report)2 but also may have been caused by differences in consolidation-maintenance approaches. In the CALGB study reported by Mayer et al,3 in addition to 4 consolidation courses of standard versus high-dose cytarabine, patients received 4 additional monthly maintenance courses with daunorubicin and cytarabine (total, 8 consolidation-maintenance courses), which leukemia experts ignored and deleted from subsequent study designs/protocols.

Among older patients (aged ≥60 years) who receive “3 + 7 regimens,” the CR rate was 54%, the median survival was 8 to 9 months, and the 5-year survival rate was 10%.6 These outcomes are similar those reported from other studies, which also included selective subpopulations (multiple exclusions for poor performance status, organ dysfunctions, and other comorbidities).7, 8 In the Swedish Acute Leukemia Registry, among patients aged ≥65 years, 23% to 67% received intensive chemotherapy, the median survival ranged from 2 to 9 months, and the 5-year survival rate ranged from 0% to 13%. Among patients who were able to receive intensive chemotherapy, the 5-year survival rate ranged from 0% to 18%. The situation was worse when all patients with AML were included, as attested by the observation from the Surveillance, Epidemiology, and End Results (SEER) database that only 33% of older patients with AML were offered chemotherapy.9 In the SEER data, the median survival of patients aged >65 years was 2.4 months, but it was 6.7 months if they received chemotherapy.9 Among older patients, the 8-week mortality rate with intensive chemotherapy ranged from 20% to 30%, and it increased up to 30% to 60% in patients who had poor performance status and/or organ dysfunctions.10

To put things into perspective, Figure 1 illustrates the results from frontline AML therapy according to patient age since 1980 from The University of Texas M. D. Anderson Cancer Center, including all patients who were referred (no exclusions for poor performance status, organ dysfunctions, comorbid conditions, prior myelodysplastic syndrome, or secondary AML). Figure 2 provides results over the decades in younger patients (Fig. 2, top) and older patients (Fig 2, middle). The CR rate in younger patients was 72% (Table 1), and 5-year survival rate was 38%. The results continued to improve in younger patients (Fig. 2, top), perhaps as a result of more systematic use of idarubicin and high-dose cytarabine during induction-consolidation and better supportive care and antibiotics. The results in older patients were poor. Although the CR rates were reasonable (48%) (Table 1), the median survival (range, 5-7 months) and the 5-year survival rate (range, 8%-10%) did not improve in a clinically significant manner (although there was a statistically significant difference) (Fig. 2, middle). Patients aged ≥70 years who received intensive chemotherapy since 1990 had reasonable CR rates (47%), but the median survival was only 4 to 5 months (Fig. 2, bottom). Thus, the assertion that a CR is essential to prolong the survival of patients with AML6 needs to be re-evaluated and has been challenged in the treatment of myelodysplastic syndrome. These results emphasize the need to improve on the “3 + 7 regimens” in younger patients (eg, with the addition of more chemotherapy or targeted therapies, better consolidations strategies, and more maintenance with targeted agents) and to explore alternate strategies to intensive chemotherapy in older patients and/or in those who are unable to tolerate intensive chemotherapy (with predicted high 8-week mortality rates ≥30%).10 In younger patients, chemotherapy agents added to anthracyclines plus cytarabine have demonstrated promise including chlorodeoxyadenosine, lomustine, clofarabine, and others.11, 12 Targeted therapies like FMS-like tyrosine kinase 3 (FLT3) inhibitors, particularly in the subset of patients with FLT3 mutations (30% of patients with AML in normal karyotype), are promising.13 Among older patients, low-intensity therapies with hypomethylating agents, clofarabine (adenosine nucleoside analog), sapacitabine (cytosine nucleoside analog), SNS 595 (topoisomerase II inhibitor), CPX351 (a 5:1-M liposomal preparation of cytarabine and daunorubicin), and others are under investigation.

Figure 1.

The survival of unrelated patients with acute myeloid patients who received frontline therapy (1980 to present; n = 3689) is illustrated according to patient age (<60 years vs ≥60 years).

Figure 2.

The survival of (Top) younger patients and (Middle) older patients with acute myeloid leukemia is illustrated by decade (from 1970 to present). (Bottom) The outcome of patients aged ≥70 years who received intensive chemotherapy after 1990 is illustrated.

Table 1. Incidence of Complete Response and Mortality With Frontline Intensive Chemotherapy (M. D. Anderson Cancer Center)
Age, yYearNo. of PatientsNo. With CR (%)% Mortality
  1. CR indicates complete response.

<601980-200919201389 (72)1014
≥601980-20091769842 (48)1929
≥701990-2009623291 (47)2535

Significance of Residual Leukemia Detected During Frontline Induction Chemotherapy

There is an increasing perception that residual AML after a course of induction therapy constitutes “primary refractory AML” with a poor prognosis. In fact, this subset of patients is extremely heterogeneous. Depending on the selection criteria (eg, residual leukemia on Day 10-14 vs Day 28-42; residual leukemia after high-dose cytarabine vs standard-dose cytarabine induction), the CR rates with a second induction course (or with salvage therapy) may vary from <5% up to 70%. Residual leukemia documented on an early bone marrow analysis (Day 10-14) may be transient (ie, may disappear without further therapy), persistent (but still responsive to a second course of similar therapy), or truly resistant. Residual AML with 80% to 90% blasts has a different prognostic implication than with residual AML with 6% to 10% blasts. Residual leukemia on Days 21 to 42 of induction therapy maybe more indicative of resistant leukemia than the same residual on Days 10 to 14. Finally, residual leukemia after higher intensity induction chemotherapy (eg high-dose cytarabine-based induction) may be more indicative of resistant leukemia than residual leukemia after a “3 + 7 regimen.” Many patients with residual leukemia on Day 10 to 14 achieve a CR without additional chemotherapy. In studies at The University of Texas M. D. Anderson Cancer Center,14 patients who had residual AML in Day 14 bone marrow with 5% to 59% blasts had estimated subsequent CR rates of 40% to 65% without additional chemotherapy. Those who had ≥60% blasts had lower CR rates of only 19% (6 of 31 patients). In contrast, residual leukemia in Day 21 bone marrow was associated with CR rates of 36% to 78% if the blasts were 5% to 39% but with a lower CR rate of only 11% (5 of 46 patients) if the blasts were ≥40%. This emphasizes the different prognostic effect of different levels of blast percentages and of residual blasts on Day 14 versus later bone marrow studies. An important question is whether a second course of induction therapy given on Day 14 of Course 1 for residual leukemia is beneficial to patients (or whether it may be detrimental). The authors make no claims either way, although a common belief among oncologists is that such an approach may be beneficial (and often is practiced). To prove such benefit, a study of patients with residual leukemia on Day 14 (stratified by the percentage of blasts) randomized to either a second induction course, further observation, or perhaps higher intensity/investigational strategy, is required with the endpoint of survival and disease-free survival. To our knowledge, no such study has been conducted to date in adult AML. However, strategies using “double induction” or “time-sequential induction” have not demonstrated a significant overall improvement in outcome.15, 16 In 1 study, an exploratory analysis of the value of a second double induction on Day 21 with mitoxantrone and high-dose cytarabine suggested a beneficial effect among patients who had unfavorable features (Day 16 bone marrow blasts >40%, unfavorable karyotype, high lactate dehydrogenase).15 In the second study, timed-sequential induction improved recurrence-free survival (compared with regular induction and double induction) in younger patients (aged ≤50 years) but did not improve survival. It is noteworthy that only 38 younger patients received time-sequential therapy.16

In the study by Rowe et al, the analysis focused on the 1272 patients who achieved a CR.1 The outcome of patients who did not achieve a CR or who had persistent disease on Days 10 to 14 was not addressed. Of 1980 patients who were treated (Table 3),1 945 patients achieved a CR after 1 course, leaving 1035 patients with residual leukemia, only 557 of whom received the second course (327 patients achieved a CR). An important question is what happened to the 478 patients (1035 - 557 patients) who did not receive a second cycle and the 230 patients (557 - 337 patients) who received a second course but did not achieve a CR? Did they all receive the second course for persistent disease? How many died within 30 days of the second induction? Was the second induction perhaps harmful to that subset of patients? How many survived the second induction, and what was their outcome? Knowing the fate of the patients who did not achieve a CR with the first course or with the second course is important to assess whether a second course of induction therapy for residual leukemia in Day 10 to 14 bone marrow was beneficial to the total study group.

Are “3 + 7 Regimens” Still the Standard Frontline Therapy in AML?

Recent studies have challenged the concept that “3 + 7 regimens” remain the gold standard of AML therapy. Kern and Estey conducted a meta-analysis of 3 studies in 1691 patients who were randomized to induction therapy with standard versus high-dose cytarabine. The authors concluded that the 4-year estimated rates of recurrence-free survival (P = .03), survival (P = .0005), and event-free survival (P < .0001) favored induction therapy with high-dose cytarabine.17 High-dose cytarabine induction therapy results in most CRs being achieved after 1 course, thus shortening the duration and cost of induction therapy. High-dose cytarabine induction regimens truly select patients with resistant AML who fare poorly after either second induction therapy or salvage therapy. In our experience in 1179 patients who received high-dose, cytarabine-based induction regimens, only 285 patients had persistent disease after 1 course of induction. Only 43 of those 285 patients (15%) responded to any form of salvage therapy, and only 21 patients (7%) were alive in CR for at least 6 months (including 14 patients who underwent allogeneic stem cell transplantation).18

Two recent studies evaluated cytarabine at doses of 100 to 200 mg/m2 daily for 7 days with randomization to the addition of either high-dose daunorubicin 90 mg/m2 daily for 3 days or standard dose daunorubicin 45 mg/m2 intravenously daily for 3 days in younger patients (aged <60 years; ECOG study)2 or older patients (ages 60-83 years; Europe).6 Both studies concluded that high-dose daunorubicin improved some of the results over standard-dose daunorubicin. In the ECOG study, high-dose daunorubicin was associated with a significant improvement in the CR rate (71% vs 57%; P < .001) and in survival (median 24 months vs 16 months; P = .003). The benefit was confined to patients aged <50 years and those with normal karyotype.2 In the European study, high-dose daunorubicin was associated with a significant improvement in the overall CR rate (64% vs 54%; P = .002) and in the CR rate after 1 cycle (52% vs 35%; P < .001). Survival was improved only in patients aged 60 to 65 years.

Pautas et al11 randomized 468 patients ages 50 to 70 years on Acute Leukemia French Association studies to 3 induction anthracycline regimens: idarubicin 12 mg/m2 daily for 3 days, idarubicin 12 mg/m2 daily for 4 days, and high-dose daunorubicin 80 mg/m2 daily for 3 days. The CR rate with idarubicin daily for 3 days was higher than with high-dose daunorubicin (83% vs 70%), as was the 3-year event-free survival rate (23% vs 16%).

Holowiecki et al from Poland randomized 673 patients (ages 18 to 60 years) to receive either daunorubicin plus cytarabine (DA), or DA plus fludarabine, or DA plus chlorodeoxyadenosine (DAC).19 The authors reported significantly higher CR rates after 1 course of induction and overall and reported better 3-year survival rates with the addition of the DAC regimen.19

In a retrospective study, Pigneux et al evaluated the potential benefit of adding lomustine (200 mg/m2 orally on Day 1) to standard induction chemotherapy in 847 older patients with AML. In multivariate analysis, they demonstrated that the addition of lomustine improved the CR rate and survival.12

In Medical Research Council studies, Burnett et al investigated the potential benefit of adding gemtuzumab ozogormycin (GO) 3 mg/m2 to chemotherapy during the induction course and first consolidation. They observed that the addition of GO improved outcome in selected patient subsets.20

Overall, the results from multiple investigations using different induction approaches (high-dose cytarabine induction; high-dose daunorubicin induction; idarubicin instead of daunorubicin; the addition of GO, chlorodeoxyadenosine, or lomustine) suggest that better regimens than “3 + 7 regimens” are available. These regimens should be investigated further in the context of AML heterogeneity and the emerging prognostic knowledge concerning cytogenetic and molecular abnormalities (FLT3 abnormalities, nucleophosmin [NPM1] mutations, others) and the properties of the AML stem cell.21-24 In addition, the significance of residual AML on Days 10 to 14 and the benefit of double-induction therapy need to be re-examined in the context of randomized studies evaluating these new approaches. Today, based on existing data, evaluating Day 10 to 14 bone marrow for residual AML and reacting to the bone marrow findings may be a common practice, but is not validated by objective data.


Supported by research funding from Bristol-Myers Squibb, Novartis, Genzyme, Eisai, and Celgene.