What Is the Optimal Therapy for Childhood AML?

In the past 3 decades, we have witnessed significant improvement in the outcomes of pediatric patients with acute myelogenous leukemia (AML). While it is clear that progress has been made in supportive care (eg, aggressive use of antibiotics for febrile neutropenia, improvements in blood product support, and higher quality parenteral nutrition), advances in the treatment of AML surely have contributed dramatically to this success. The major cooperative groups in the United States have taken different approaches to AML therapy, and both have yielded important insights. European groups have seen similar improvements in their treatment of AML in children.

The purpose of this article is to summarize not only the current status of AML therapy for children, but also to review the studies that led to the current treatment regimens, and to suggest possible future directions for the treatment of AML in children and adolescents.

The Past

The most important lesson from the recent past is that, in contrast to the prolonged, low-intensity regimens used for pediatric acute lymphoblastic leukemia (ALL), high-intensity therapy is critical to treating AML successfully. This was clearly demonstrated by Children’s Cancer Group (CCG) study 213, which compared intensive postremission therapy with a more prolonged maintenance treatment.[1] Children who achieved a remission and were treated with an intensive consolidation regimen containing cytarabine(Drug information on cytarabine) (Ara-C) and l-asparaginase (Elspar) had a 68% 5-year survival rate (from the end of consolidation), compared with 44% for those children treated with less intensive consolidation followed by maintenance therapy.

The major American cooperative groups have pursued intensification along parallel paths. While the CCG has intensified the timing of successive courses of induction therapy, the Pediatric Oncology Group (POG) has intensified drug doses. Each of these approaches has led to increasing rates of remission induction and event-free survival in successive clinical trials. Other attempts at dose intensification include trials from St. Jude Children’s Research Hospital aimed at maintaining specific target plasma concentrations of cytarabine and etoposide(Drug information on etoposide),[2], and the utilization of high-dose cytarabine regimens at the Dana-Farber Cancer Institute.[3] The efforts of the cooperative groups will be discussed in more depth below.

Pediatric Oncology Group

The general approach taken by POG since the early 1980s has been to intensify therapy through the use of higher drug doses, as exemplified by the use of high-dose Ara-C (HDAC). Since cytarabine has a very short half-life and is only active against cells in the S phase of the cell cycle, its efficacy is very dose- and schedule-dependent. It is most commonly administered as a prolonged continuous infusion, in an attempt to maximize exposure to cells in cycle. The sensitivity of leukemic blasts to cytarabine is in part dependent on intracellular concentration of the drug. Impaired drug import is a major mechanism of resistance, and this can be overcome by using doses at least 1 log greater than conventional dosing.[4]

Another important contributor to intracellular drug concentration is cytidine deaminase activity. Cytidine deaminase is the major degradative enzyme of cytarabine, but prolonged infusions, rather than high doses, are thought to overcome this activity, because enzymatic activity is not saturated at the dose ranges used.[5]

POG 8498

In 1991, POG published the results of study 8498, which evaluated the efficacy of HDAC in the treatment of pediatric AML.[6] When originally opened, patients enrolled in the study were treated with two cycles of DAT (daunorubicin for 3 days, cytarabine as a 7-day infusion, and thioguanine for 7 days). For patients in remission, this was followed by (1) four doses of HDAC followed by asparaginase(Drug information on asparaginase), (2) etoposide with azacytidine, (3) prednisone(Drug information on prednisone), vincristine, methotrexate, and mercaptopurine(Drug information on mercaptopurine) (Purinethol), and (4) a 5-day continuous infusion of cytarabine. After 2 years, the protocol was amended to substitute six doses of HDAC for the second DAT induction course and six doses of HDAC instead of the four-dose HDAC/asparaginase course.

The remission induction rate was 85% in both groups of patients, and there was no significant difference in induction deaths or deaths in remission. At the time of publication, patients treated with HDAC during induction had a higher 3-year event-free survival (34% vs 29%) and disease-free survival (42% vs 34%), although these numbers did not reach statistical significance.

POG 8821

The trend toward improved remission induction and long-term survival in patients treated with HDAC led to the incorporation of this agent into subsequent POG AML trials. The results of POG 8821 were published in 1996.[7] The remission induction regimen in this trial was identical to that used for the second half of POG 8498 (ie, DAT followed by HDAC). Patients who achieved a remission and had an HLA-identical sibling were treated with allogeneic bone marrow transplantation. Other patients were randomized to receive either intensive consolidation chemotherapy or autologous bone marrow transplantation.

The remission induction rate in this study was 85%. The 3-year event-free survival rate was 34%, and overall survival was 42%. There was no significant difference in outcome between patients treated with chemotherapy and those treated with autologous transplantation, with both groups achieving a 3-year event-free survival of approximately 37%. In contrast, the event-free survival for patients treated with an allogeneic transplant was 52% at 3 years—a significant improvement.

Thus, children treated with autologous transplantation did no better than those treated with chemotherapy alone, while survival was superior for children treated with an allogeneic transplant. Interpretation of this study, however, is complicated by the poor compliance rate. Only 62% of the children who achieved a remission were eligible for randomization, and 57% of the children who were not eligible dropped out for reasons other than lack of a suitable donor. Additionally, of those children eligible for randomization, 32% were not randomized because of lack of parental consent.

Children’s Cancer Group

The CCG has pursued the use of "timed sequential therapy" in pediatric AML.[8] This concept arose from the work of Burke and colleagues at the Johns Hopkins Hospital, beginning in the 1970s. Their initial observation was that most AML has a low growth fraction at the time of diagnosis,[9] and that there is a factor in the serum of patients with AML that suppresses the proliferation of hematopoietic cells in vitro.[10] After the administration of a single cycle of chemotherapy, this activity disappears and is replaced by a factor that stimulates proliferation.[10] They went on to show that the administration of a second cycle of chemotherapy timed to coincide with the peak of this proliferative activity in patients’ serum significantly increased the rate of remission induction, and they called this approach timed sequential therapy.[11]

CCG 213P

The first CCG study addressing the importance of timing successive cycles of chemotherapy was CCG 213P.[1] This study compared the efficacy of the Capizzi regimen (HDAC followed by asparaginase) given at 28-day intervals with that of the same regimen given at 7-day intervals to children who were just recovering from remission induction therapy. From the end of induction, children treated with two courses 7 days apart had a 5-year overall survival rate of 58%, compared with 41% for the children treated with a 28-day interval. This study clearly confirmed the efficacy of intensively timed therapy, at least as a component of postremission AML treatment.

CCG 2861

The feasibility of timed sequential remission induction therapy was first tested by the CCG in a pilot study, CCG 2861.[12] Children were treated with a five-drug induction regimen—DCTER (dexamethasone, cytarabine, thioguanine, etoposide, daunorubicin(Drug information on daunorubicin) [Rubomycin])—over 4 days, with a second, identical cycle given after 6 days of rest. The second cycle was given on day 10, regardless of hematologic status at the time.

Seventy-six percent of the children enrolled in this study achieved a remission, with half of the remainder dying of progressive disease and half dying of regimen-related toxicity. For all patients, the 3-year event-free survival rate from the time of diagnosis was 37%, which was comparable to the 5-year survival rate of 36% seen in CCG 213,[13] a concurrently running trial that compared the efficacy of DCTER with that of 7 days of cytarabine and 3 days of daunorubicin (7+3).

CCG 2891

Having demonstrated that it is feasible to treat children with intensively timed cycles of DCTER, the CCG went on to compare this approach with the same regimen delivered with standard timing, in trial 2891.[14] Children with newly diagnosed AML were randomized to receive DCTER followed by a second cycle of DCTER given either after blood count recovery or after 6 days’ rest, regardless of hematologic status.

When rates of remission induction were compared, there was no difference between the two groups. Seventy-five percent of patients treated with intensive timing achieved a remission, compared with 70% of those treated with standard timing. This difference was not statistically significant. In contrast, there was a significant difference in the 3-year event-free survival rate between the two groups—42% for the intensive timing group, compared with 27% for the standard timing group.

These results were recently updated, and the 8-year overall survival rate for all patients who achieved a remission is 54%.[15] Even at this long follow-up, the superiority of intensive timing is preserved, with significantly better survival for patients treated with intensively timed therapy (49% surviving from initial diagnosis, compared with 34% survival for patients treated with standard timing), regardless of their postinduction treatment.

Interestingly, the 5-year event-free survival rate for children enrolled in CCG 213,[1] which compared DCTER with 7+3, was 31%—equivalent to the 27% 3-year event-free survival rate seen in CCG 2891.[14] This raises the possibility that the results of CCG 2891 might be improved further with the use of more effective chemotherapy. Nevertheless, this trial conclusively demonstrated the superiority of intensive timed sequential therapy.