- © 2003 by American Society of Clinical Oncology
Mitoxantrone and Cytarabine Induction, High-Dose Cytarabine, and Etoposide Intensification for Pediatric Patients With Relapsed or Refractory Acute Myeloid Leukemia: Children’s Cancer Group Study 2951
- Robert J. Wells,
- Mary T. Adams,
- Todd A. Alonzo,
- Robert J. Arceci,
- Jonathan Buckley,
- Allen B. Buxton,
- Kathryn Dusenbery,
- Alan Gamis,
- Margaret Masterson,
- Terry Vik,
- Phyllis Warkentin and
- James A. Whitlock
- From the M.D. Anderson Cancer Center, Houston, TX; LifeSource Upper Midwest Organ Procurement Organization, Inc, St Paul; University of Minnesota, Minneapolis, MN; University of Southern California Keck School of Medicine, Los Angeles; Children’s Oncology Group, Arcadia, CA; Johns Hopkins Hospital, Baltimore, MD; Children’s Mercy Hospital and Clinics, Kansas City, MO; Cancer Institute of New Jersey, New Brunswick, NJ; Indiana University, Riley Children’s Hospital, Indianapolis, IN; University of Nebraska Medical Center, Omaha, NE; and Pediatric Hematology/Oncology, Vanderbilt University Medical Center, Nashville, TN.
- Address reprint requests to Robert J. Wells, MD, Children’s Oncology Group, 440 E Huntington Dr, Suite 300, PO Box 60012, Arcadia, CA 91066-6012; email: rjwells{at}mdanderson.org.
Abstract
Purpose: To evaluate the response rate, survival, and toxicity of mitoxantrone and cytarabine induction, high-dose cytarabine and etoposide intensification, and further consolidation/maintenance therapies, including bone marrow transplantation, in children with relapsed, refractory, or secondary acute myeloid leukemia (AML). To evaluate response to 2-chlorodeoxyadenosine (2-CDA) and etoposide (VP-16) in patients who did not respond to mitoxantrone and cytarabine.
Patients and Methods: Patients with relapsed/refractory AML (n = 101) and secondary AML (n = 13) were entered.
Results: Mitoxantrone and cytarabine induction achieved a remission rate of 76% for relapsed/refractory patients and 77% for patients with secondary AML, with a 3% induction mortality rate. Cytarabine and etoposide intensification exceeded the acceptable toxic death rate of 10%. The response rate of 2-CDA/VP-16 was 8%. Two-year overall survival was estimated at 24% and was better than historical control data. Patients with secondary AML had similar outcomes to relapsed or refractory patients. Initial remission longer than 1 year was the most important prognostic factor for patients with primary AML (2-year survival rate, 75%), whereas for patients with primary AML, with less than 12 months of initial remission, survival was 13% and was similar to that of refractory patients (6%).
Conclusion: Mitoxantrone and cytarabine induction is effective with reasonable toxicity in patients with relapsed/refractory or secondary AML. The cytarabine and etoposide intensification regimen should be abandoned because of toxicity. Patients with relapsed AML with initial remissions longer than 1 year have a relatively good prognosis.
OUTCOMES OF pediatric acute myeloid leukemia (AML) have improved gradually over the last 20 or more years. Approximately 75% to 85% of patients now have initial complete remissions, and 50% to 70% of those patients will survive for longer than 3 years.1–,6 Prognosis remains bleak for patients who do not achieve complete remissions with first therapy or who experience relapse after achieving remissions. Review of outcome data from Children’s Cancer Group (CCG) studies 251, 213, and 2861, conducted from 1979 to 1989, indicate that only 12% (SE, ± 2%) of patients were alive 3 years after relapse or inability to achieve remission when given the initial therapy (J.W. Lee, personal communication, November 1995).7–,9 In the more recent study CCG-2891 (1989 to 1995), 3-year survival was slightly better at 17% (SE, ± 2%; J.W. Lee, personal communication, November 1999).
The outcome data from CCG studies 251, 213, 2861, and 2891 have limited usefulness because only survival outcome is available. Little information is available about therapies associated with survival or treatment failure. Most reports in the literature that address patients with relapsed or refractory pediatric AML are essentially phase II trials of induction therapies or transplantation regimens for those who have attained second complete responses.10–,13
This Children’s Cancer Group study, CCG-2951, was designed to test the efficacy of a promising induction therapy that used mitoxantrone and cytarabine. It was combined with very high–dose cytarabine and etoposide intensification. This combination was being considered for front-line pediatric AML therapy.14,15 Patients who did not have complete responses to mitoxantrone cytarabine induction were given an alternate induction of 2-chlorodeoxyadenosine (2-CDA) and etoposide. We believed that most patients who had remissions would proceed to some form of bone marrow or stem-cell transplantation. It was the hope of the study committee that this would lead to an overall improvement in survival for this group of patients compared with historic control data. CCG-2951 included a separate cohort of patients with newly diagnosed, untreated (for AML) patients with secondary AML. This report describes outcomes for the induction and intensification portions of CCG-2951 and some preliminary follow-up information about the outcome of patients with relapsed or refractory AML.
PATIENTS AND METHODS
CCG-2951 opened on August 18, 1997. The study was suspended to new patient entry on January 26, 2000, and later closed (May 11, 2001) without additional patients entered. Patients with AML who did not have complete remissions with their first therapy (refractory) or who had relapses after initial remission were eligible for this study, provided that they received no additional therapy for AML. Patients also had to be younger than 22 years at enrollment and were required to have normal or near normal renal, hepatic, and cardiac function. Patients or their parents or guardians were required to sign an informed consent document, and the treating institution’s institutional review board must have approved this protocol before enrollment. Patients also were enrolled in this protocol who had AML as a second malignant neoplasm, provided no prior therapy for AML had been given. Patients with Fanconi’s anemia were specifically excluded from participation.
Therapy
Induction therapy.
Patients received cytarabine 1.0 g/m2 as a 2-hour infusion every 12 hours for 4 days (eight doses) starting at hour 0, day 0. Mitoxantrone 12 mg/m2 was given as a 1-hour infusion every 24 hours for 4 days beginning on day 2, hour 11. Cytarabine was given intrathecally before we started systemic therapy based on an age-adjusted regimen.8 Granulocyte colony-stimulating factor 5 μg/kg was given starting the day after the last dose of mitoxantrone.
Salvage induction therapy.
Patients who had more than 15% blasts in their bone marrow at the end of induction were considered to have failed to achieve remission. These patients were given salvage induction therapy comprising 2-CDA 9 mg/m2 as a 2-hour infusion for 5 days and etoposide 150 mg/m2 as another 2-hour infusion for 5 days each. Therapy was to be repeated once if patients did not achieve bone marrow remission after the first course. Intrathecal cytarabine was given once per course of therapy.
Intensification.
Patients who achieved responses (M1 or M2a marrow [< 15% blasts]) with induction or salvage induction were eligible to receive intensification therapy that consisted of cytarabine 3.0 g/m2 as a 3-hour infusion every 12 hours for 6 days (12 doses of 3 g/m2) followed by daily infusions of etoposide 400 mg/m2 as a 2-hour infusion for 3 days starting 12 hours after the last dose of cytarabine. No intrathecal therapy was given during this phase.
Continuation therapy.
After intensification, patients were treated at the discretion of responsible physicians with no further therapy, a suggested consolidation with chemotherapy (2-CDA/etoposide similar to the salvage induction), other chemotherapy, or with various forms of stem-cell or bone marrow transplantation (BMT). Investigators were responsible for reporting which option they chose and the outcome, such as death or relapse, so that survival and disease-free survival (DFS) could be calculated.
Statistics
Data from CCG-2951 through November 8, 2001, were included in this analysis. The Kaplan-Meier method was used to calculate estimates of overall survival (OS) from study enrollment and from remissions after two courses of induction therapy; event-free survival (EFS), defined as the time from study enrollment to induction failure, relapse, or death; and DFS, defined as the time from remission induction to relapse in marrow or death.16,17
Patients lost to follow-up were censored at their last known points of study, with a cutoff of May 8, 2001. Confidence intervals were calculated using Greenwood’s formula.18 Differences in OS, EFS, and DFS were tested for significance using the log-rank statistic.19
Univariate analyses of potential prognostic factors were used to compare induction outcomes using Fisher’s exact test. Continuous variables were considered using defined discrete categories. Factors univariately significant at P < .05 were considered for inclusion in multivariate models. Multivariate models were constructed for induction outcome and survival using stepwise logistic regression and Cox regression,20 respectively. Forward and backward elimination procedures were used. The likelihood ratio test was used to determine whether variables should be added or dropped from the multivariate model. Exact logistic regression was used with sparse data.21 Multivariate analyses for complete response (CR) and OS included patients with complete covariate data.
RESULTS
In the approximately 2.5 years that this study was open for accrual, 116 patients were entered, 101 with primary AML (ie, refractory or first-relapse AML), 13 with secondary AML, one incorrect diagnosis, and one enrolled when the study was suspended. The median follow-up on 38 patients alive at last contact was 16.3 months (range, 0.2 to 36.8 months). Patient outcomes for the phases of therapy are described in Figures 1⇓ and 2⇓ for patients with primary and secondary AML, respectively.
Patient characteristics are listed in Table 1⇓. The median age of the 101 patients with primary AML was 10.2 years (range, 0.7 to 19.9 years). The most prevalent French-American-British (FAB) classification of acute leukemias was M2 (25.9%), followed by M4 (18.8%). Sixteen percent of patients had refractory AML, 27% had prior remissions of less than 7 months, 35% had prior remissions of 7 to 12 months, and 23% had prior remissions greater than 12 months. Prior malignancies for 13 patients with secondary AML included Ewing’s sarcoma (two patients), acute lymphoblastic leukemia (ALL; three patients), non-Hodgkin’s lymphoma (NHL; two patients), glioblastoma (one patient), osteosarcoma (three patients), medulloblastoma (one patient), and ependymoma (one patient). Their median age at study enrollment was 14.4 years (range, 6.6 to 18.7 years). Most patients with secondary AML were classified as having FAB M0 disease (50%), whereas 25% had FAB M4 disease. The four patients with secondary AML who were alive at last contact had a median follow-up time of 16.6 months. The lengths of follow-up for these patients were 1.8, 10.4, 22.7, and 34.4 months.
Induction Outcome
Among 101 primary patients entered, one withdrew before completing induction therapy. Two more could not be evaluated for response because responsible investigators did not collect end-of-induction bone marrow samples. Ninety-eight patients were evaluated for induction outcome. Seventy-two patients had M1 (< 5% blasts) bone marrow and three patients had M2a (6% to 15% blasts) bone marrow at the end of induction, with slightly hypocellular to cellular biopsy results, for a response rate of 77% (95% confidence interval [CI], 67% to 85%). Among those, 40 with M1 and 0 with M2a also had full recovery of peripheral-blood counts, with absolute neutrophils greater than 1,000/μL and platelets greater than 100,000/μL before they were given the next course of therapy. Five patients had partial responses (M2b bone marrow [16% to 25% blasts]), three patients died during induction therapy, and 15 patients survived induction therapy but did not respond (> 25% blasts in bone marrow at the end of induction). The deaths were on days 5, 11, and 12 of induction. Two of the three deaths were related to infectious complications (bacterial sepsis) and the other was from disease progression.
Among patients with secondary AML, 10 patients (77%; 95% CI, 46% to 95%) had M1 marrow responses, one patient had a partial response (M2b bone marrow), one patient died during induction, and one patient survived induction but did not respond. The induction death was caused by disseminated fungemia (aspergillosis) at day 57 of induction. That patient’s bone marrow was severely hypocellular after induction therapy and never showed evidence of recovery.
Salvage Induction Outcome
Fourteen of 20 potentially eligible patients with primary AML received treatment with at least one or two courses of 2-CDA and etoposide. One patient (who entered salvage induction with M2b marrow) achieved an M1 marrow response after one course of treatment, but died 20 days later from toxicity. One patient withdrew and was not evaluated, 11 patients did not respond to therapy, and one patient did not have sufficient data for response evaluation. Therefore, among 12 patients with primary AML with outcome information, one patient responded (8%; 95% CI, 0% to 38%).
Two patients with secondary AML also were treated with salvage induction. They did not respond to the therapy.
Intensification
Fifty-eight of 75 patients who were eligible for intensification (M1 or M2a marrow at the end of induction) actually received it. Twelve patients who completed induction were next treated with BMT. One patient went from induction to chemotherapy consolidation, and four patients had no postinduction therapy information available.
Among 58 patients who received intensification, 42 patients (72%; 95% CI, 59% to 83%) completed intensification in remission (M1 or M2a marrow), 10 patients (17%; 95% CI, 9% to 29%) died, generally from infectious complications, and four patients relapsed by the end of intensification. The number of deaths during intensification exceeded study toxicity limits designed to assure that the number of toxic deaths during that phase did not exceed approximately 10% mortality, which was normal with high-dose cytarabine and asparaginase consolidation in several CCG AML therapies.1,8 The toxicity led to suspension of new patient accrual and then study closure. Four patients (three patients with primary AML and one patient with secondary AML) were entered on the induction phase at the time of suspension, so intensification was omitted from treatment of those four patients when they attained CRs at the end of induction.
Six patients with secondary AML also received intensification therapy. Three patients had CRs at the end of induction, one patient died of sepsis, one patient was inaccessible, and one patient had no data available. Among the four remaining patients with secondary AML with CRs at the end of induction, three patients received BMT. The reason the other patient did not receive intensification is not known.
Follow-On Therapy
Thirty-five patients with primary AML underwent BMT from a variety of donor sources and preparative regimens after achieving CRs during CCG-2951 therapy (12 patients after induction and 23 patients after intensification). Most transplantation procedures were performed with unrelated donors. Thirty-five patients (100%) with primary AML (30 unrelated donors, four related, and one unknown donor) and three (75%) of four patients with secondary AML (one unrelated cord blood, one sibling, and one unknown donor) received allogeneic BMT. The remaining patient with secondary AML received an autologous BMT. Two more patients underwent transplantation without obtaining CRs. Thirteen patients with primary AML received further chemotherapy after achieving CRs with induction; all but one received intensification before further chemotherapy.
Seventeen of 35 patients who underwent BMT were alive with a median follow-up of 17.6 months (range, 0.3 to 36.8 months), and 18 patients died (five from progressive or recurrent leukemia, five from infection, one from hemorrhage, two from toxicity, four from graft-versus-host disease, and one from other cause), with median survival of 8.0 months (range, 3.4 to 18.4 months). The two patients who underwent transplantation without achieving CRs died (one from progressive disease 8 months after transplantation and one from veno-occlusive disease 2 months after transplantation).
The status of 13 patients with primary AML who received chemotherapy is that eight were alive with median follow-up of 15.5 months (range, 3.5 to 34.8 months) and five died (three from progressive leukemia and two from infections) after a median of 14.5 months (range, 4.6 to 16.5 months).
For the secondary AML group, three patients ended intensification with M1 or M2a bone marrow. One patient underwent BMT, another patient had chemotherapy, and the follow-on treatment of the third patient is unknown. They died at 5.7 (toxicity), 13.1 (progressive disease), and 5.8 (progressive disease) months from start of therapy, respectively.
Among four patients with secondary AML who did not receive intensification, three underwent BMT and were alive at 10.5, 22.7, and 34.4 months from start of therapy. The remaining patient died of progressive leukemia at 2.4 months.
Survival
OS from the start of therapy was similar between patients with primary and secondary AML (P = .788). Corresponding estimates at 2 years were 24% (95% CI, 15% to 34%) and 21% (95% CI, 4% to 48%), respectively. Primary AML patients with greater than 1-year initial remission had 75% OS at 2 years (95% CI, 46% to 90%). For patients with less than 12 months of initial remission, survival was 13% (95% CI, 5% to 26%), similar to that of refractory patients (6%; 95% CI, 0% to 25%; P = .423; Fig 3⇓).
DFS was similar to OS. For patients with primary AML, survival was 31% (95% CI, 20% to 44%) at 2 years from the end of induction and 45% (95% CI, 27% to 61%) at 2 years from the end of intensification (Table 2⇓). Survival from the end of induction for patients who received intensification (33%; 95% CI, 20% to 47%; n = 58) was similar to that of those who did not receive intensification according to the protocol or investigator choice (31% [95% CI, 10% to 55%] at 2 years; P = .676; n = 17). DFS also was similar (P = .890).
OS from start of therapy for the 39 patients with primary AML with M1 marrow who had end-of-induction BMTs with blood count recoveries (absolute neutrophil count [ANC] >1,000/μL and platelets >100,000/μL) was similar to OS for the 32 patients with primary AML with M1 marrow who had end-of-induction BMTs without full blood count recoveries before the start of the next course of therapy (P = .692).
OS (Fig 4⇓) and DFS from the end of intensification were similar between follow-on BMT (n = 23) and chemotherapy (n = 12) patients (40% [95% CI, 18% to 61%] v 50% [95% CI, 15% to 77%], P = .384%; and 43% [95% CI, 22% to 63%] v 51% [95% CI, 16% to 78%], P = .475, respectively). Patients who received follow-on BMT included those who received intensification (n = 23) and some who did not (n = 12). From the end of induction, 2-year survival estimates were as follows: OS, 40% versus 44% (P = .707) and DFS, 43% versus 44% (P = .769). Those outcomes indicated that intensification as given in this study did not improve outcomes of subsequent BMTs.
Toxicity
The primary toxicity during induction, salvage induction, and intensification in this study was bone marrow suppression caused by chemotherapy or leukemia, indicated by prevalence of ANC (weeks < 500/μL) and platelet (weeks < 105/μL) toxicities during those treatments. For induction, ANC greater than 500/μL was achieved by 50% of patients by day 21 and by 91% of patients by day 35. For intensification, ANC greater than 500/μL was achieved by 38% of patients by day 21 and by 81% by day 35. Platelet recovery was 60% recovered to greater than 50,000/μL by day 28 and 81% by day 42 for induction and 48% and 78%, respectively, for intensification. During induction and intensification, 54% and 78% of patients, respectively, had at least one infection. Gastrointestinal (23%) and hepatic (15%) toxicities were the next most common. Fifty percent of patients reported at least grade 3 toxicities during induction and 81% reported at least grade 3 toxicities during intensification. Gastrointestinal toxicities and hepatic toxicities were most common.
Prognostic Factors
Several potential prognostic factors were analyzed to determine whether any could be correlated with induction remission rates and survival from beginning of therapy. Among those were sex, age, WBC count at presentation, platelet count, race, splenomegaly, hepatomegaly, FAB classification, length of initial remission (0, < 7, 7 to 12, and > 12 months), and cytogenetics. On univariate analysis, absence of splenomegaly (80.2% v 50.0%; P = .031), FAB M1 classification (100% v 72.0%; P = .032), and remission length of more than 12 months (100% v 69.7%; P = .001) were correlated with improved remission rates. Initial remission length of zero (50% v 81.7%; P = .020) or less than 7 months (57.7% v 83.3%; P = .015) was correlated with lower initial remission rates. For OS, only initial remissions greater than 12 months (P < .001) were correlated with improved survival; no prior remissions (P = .030) and remissions less than 7 months (P = .021) were correlated with inferior survival. Using a multivariate analysis, only the initial remission length and FAB M1 were correlated with response rate, and only initial remission length had a significant association with survival.
Prior Studies
Figure 5⇓ compares survival experience on this induction to that of relapsed or refractory patients on some prior CCG studies. Survival on CCG-2951 was better than that of CCG-251 (P < .001) and the aggregate of CCG-251, CCG-213, CCG-2861, and CCG-2891 (25% v 14% at 3 years; P = .011).5–,9
DISCUSSION
This study provided mixed results overall. Patients seemed to have slightly better outcomes than our historical control population. The activity of mitoxantrone and cytarabine induction was confirmed clearly. That induction was able to induce a remission in all patients whose initial remissions were longer than 12 months (Table 4). It also was able to induce remissions in 57% of patients who did not have remissions after initial treatment with anthracycline, standard-dose cytarabine, and etoposide regimens and who were clinically resistant to those drugs. The mitoxantrone and cytarabine regimen also induced responses in 77% of patients with secondary AML, a traditionally difficult group to treat. For example, in our last study (CCG-2891), only 12 of 24 patients who had secondary AML achieved remissions.22 The toxicity-related death rate during induction using this therapy was quite low (3%) compared with that of most AML induction therapies.
This study was not successful in combining induction with what we hoped would be a better intensification. The high-dose cytarabine and etoposide regimen proved to be too toxic and forced closure of the study. The high relapse rate in patients who survived the treatment indicates that even highly toxic high-dose chemotherapy with standard agents is not effective in producing long-lasting remissions in this group of patients. Our prejudice that some form of bone marrow transplantation could produce a high survival rate when used for patients in second remissions also was not supported by our results. Just under 50% of those patients who underwent transplantation survive, and as a group, patients who underwent transplantation did no better than those who received follow-on chemotherapy. However, follow-up is relatively short, and the number of patients who received chemotherapy in this study is quite small. The salvage induction of 2-CDA and etoposide was disappointing; however, the patient population treated is a very refractory group likely to be clinically resistant to most conventional therapy. The OS rate remained unsatisfactory.
This study confirmed observations of many others that the most important prognostic factor for patients with AML who experience relapse is the length of initial remission. It also found that many patients who have initial remissions of more than 12 months can be successfully treated with salvage therapy and have a reasonable chance for long-term survival.14,23,24 No other prognostic factors seemed to have much significance, although the FAB M1 group identified here may be useful in future studies.
We confirmed activity and acceptable toxicity of mitoxantrone and cytarabine induction and recommend that therapy be incorporated into initial treatment regimens for AML. Preliminary data from this study on patients with secondary AML were also somewhat encouraging and should be extended. The intensification regimen should be abandoned, as should the salvage induction. We did not find benefit of allogeneic bone marrow transplantation over postremission chemotherapy in patients who had AML in second remissions after mitoxantrone and cytarabine reinduction therapy.
APPENDIX
Children’s Cancer Group Participating Institutions
CCG-2951 Enrollment Characteristics for Patients With Primary and Secondary Acute Myeloid Leukemia
CCG-2951 Survival Estimates (± 2* [SE]) for Patients With Primary Acute Myeloid Leukemia
CCG-2951. Flow diagram for patients with primary acute myeloid leukemia. CR, complete response; PR, partial response; BMT, bone marrow transplantation.
CCG-2951. Flow diagram for patients with secondary acute myeloid leukemia. CR, complete response; PR, partial response; PD, progressive disease; BMT, bone marrow transplantation.
CCG-2951. Overall survival from start of therapy for patients with primary acute myeloid leukemia.
CCG-2951: Bone marrow transplantation (BMT; n = 23) versus chemotherapy (n = 12) follow-on therapy for survival.
Overall survival for patients with relapsed or refractory acute myeloid leukemia in Children’s Cancer Group (CCG) studies 251 (1978 to 1983, n = 292), 213 (1985 to 1989, n = 290), 2861 (1988 to 1989, n = 63), and 2891 (1989 to 1995, n = 397) with CCG-2951.
- Received June 21, 2002.
- Accepted May 15, 2003.
