Methotrexate in Relapsed Childhood Acute Myeloid Leukemia: A Single Agent Therapeutic Window Study.

Prognosis of childhood AML has improved, but still one third of patients relapse. Conventional drugs not being used in AML yet might be useful to reduce the relapse rate. Preclinical and clinical studies suggest that methotrexate may be beneficial in AML. We previously showed (Rots et al., Eur J Cancer 2001, 37: 492) that in vitro methotrexate inhibits the key enzyme thymidylate synthase at least as effectively in AML as in ALL, if the exposure time is long instead of short. Older clinical studies with methotrexate at doses up to 300 mg/m² showed responses in up to 40% of patients with AML. We therefore initiated a single agent methotrexate window study in children with relapsed AML, in the setting of the I-BFM-SG. Children and adolescents up to 21 years of age with relapsed AML with bone marrow (BM) involvement were eligible. Children with Down syndrome were not eligible. Because of the possibility of toxicity postponing reinduction chemotherapy known to be effective, and the experience with methotrexate given at 1 gr/m² over 36 hours (10% given as a loading dose) in relapsed ALL, that regimen was chosen instead of higher-dose methotrexate. Folinic acid rescue started at 48 hours after the start of the methotrexate infusion for at least 2 doses at 15 mg/m². Hyperhydration and urine alkalanisation were also required. Clinical response was determined by a day 7 BM examination and was arbitrarily defined as good in case of a relative reduction in BM leukemic blasts of at least 50%, as partial in case of a 10–50% reduction, as progressive disease in case of an increase of at least 10%, and as stable in others. In addition, peripheral blood (PB) was examined until the day of the BM aspiration. The study was designed according to the 2-stage method of Gehan. The chance to reject the drug with a good response rate of at least 30% is less than 5% in case of no good responder among the first cohort of 9 patients. Disappointingly, the first 9 patients did not show a single good responder. Moreover, 3 patients had progressive disease (all 3 early relapsed AML FAB type M0 or M7). Therefore, it was decided to close the study. At that time, 4 more patients had been enrolled. The 13 patients all had first relapsed AML, 6 early (within 1 year from initial diagnosis; 1 FAB type M0, 2 M4, 1 M5, 2 M7) and 7 late (2 FAB type M0, 1 M1, 1 M2, 1 M4, 1 M5, 1 M7). In total, 1 patient (late relapsed M7) showed a good response, 2 a partial response (both late relapsed AML, FAB type M1 and M2), and 5 progressive disease. Based on PB, responses were seen in all patients with circulating blasts. Five grade III/IV toxicities were reported in 4 patients (mucositis, transaminatitis, infection, erythema and vomiting). In conclusion, the response rate (23%, only 1 out of 13 with a good response) to single agent methotrexate at 1 gr/m² over 36 hours given once was low in this cohort of children with first relapsed AML. The relatively low and single dose may have contributed to this result. However, 5 out of 13 patients had progressive disease based on BM examination. Future studies of the AML committee of the International BFM Study Group will focus on novel agents, such as clofarabine and gemtuzumab ozogamicin. This study did prove the feasability to perform such studies in an international setting.