Low Leukocyte Counts with Blast Cells in Cerebrospinal Fluid of Children with Newly Diagnosed Acute Lymphoblastic Leukemia

BACKGROUND: Treatment of the central nervous system is crucial to the successful treatment of acute lymphoblastic leukemia in children. The intensity and timing of the therapy are based on the presence or predicted risk of central nervous system leukemia as assessed according to criteria that remain controversial. METHODS: The clinical importance of leukemic blast cells detected in cerebrospinal fluid at the time of diagnosis was evaluated in 351 children with acute lymphoblastic leukemia in a randomized trial of intensive chemotherapy. All patients received intrathecal chemotherapy during the first year. Patients considered to be at high risk of relapse because of their clinical and cytogenetic features also received cranial irradiation and intrathecal chemotherapy one year after remission. Patients were considered to have central nervous system leukemia at diagnosis if they had at least 5 leukocytes per microliter of cerebrospinal fluid, with leukemic blast cells apparent in cytocentrifuged preparations, or cranial-nerve palsy; they received additional intrathecal injections of chemotherapeutic agents and cranial irradiation. Patients were retrospectively classified on the basis of cerebrospinal fluid findings: 291 patients had no detectable blast cells, 42 had fewer than 5 leukocytes per microliter and blast cells, and 18 had central nervous system leukemia as defined above. The clinical characteristics and outcomes of treatment in these groups were analyzed. RESULTS: The five-year probability of survival free of relapses confined to the central nervous system in patients with detectable blast cells and fewer than 5 leukocytes per microliter of cerebrospinal fluid was lower than in patients without blast cells (mean +/- SE, 87 +/- 13 vs. 96 +/- 2 percent), but was not different from the probability in patients with central nervous system leukemia at diagnosis. All such isolated relapses of leukemia in patients with detectable blast cells occurred during the first year of treatment, before scheduled cranial irradiation. In a multivariate analysis, the presence of cerebrospinal fluid blast cells with fewer than 5 leukocytes per microliter was independently related to the risk of relapse confined to the central nervous system. CONCLUSIONS: Patients with leukemic blast cells in their cerebrospinal fluid are at increased risk for central nervous system relapse when cranial irradiation is delayed. Such patients require intensified central nervous system treatment early in the course of therapy.