Induction failure in acute lymphoblastic leukemia of childhood.

BACKGROUND: Although it is widely accepted that failure to achieve complete remission (CR) portends a poor prognosis in childhood acute lymphoblastic leukemia (ALL), there is variability in the precise definition of induction failure and, to the authors' knowledge, few published data exist regarding the outcome of patients who are slow to achieve CR.
METHODS: Between 1987-1995, 774 children with ALL were treated on 2 consecutive protocols and were evaluable to assess the time required to attain CR. The authors compared presenting characteristics and outcomes of patients based on their remission status after 1 month of induction chemotherapy: CR (n = 656), protracted hypoplasia (low peripheral blood counts and/or hypocellular marrow) (n = 95), and persistent leukemia (M2 or M3 bone marrow and/or evidence of extramedullary leukemia) (n = 23). The median follow-up was 5.2 years.
RESULTS: Presenting features that predicted persistent leukemia included a leukocyte count > 100,000/mm3 and T-cell phenotype. Approximately 91% of patients with persistent leukemia and 100% with protracted hypoplasia eventually achieved CR. The 5-year event free survival (EFS) (95% confidence intervals [95% CI] in parentheses) for patients with persistent leukemia after 1 month was 16% (95% CI, 0%, 31%), which was significantly worse (P < 0.001) than that for those who achieved CR within 1 month (5-year EFS, 82%; 95% CI, 79%, 86%) and that for those with protracted hypoplasia (5-year EFS, 79%; 95% CI, 70%, 87%). For patients with persistent leukemia, there was no significant difference in survival based on bone marrow status (M2 or M3) after 1 month or on the number of induction cycles received before achieving CR.
CONCLUSIONS: Patients with persistent leukemia at the end of 1 month of therapy have a dismal prognosis, regardless of when they subsequently achieve CR. More intensive and/or novel therapies should be considered for this subset of patients.