Great oaks from little acorns

Comment on Pui et al, page 2690

Prognosis in childhood ALL has been improved with early intensification of therapy; stratification of treatment intensity according to risk of relapse can preserve this efficacy while limiting unnecessary toxicities.

Prognosis in any illness depends upon interactions between the patient, the disease, and the treatment. In the past 4 decades, event-free survival of children with acute lymphoblastic leukemia (ALL) has increased from less than 5% to almost 80%. Key clinical contributors to this progress have included new drugs, better use of old drugs, central nervous system prophylaxis, and an early intensive postremission phase of therapy.FIG1 

Improved outcomes with intensified therapy have been accompanied by a higher incidence of unwanted side effects. It would be ideal to individualize the intensity of treatment according to a patient's risk of relapse. In such a scenario, patients with a lower risk of relapse would receive “minimalized” therapy, whereas patients at a higher risk of treatment failure would receive more intensified therapy. Key clinical predictors of relapse include age, white blood cell (WBC) count, and the presence of extramedullary disease at diagnosis. Key laboratory predictors of outcomes include immunophenotype, cytogenetics, pharmacokinetics, pharmacodynamics, pharmacogenetics, rate of response (minimal residual disease), and gene expression profiles.

In this issue, Pui and colleagues report the results of St Jude study XIIIB. Using a revised classification system, 53% of patients were classified as lower risk compared with 12% in the previous St Jude study. Other changes from previous studies included (1) using cranial irradiation only for patients with central nervous system (CNS) leukemia at diagnosis or T-ALL and a WBC count lower than 100 × 10⁹/L (100 000/μL), (2) replacement of prednisone with dexamethasone after induction, (3) earlier use of reinduction, and (4) elimination of asparaginase from etoposide cycles in higher-risk patients. Five-year event-free survival was 81% (89% for lower risk, 74% for higher risk). Isolated CNS relapse occurred in 1.7% of patients and combined CNS/marrow relapse in 3% of patients.

How should we interpret the data from this relatively small nonrandomized study? Are all of the components used necessary to achieve the excellent results seen in the lower-risk cohort? Does early intensive CNS prophylaxis really obviate the prognostic significance of CNS-2 or initial traumatic lumbar puncture with blasts status? Is repeated etoposide use in higher-risk patients worth the risk of developing secondary acute myeloid leukemia (AML)? Can gene expression profiles be used to further refine treatment stratification strategies?¹ 

These and other questions can only be answered definitively in larger randomized trials. However, the St Jude team has planted many seeds for subsequent studies. We should remember that many components of modern therapy (CNS prophylaxis, intermediate dose methotrexate, etc) began with nonrandomized trials. With proper sunlight, water, and fresh air, great oaks from little acorns may grow.