“Lesser” ALL: treat less or study less?

In this issue of Blood, Chauvenet and colleagues report the results of a recent Pediatric Oncology Group (POG) trial for children with “lesser” risk B-lineage acute lymphoblastic leukemia (ALL), in whom they observed a 97% long-term survival rate.

These results are quite favorable, compared with contemporary trials. The Children's Cancer Group (CCG) 1922 trial reported a 93% 6 year survival rate,¹  but utilized agents with potential for serious long-term effects that were avoided in the POG trial: doxorubicin, cytoxan, and dexamethasone. The CCG trial also utilized a duration of therapy that was longer (26 months for girls, 38 months for boys) than the POG trial (30 months for all patients).

Upon closer inspection, however, it becomes apparent that the difference in therapy between these 2 trials is not great. For example, 75mg/m² cumulative doxorubicin (versus none), 1 g/m² cumulative cyclophosphamide (versus none), and 1.0 to 1.4 g/m² dexamethasone (versus 3.4 g/m² prednisone) were used for the CCG and POG trials, respectively. In addition, the POG trial used more intrathecal chemotherapy: 21 injections of triple therapy (cytosine arabinoside plus methotrexate plus hydrocortisone), compared with 15 (girls) to 19 (boys) injections of methotrexate alone; and more parenteral methotrexate (6 g/m² intravenously during consolidation and 20 mg/m² intramuscularly weekly for 105 weeks of maintenance) compared with none intravenously or intramuscularly in the CCG approach. Lastly, the POG trial probably included patients with a better intrinsic prognosis in that only patients with either favorable trisomy cytogenetics or a favorable DNA index were eligible. No such restrictive inclusion criteria were used in the CCG trial.

Unfortunately, the balance of the toxicities of these 2 contrasting approaches cannot be assessed well from the reported data: more fatalities from sepsis (2.5% versus none) in the CCG trial, but more hepatic toxicity (52% versus 13%-19%) and probably more neurotoxicity (8.8% overall with 2.2% seizures) in the POG approach. Second malignancies were similar (0.6% CCG, 0.3% POG). Regrettably, there was no reporting of transfusion use, or of hospital days for either trial.

So, what are we to make of all this? The only way to determine if the POG approach is superior for the exact same favorable group of ALL patients is to perform a randomized controlled trial. However, in the current age of declining funding for research, is it worth it to study a disease in which at least 95% of the patients would be expected to survive, regardless of treatment regimen assigned? In addition, we are encountering an increasing number of parents and patients who are not willing to risk some additional toxicity for the small potential gain that might be achieved by being treated in current investigational arms of National Cancer Institute (NCI)–sponsored trials for newly diagnosed ALL. At the Children's Hospital Oakland (Oakland, CA) in 2006, a 40% refusal rate was witnessed for participation in a standard-risk ALL trial. If the POG approach were to become the investigational regimen for the next trial in standard/favorable risk ALL, this toxicity concern might be ameliorated and thus allow the continuing quest for the eventual total cure of childhood ALL.