Equivocal conclusions
As noted above, CNS relapse, the primary endpoint for the interim maintenance randomization, did not differ between the groups, with identical 5-year relapse rates. Similarly, 5-year event-free survival (EFS) rates were 90.3% in the high-dose group and 89.5% in the standard group, a difference that was not statistically significant (P = .68).
There was, however, an interaction between the first (short- vs. standard-course dexamethasone) and the interim maintenance randomizations, indicating significantly inferior EFS outcomes for patients who had received the short dose of dexamethasone followed by high-dose methotrexate, especially among patients who did not receive pulses (P = .006).
An analysis of patients treated with standard dexamethasone showed that those who received high-dose methotrexate had a lower risk for bone marrow relapse, with a hazard ratio of 0.62 (P = .029), and trends, albeit nonsignificant, toward better EFS and overall survival.
In addition, the overall results suggested that steroid pulses could be safely omitted without leading to an increase in bone marrow relapses: the 5-year rates of bone marrow relapse were 10.2% with pulses and 12.2% without, although omitting pulses was associated with a slight but significant decrease in EFS overall (P = .01). The effect was attenuated among patients who had received standard-course dexamethasone and high-dose methotrexate. Leaving out the pulses also reduced rates of grade 3 or 4 adverse events, including febrile neutropenia, Ms. Kirkwood noted in her presentation.
The investigators plan to analyze quality-of-life outcomes related to dexamethasone-vincristine pulses to see whether doing so could tip the balance in favor of leaving them out of therapy, and they will continue to follow patients to see whether their findings hold.
UKALL2011 was funded by Children with Cancer UK, Blood Cancer UK, and Cancer Research UK. Ms. Kirkwood disclosed consulting for and receiving honoraria from Kite. Dr. Vora reported no relevant financial relationships.
A version of this article first appeared on Medscape.com.