High-dose methotrexate in children and young adults with ALL and lymphoblastic lymphoma: Results of the randomized phase III study UKALL 2011

Purpose
UKALL 2011 randomly assigned children and young adults (younger than 25 years) with ALL or lymphoblastic lymphoma. The aims were to reduce induction toxicity (randomization 1 [R1]), CNS relapse risk (randomization 2 [R2]–interim maintenance [R2IM]), and maintenance morbidity (R2pulses).

Methods
R1 compared induction dexamethasone (dex) for 28 days (6 mg/m²; standard) with 14 days (10 mg/m²; short). R2 was a factorial randomization resulting in four arms: high-dose methotrexate (HDM) with pulses, HDM without pulses, standard interim maintenance (SIM) with pulses (standard of care), and SIM without pulses. The primary end points were reduction in steroid-related toxicity (R1), CNS relapse rate (CNSR, R2IM), and bone marrow relapse rate (BMR, R2pulses; ALL only, noninferiority margin 5%). Event-free survival (EFS) was an additional primary end point for both randomizations.

Results
Of 2,750 eligible patients registered between April 2012 and December 2018, 1,902 were randomly assigned to R1 and 1,570 to R2. Median follow-up is 99 (R1) and 87 months (R2). There were no differences in steroid-related toxicity between short and standard dex (23.8% v 25.5%; P = .41) and CNSR between SIM and HDM (0.98 [95% CI, 0.65 to 1.49]; P = .94; 5-year rates: SIM 5.3% and HDM 5.5%). EFS was no different between R1 and R2IM arms. BMR in the no pulses arm was noninferior (+1.7% increase at 5 years [95% CI, –1.5 to 4.1]; hazard ratio [HR], 1.19 [95% CI, 0.87 to 1.62]; P = .27). Although the EFS in the no pulses arm was inferior (1.34 [95% CI, 1.05 to 1.73]; P = .021), this was not significant for relapse (HR, 1.24 [95% CI, 0.96 to 1.62]; P = .10).

Conclusion
Shorter duration of induction dex does not reduce steroid-related toxicity and HDM does not improve CNSR within a UKALL treatment backbone. Omission of pulses is noninferior for BMR.