Background: Brain tumours rarely occur in survivors of childhood acute lymphoblastic leukaemia after cranial radiotherapy. An unusually high frequency of brain tumours seen among children enrolled in one of our leukaemia treatment protocols, Total Therapy Study XII, prompted us to identify the potential causes of this complication.
Methods: We assessed clinical, biological, and pharmacokinetic features in all 52 children who received prophylactic cranial radiotherapy. We compared the cumulative incidence of brain tumours between subgroups, and with that of 421 children who received radiotherapy in previous studies.
Findings: The incidence of brain tumours among irradiated children (six of 52, 12.8% [SE 5.0]) was high compared with patients in the same study who did not receive radiotherapy (none of 101; p=0.0008) and with other protocols that included cranial radiotherapy (p<0.0001). Of the six children, four had erythrocyte concentrations of thioguanine nucleotide metabolites higher than the 70th percentile for the entire cohort, and three had a genetic defect in thiopurine catabolism. The 8-year cumulative incidence of brain tumour among children with defective versus wild-type thiopurine methyltransferase phenotype was 42.9% (SE 20.6) versus 8.3% (4.7; p=0.0077). This protocol differed from previous protocols, in that more intensive systemic antimetabolite therapy was given before and during radiotherapy.
Interpretation: These data support the elimination of prophylactic radiotherapy for acute lymphoblastic leukaemia except in patients at high risk of central-nervous-system relapse. Underlying genetic characteristics and treatment variables may be associated with an increased risk of radiation-associated brain tumours.