Importance of radiation time and dose factors on outcome for childhood medulloblastoma

The purpose of this study was to investigate the relationship of posterior fossa radiation therapy duration (PFRTD) and relapse-free survival (RFS) following adjuvant craniospinal RT for childhood medulloblastoma. A retrospective audit was performed assessing all children aged <18 years managed with adjuvant craniospinal RT for medulloblastoma in Australia and New Zealand in 1980-1993. Children receiving prolonged (>180 days) pre-RT chemotherapy were excluded. Data were obtained for potential prognostic factors in domains of patient, tumour and treatment factors. Radiation therapy time factors assessed were PFRTD and time interval from surgery to commencement of RT (SRTD). The end-point assessed was RFS and analysis was performed using Cox regression and Kaplan-Meier survival. One hundred and eighty-nine children were identified from 10 oncology units, with data available from 182 children for analysis. Median follow up was 5.3 years. Seventy-three per cent of children presented with disease confined to the cerebellum; 13% had initial neuraxis disease. Macroscopic resection was described in 54%; 42% received adjuvant chemotherapy. Median RT dose and RT duration to PF was 55 Gy and 45 days, respectively. Seventy-eight relapses occurred with a 10-year actuarial RFS of 58.2% (standard error +/- 4%). On univariate analysis, increasing PF dose (P = 0.002), age >5 years (P = 0.006), and more thorough extent of surgical resection (P = 0.043) were associated with improved RFS; PFRTD (P = 0.20) and SRTD (P = 0.51) were not associated with RFS. On multivariate analysis, although both PF dose (P = 0.004) and extent of surgery (P = 0.045) remained strongly significant, RT duration was now associated with RFS (P = 0.049). Other factors assessed that did not reach significance were patient age, local tumour extent, presence of internal shunt and use of chemotherapy. The importance of local treatment factors was confirmed in this audit with established prognostic factors such as primary tumour macroscopic resection and adequate PF RT dose being associated with RFS. A treatment time effect is weakly suggested, although less significant than RT dose delivered.