The influence of radiation therapy quality in the patterns of failure of pediatric medulloblastoma

Purpose: In this study factors are analyzed that may potentially influence the site of failure in medulloblastoma children. Patient, disease, and treatment related variables are analyzed with a special focus on radiotherapy (RT) time-dose and technical factors. Methods and Materials: Eighty-six children and adolescents with a diagnosis of medulloblastoma were treated in Switzerland during the period 1972-1991. Postoperative megavoltage RT was delivered to all patients. Simulation and portal films of the whole brain irradiation (WBI) fields were retrospectively reviewed in 77 patients. The distance from the field margin to the cribiform plate and to the floor of the temporal fossa were carefully assessed and correlated with supratentorial failure-free survival. Fields were considered acceptable if the portal limits surrounded the basal portion of the brain with at least 5 mm margin for x-rays and 10 mm margin for Co-60. A marginal miss was concluded if the margins were 1-4 mm for x-rays or 4-9 mm for Co-60, while a major miss was considered to exist if the margins were ≤0 mm (field limit inside the target) for x-rays, or ≤3 mm for Co-60. The above mentioned limits were chosen on the assumption that the field margins associated with a marginal miss and with a major miss would underdose the excessively shielded regions by at least 25-50% and 50-75%, respectively. The correctness of the field margins around the convexity was also assessed. Field limits inside the brain convexity were considered unacceptable. In 19 children the spine was treated with high-energy electron beams the reminder with megavoltage photons. Simulation and port films of the posterior fossa fields were also reviewed in 72 patients. The surface of the treatment region was calculated to evaluate the potential relationship between field size and posterior fossa failure. The median field size was 72 cm2 (40-234, range). In addition to field size, a technique was judged to be acceptable only if the anterior field limit reached the posterior clinoids and the inferior limit included C1 (first cervical vertebra). The field size and the field limits were evaluated and correlated with posterior fossa failure-free survival. Results: Five- and 10-year overall survival was 57% and 39%, respectively. The median follow-up for the surviving patients was 8.6 years (3.1-16.8, range). Forty-nine patients failed either locally or distantly. Twenty-four patients failed in the posterior fossa with or without distant failure. Twelve patients failed in the supratentorial region only. The lower portion of the frontal and/or temporal lobes was the site of failure in 8 cases. The supratentorial irradiation technique was considered acceptable in 13 patients (17%). A marginal miss was concluded in 28 patients (36%), while a major miss was observed in 36 patients (47%). Twelve patients failed in the supratentorial region and 9 of these patients belonged to the group of 36 children in whom the inferior portion of the brain had been clearly underdosed. On multivariate analysis only base of skull field correctness was retained as being significantly correlated with supratentorial failure-free survival (p=0.049). Fields were judged too tight around the convexity (i.e., field limits inside the brain convexity) in 7 patients (5 with an additional major miss of the base of skull). None recurred in the supratentorium. Neither the total dose to the spinal theca nor the treatment technique (electron versus photon beams) were significantly correlated with outcome. Posterior fossa failure-free survival was not influenced by total dose, overall treatment time, field size, or field margin correctness. Conclusion: A correlation between WBI field correctness and supratentorial failure-free survival was observed. Underdosage of the basal brain portion was a frequent event among the patients in this series. This, and a gravity-related sanctuary effect may explain why the meningeal region close to the base of skull (in addition to the ventricles) is at higher risk for relapse. Treating the spine with electron beams was not deletereous. A significant correlation between local control and other technical factors was not observed, including those relating to posterior fossa treatment