Prognostic criteria for experimental protocols in pediatric brainstem gliomas

Prognostic criteria were developed to select children with high-risk brainstem gliomas for experimental treatment with hyperfractionated radiation. Patients were required to have (a) tumors confined primarily to the brainstem; (b) positive findings in two of three groups of neurologic signs (cranial nerve deficits and long-tract and cerebellar signs), and (c) symptom duration of less than 6 months. The 18-month survival of only 1 of 33 evaluable patients suggests that these criteria did in fact select tumors of poor prognosis. Such noninvasive predictors provide a valuable means of selecting for experimental treatment those children who would not be expected to benefit from standard therapies for malignant brainstem tumors. These criteria do not require routine biopsy for prognostic purposes, and avoid exposing the children to unnecessary risk.