Management of childhood craniopharyngioma: can the morbidity of radical surgery be predicted?

Seventy-five children treated for craniopharyngioma between 1973 and 1994 were studied to demonstrate which pre- and intraoperative factors were indicative of a poor outcome as defined by a quantitative assessment of morbidity. This involved a retrospective review of 65 patients and a prospective study of 10 patients focused on clinical details and cranial imaging and a follow-up "study assessment" of 66 survivors performed over the last 2 years. As part of the assessment, 63 patients underwent magnetic resonance imaging with a three-dimensional volume acquisition sequence 1.5 to 19.2 years after initial surgery. Predictors of high morbidity included severe hydrocephalus, intraoperative adverse events, and young age ( < or = 5 years) at presentation. Predictors of increased hypothalamic morbidity included symptoms of hypothalamic disturbance already established at diagnosis, greater height ( > or = 3.5 cm) of the tumor in the midline, and attempts to remove adherent tumor from the region of the hypothalamus at operation. Large tumor size, young age, and severe hydrocephalus were predictors of tumor recurrence, whereas complete tumor resection (as determined by postoperative neuroimaging) and radiotherapy given electively after subtotal excision were less likely to be associated with recurrent disease. Based on these findings, the authors propose an individualized, more flexible treatment approach whereby surgical strategies may be modified to provide long-term tumor control with the lowest morbidity.