Magnetic resonance imaging in the evaluation of treatment-related central nervous system damage.

Neurologic and neuropsychologic treatment related sequelae are increasingly encountered in children with cancer, but conventional means of neurologic investigation are insensitive to the presence and extent of damage. Magnetic resonance imaging (MRI) has shown brain damage not demonstrable by other means of investigation. For this reason, 11 children with cancer and with nontumor-related neurologic dysfunction were studied on a 1.5 Tesla MRI unit. All had concurrent computed tomography (CT). MRI abnormalities were seen in all (100%) patients. In 10 of 11 patients, abnormalities were of greater extent on MRI than on CT. White matter changes were frequently seen on MRI without corresponding CT abnormality. Those patients with the most severe forms of neurologic compromise had the most extensive changes on MRI. Focal neurologic findings correlated well with regions of focal signal change. Milder forms of neurologic compromise occurred in patients with definite, but less extensive, periventricular and/or subcortical change on MRI. MRI is more sensitive than CT in demonstrating treatment-related neurologic damage in children with cancer, and the type of change seen on MRI seems to correlate well with the type and severity of neurologic dysfunction present.