Charts of all children referred to the pediatric neurology clinic, Schneider Children’s Hospital, New Hyde Park, NY, for evaluation of headaches over a 2-year period were reviewed retrospectively for headache characteristics, indications for performing CT and MRI studies, and imaging results. Of 133 patients ages 3 to 18 years, 52% had migrainous headaches, 21% chronic tension headaches, and 19% were unclassified. The indications for brain imaging in 78 patients examined (MRI 45, CT 27, both 6) were not specified in 17, atypical headache pattern in 12, parental concern 12, physician concern about cerebral tumor 11, systemic symptoms of fatigue and weight loss in 11, focal symptoms or signs during headaches in 7, neurologic or ocular abnormalities 6, and increasing severity or frequency of headaches in 5. None of the scans showed brain tumor, vascular abnormality, or hydrocephalus that required neurosurgical intervention. Abnormal scans in 11 patients included evidence of chronic sinusitis in 7, a neuroepithelial cyst adjacent to the foramen of Monroe treated conservatively in 1, right temporal arachnoid cyst in 1, left cerebral hemiatrophy in 1, and Dandy-Walker malformation in 1. 
COMMENT. The authors conclude that brain imaging has very limited value in the management of childhood headaches in the absence of clinical signs of structural brain lesion. Although none of 11 (14%) positive scans was considered indicative of a treatable disease at the time, the abnormalities uncovered might potentially have proved significant. I would agree that brain imaging is not warranted in patients with well-defined migraine and that routine use of neuroimaging is to be discouraged. However, patients referred to a pediatric neurologist for chronic headache have usually been screened by their pediatrician or family physician and the pressures for neuroimaging may be more demanding.
An MRI should be considered especially in those with atypical recurrent headaches, a recent change in the character of the headache, persistent vomiting, a history of seizures, abnormal neurologic findings, and in younger age groups. However, the heavy sedation required for the MRI in a young child and risks of an adverse reaction to contrast medium with CT must be weighed against the benefits of the study. Headache as the sole manifestation of brain tumor is a rare occurrence. As a further consideration, the luxury of follow-up evaluation and observation over time may not be available to the neurologist who examines a patient in consultation, and the deferral of imaging may not be practical or judicious. For further commentary on the indications for imaging in headache and the value of an EEG as a preliminary test in diagnosis, see Progress in Pediatric Neurology II, PNB Publishers, 1994, ppl64-6).