The utility of neuroimaging in the evaluation of children with migraine and chronic daily headache, with normal neurologic examination, was determined by a retrospective review of 302 patients, aged between 6 and 18 years, seen in the Pediatric Neurology Clinic, Children’s Hospital, Eastern Virginia Medical School, Norfolk, 1997-1999. Other types of headache among the patients reviewed included: migrainelike symptoms (10%), chronic daily headache and abnormal neurologic examination (2%), secondary headache (17%), complicated migraine (7%), posttraumatic headache (7%), seizure-related headache (4%), brain tumors (3.6%), tension-type headache (3%), and pseudotumor cerebri (1.3%).
Of 107 (35% of total) with uncomplicated migraine, 42 (39%) received CT scans, and 2 (5%) were abnormal (arachnoid cyst in 1 and dilated Virschow-Robin space in 1). Of 12 (11%) who received an MRI, 2 (17%) were abnormal, both having a Chiari type 1 malformation. Of 30 patients with chronic daily headache, 17 (57%) received CT scans, and 3 (18%) were abnormal (a maxillary opacification, a mucous retention cyst, and an occult vascular malformation). Of 8 (27%) who had an MRI, 2 (25%) were abnormal, 1 a Chiari type 1 malformation and the other an occult vascular malformation. The yield of neuroimaging (CT and MRI) in children with uncomplicated migraine and chronic daily headache was 4% and 17%, respectively. It was concluded that since none of the abnormalities was associated with headache presentation or necessitated surgical intervention, the neuroimaging was not warranted in these patients. 
COMMENT. In children presenting with uncomplicated migraine and chronic daily headache and having a normal neurologic examination, the yield of neuroimaging studies is 4% and 17%, respectively. In the above retrospective analysis of cases, the types of CT and MRI abnormalities detected were not considered significant in the etiology and management of the headache disorder. It was concluded that neuroimaging is unwarranted in these specific headache syndromes. The occurrence of Chiari 1 malformation in 3 cases cannot be dismissed as a coincidental finding, however, since headache may be the presenting symptom and surgery has occasionally been advocated (Stovner IJ et al. 1992; see Progress in Pediatric Neurology II. 1994;p158). The authors admit that further research is required, to include large prospective studies and the role of repeated neuroimaging in previously negative studies. The importance of neuroimaging in children with headache associated with abnormal neurological or significant physical findings is accepted.
While routine neuroimaging may not be warranted for the pediatrician or family practitioner who has referred the child with recurrent headache for consultation, practice guidelines for the pediatric neurologist must also include the availability of the patient for follow-up. If MRI is deferred, more frequent clinical neurologic evaluations may be necessary to exclude some underlying neurosurgical lesion. Headache, without localizing neurologic abnormalities or signs of increased intracranial pressure, may be an uncommon presenting symptom of brain tumor. The luxury of observation over time is not always available to the neurologist, and the deferral of imaging may not be practical or judicious. (See Progress in Pediatric Neurology III. 1997;p185, for previous studies on brain imaging indications for headaches and commentary).