Should You Order Neuroimaging for Migraine?

Randolph W. Evans, MD


January 16, 2019

Editorial Collaboration

Medscape &

Should neuroimaging be obtained in patients with migraine? A 2000 report of the Quality Standards Subcommittee of the American Academy of Neurology makes the following recommendation: "Neuroimaging is not usually warranted in patients with migraine and a normal neurologic examination (Grade B)."[1] The American Headache Society's Choosing Wisely list in 2013 recommended: "Don't perform neuroimaging in patients with stable headaches that meet criteria for migraine."[2] When neuroimaging for migraine is performed, commonly incidental or normal anatomical findings are present.[3] Here, we review the evidence for or against neuroimaging for migraine, based on recent studies.

Recent Studies

Two recent studies provide additional information about the yield of neuroimaging in patients with migraine. Mullally and Hall[4] performed a prospective study on 100 patients with a diagnosis of migraine (45 without aura, 14 with aura, and 41 with chronic migraine) and normal results on neurologic examination. MRI scans of the brain were performed solely at their request. The duration of headaches ranged from 4 months to 40 years. MRI scans were normal in 82 patients and found clinically insignificant abnormalities in 17 patients. MRI was abnormal in one patient (chronic migraine without aura), finding a meningioma requiring surgery and radiotherapy, which is similar to the yield of brain tumor in the general asymptomatic population. The authors conclude, "Brain MRI obtained at the specific request of patients with a diagnosis of migraine in the presence of normal neurologic examination results has a yield that is equivalent to that of the general asymptomatic population. Patients do not seem to have more insight than the examining clinician with regard to detecting underlying structural abnormalities, and brain MRI should not be performed as part of the routine evaluation of migraine without a clear clinical indication."

A meningioma is not necessarily an incidental finding in an individual with migraine. Evans and colleagues[5] reported a 47-year-old woman with a left frontal secretory meningioma, which mimicked transformed migraine with and without aura. There is potential harm to the patient and the physician's medicolegal liability if these rare cases are not detected.

Wang and coworkers[6] recruited 1070 healthy controls and 1070 primary headache patients (including 665 with migraine) from the Chinese People's Liberation Army General Hospital who then underwent either CT or MRI scans. Abnormal scans were found in the following: 0.67% in migraine (3/665 with MRI and 0/291 with CT) compared with 0.73% abnormal scans in controls. Abnormalities in patients with migraine were two with hydrocephalus and two with tumors of the throat and nose. The researchers conclude, "The present study found that neuroimaging was unnecessary for the primary headache patients."

Reasons to Consider Neuroimaging for Presumed Migraine

Some neurosurgeons have a different perspective. Hawasli and colleagues[7] performed a retrospective review of patients who were diagnosed with a brain neoplasm from an open brain biopsy at Washington University. They found that 11/95 (11.6%) had isolated headaches, including 3/11 (27.3%) with diagnoses of migraine. No further details of the headaches were provided. "Although we do not recommend routine screening for the general population," they note, "we do contend that a substantial number of patients with brain tumors will present with isolated headaches."

Following guidelines does not indemnify physicians or protect them in a malpractice suit.[8,9,10] I treated a pregnant patient for a severe headache consistent with migraine status with a normal neurologic exam due to a pilocytic astrocytoma with mass effect.[11] She successfully sued her prior neurologist (and received $250,000 at settlement), who appropriately declined months earlier to obtain a scan, as the authors recommend, when she presented with a history of typical migraines and a normal neurologic exam, where guidelines do not recommend imaging. Until physicians are indemnified in malpractice cases when they follow guidelines, you might consider what a jury, rather than our peers, would consider indications for neuroimaging.

Many anxious patients and their family members are not reassured even after a long discussion about the low yield of neuroimaging. Howard and coworkers[12] performed a randomized controlled trial in a London headache clinic of 150 patients with chronic daily headache (76 randomly assigned to the offer of a brain scan and 74 to treatment as usual). Patients offered a scan were less worried about a serious cause of the headaches at 3 months, although this was not maintained at 1 year. However, patients with high levels of psychiatric morbidity who were offered a scan had significantly less overall costs due to lower utilization of medical resources as a result of being reassured.

Finally, you may wish to consider neuroimaging for presumed migraine where a migraine mimic[13] may be present, for the following reasons: unusual, prolonged, or persistent aura; increasing frequency, severity, or change in clinical features; first or worst migraine; migraine with brainstem aura; confusional; hemiplegic; late-life migraine accompaniments; aura without headache; side-locked headache; and post-trauma headache. A new publication reviews the sensitivity, specificity, and predictive value of red and orange flags.[14]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.