Michael Eller; Peter J Goadsby


Expert Rev Neurother. 2013;13(3):263-273. 

In This Article

How to Walk the Line?

There is a rational and reasonable way to stratify the requirement for an MRI in patients with headache. Most of the evidence that follows supports a probabilistic guide to decision making in the common headache phenotypes. Additional features, often termed red flags (Box 1), are a helpful lens to focus the International Classification of Headache Disorders (ICHD)[5] to a higher pretest probability. These essentially suggest a secondary cause for headache, even if the phenotype is of a primary headache disorder like migraine. For example, a progressive throbbing headache with nausea and photophobia – completely novel for a patient – reaching a crescendo over months, necessitates an MRI brain scan to exclude an intracranial lesion.

Red flags include any novel and persistent headache (qualifiers follow below), and any new or progressive symptoms that can be ascribed to the CNS. Symptoms suggestive of raised or lowered intracranial pressure, particularly if progressive or persistent, or both; any new neurological signs; and any abrupt onset headache all require elective or emergent imaging. For example, sudden-onset headache quite atypical for a given patient will necessitate the urgent exclusion of subarachnoid hemorrhage (SAH) among other potential causes.

In the authors view, the ordering of an MRI brain scan implies that a neurological examination has been completed. A primary headache diagnosis such as migraine self-evidently implies a normal neurological exam in the context of the history.

Any change in headache semiology for a given patient must also be seen within a wider context. For example, a new diagnosis of episodic headache with typical features of migraine with aura, in the context of cyclical vomiting as a child and family history of migraine, does not mean the patient requires imaging, even though the headache is 'new'. Symptoms should never be interpreted exclusively within the silo of a specialty or subspecialty; relevant medical comorbidities must be taken into account. For example, a patient that is immunosuppressed secondary to an iatrogenic cause, such as an allogeneic bone marrow transplant, requires an MRI brain scan in the case of a novel headache, even in the absence of other accompanying symptoms. This goes some way towards excluding infection, posterior reversible encephalopathy syndrome, drug toxicity and other lesions such as a subdural hemorrhage.

The age of the patient must also be considered. Older patients with novel headache are much more likely to be suffering from intracranial pathology, such as an ischemic stroke.[6] In the pediatric population, intracranial tumors can present with headache a third of the time.[7] These are almost always associated with focal neurological signs, or signs and symptoms of raised intracranial pressure (ICP). As morbidity is attached to the procedural risks that go with imaging in young children, and brain tumors are an uncommon cause of headache, usually imaging is only appropriate when focal signs and/or symptoms are present.[8]

The evidence and rationale for the decision to image in the common headaches will be emphasized, as this represents a recurrent dilemma for clinicians. The bulk of this article will be devoted to this topic. The use of MRI in assessing headache in an acute setting will be briefly discussed, particularly utilizing the framework of 'red-flag' symptoms. Other headache phenotypes that are encountered less frequently in general practice or general neurology clinics are covered separately.[9] Finally, rare conditions that lead to significant and often pathognomonic changes on MRI will be discussed.