Dr. Kranz and Dr. Provenzale are at the Department of
Radiology, Duke University Medical Center, Durham, NC. Dr. Provenzale
is also at Departments of Radiology, Oncology and Biomedical
Engineering, Emory University School of Medicine, Atlanta, GA.
Headache is a common problem that results in over 2 million visits to the emergency department (ED) every year.1 Most headaches are benign, self-limiting conditions or manifestations of chronic headache syndromes.2 Even among patients presenting with the “worst headache of my life,” benign causes far exceed life-threatening causes.3
several serious and potentially life-threatening etiologies of headache
can and do cause patients to present to the ED. Physicians are tasked
with differentiating the benign causes from the more serious, and
imaging often plays a major role in making this distinction.
head computed tomography (NCT) is the most common initial imaging test
ordered for headache patients presenting to the ED.1 In most
cases, this initial NCT will be normal. However, several potentially
life-threatening causes of headache do appear as abnormalities on NCT,
and these manifestations can range from subtle to highly conspicuous. In
light of the high prevalence of normal CT examinations for headache,
maintaining vigilance when reviewing these studies can be a challenge.
An active search pattern may help radiologists to avoid missing subtle
but potentially serious diagnoses.
This 2-part series aims to
review an approach to imaging patients presenting with severe headaches
using a method that mirrors the everyday experience of radiologists.
This first part focuses on potentially life-threatening diseases that
commonly produce positive findings on the initial NCT. The goal of this
article is to facilitate development of an active search pattern for
NCT, which is encountered most frequently when evaluating headaches.
2 will discuss diseases that often do not show findings on NCT. Because
NCT may appear normal in these patients, these conditions require a
heightened level of clinical suspicion to make the diagnosis, as well as
a solid knowledge of the strengths and limitations of various imaging
techniques. The goal of the second part is to explain how imaging
modalities other than NCT may be used to review particular elements of
the clinical presentation that may prompt radiologists to suggest such
further imaging and make an accurate diagnosis.
Subarachnoid hemorrhage (SAH) is
often the first diagnosis considered when evaluating a severe headache.
However, most patients with this complaint do not have SAH. In one
prospective investigation, 78% of patients presenting with the “worst
headache of my life” did not have subarachnoid hemorrhage.4
Among unselected patients with headaches of varying severity who present
to the emergency department, SAH is even less common, with several
studies reporting an incidence of 1% or less.1,3,5
these facts, SAH is a frequently pursued diagnosis because of its high
rates of rebleeding and poor outcomes associated with untreated
hemorrhage.6 Stated differently, the negative consequences of
failure to detect SAH far outweigh the costs of screening every such
patient. Perhaps because of the sheer volume of patients who present
with headache, clinical misdiagnosis remains a problem, with initial
misdiagnosis of SAH seen in 25% to 51% of cases.5 Even in the ED, one of every 20 cases of SAH may fail to be diagnosed.7
literature review found CT to be highly sensitive for detecting SAH,
with rates of 91% to 98% within the first 12 to 24 hours.8 Significantly,
after 12 hrs to 24 hrs, CT’s sensitivity for SAH declined 82% to 84%,
and the sensitivity of unenhanced CT at 1 week fell to 50%.8 Because
of the potentially catastrophic consequences of missing SAH and the
inability of imaging to detect all cases of SAH, lumbar puncture is
still recommended in cases of clinically suspected SAH with a negative
CT.6,9 Nonetheless, by focusing on certain imaging features
on unenhanced CT, the diagnostic yield of unenhanced CT can be increased
even several days after initial presentation. In particular, the
dependent portions of the subarachnoid space and ventricular system
should be carefully examined, as these areas may reveal subtle levels of
SAH that have settled in the cerebrospinal fluid (CSF). Because
patients are usually scanned in the supine position, particularly
important locations include the interpeduncular cistern, the occipital
horns of the lateral ventricles, the quadrigeminal plate cistern, and
the dependent portions of the Sylvian fissures (Figure 1).
Once SAH is detected, the cause must be sought. In 80% of cases, the etiology is a ruptured aneurysm.10
In these patients, the pattern of hemorrhage on the initial NCT can
help predict the site of the ruptured aneurysm, as indicated in Table 1.11
In some cases, the aneurysm itself can be seen as a filling defect
against a background of subarachnoid blood (Figure 2). Imaging with
conventional angiography, CT angiography (CTA), or less commonly,
magnetic resonance angiography (MRA) is mandatory for definitively
localizing the aneurysm.
In the remaining 20% of cases, a
nonaneurysmal cause is responsible. Roughly half of such cases are due
to nonaneurysmal perimesencephalic hemorrhage, which is thought to
result from venous bleeding.10,12 In these patients, the
hemorrhage is located in the interpeduncular cistern and immediately
anterior to the brainstem (Figure 3).13 Recognizing this pattern is important, as it helps to determine prognosis and to guide subsequent imaging.
with isolated perimesencephalic hemorrhage almost always have no
evidence of aneurysm on angiography, fare much better clinically than
patients with aneurysmal hemorrhage, and are not at risk for recurrent
hemorrhage.14 Other causes of angiogram-negative SAH include
trauma, drug abuse (especially cocaine abuse), sickle cell disease, and
When adequate clinical information is
available, CT diagnosis of SAH does not suffer from high rates of
false-positive interpretations.Nevertheless, several mimics of SAH can
manifest on CT as increased density in the subarachnoid space.
Pseudo-subarachnoid hemorrhage is one such mimic; it can occur in the
setting of markedly increased intracranial pressure, such as diffuse
cerebral edema (Figure 4). The increased density in the subarachnoid
space in pseudo-subarachnoid hemorrhage has been postulated to be due to
engorgement of pial vasculature combined with heightened vascular
conspicuity due to decreased parenchymal attenuation.16
spread of tumor may cause increased attenuation of the subarachnoid
space on CT, particularly in neoplasms with highnuclear-to-cytoplasmic
ratios, such as lymphoma (Figure 5). Finally, myelographic contrast
material can mimic subarachnoid blood. Although one might expect a
history of recent myelogram to be readily available from patients
presenting with headache, the authors’ experience is that such details
occasionally may not be immediately available when the myelogram has
been performed at another medical facility.
Many entities can cause brain
parenchymal hemorrhage. An in-depth discussion of the pathophysiology
and imaging of brain hemorrhageis beyond the scope of this article; a
number of excellent review articles are available to explore this topic
further.17,18 The major question faced by clinicians and
radiologists is whether an underlying brain lesion, commonly vascular in
nature, exists as a cause of hemorrhage. The answer to this question
can often be quickly provided noninvasively by CTA or MRA (Figure 6).
Clinicians can best select patients for angiographicimaging by
considering several demographic, historical, and anatomical factors,
most notably age, blood pressure, and hemorrhage location. In one study,
CTA defined a vascular cause in 15% of unselected patients presenting
with parenchymal hemorrhage.19 The incidence of a vascular
cause of spontaneous parenchymal hemorrhage in this study rose to 47%,
however, among patients < 46 years old.19 In addition to
patients under the age of 50, other factors associated with a vascular
etiology include absence of hypertension, presence of subarachnoid or
intraventricular hemorrhage, and hemorrhage location in either the
temporal or frontal lobes.19
Hydrocephalus should be considered as a
potential etiology in patients presenting with severe headache. In some
cases, especially when acute, untreated hydrocephalus can be fatal.
Therefore, evaluation of any brain imaging study must focus on the
caliber of the ventricularsystem. If prior imaging studies are
available, clinicians should carefully compare ventricular for interval
changes that might signal new onset of hydrocephalus.
the ventricular size is abnormally increased in a given patient often
requires the subjective judgment of the radiologist. In earl
yhydrocephalus, or in patients with brain volume loss due to aging or
parenchymal disease, correctly identifying hydrocephalus may be
difficult.Focusing on one particular location, the temporal horn of the
lateral ventricle may prove helpful. Disproportionate enlargement of the
temporal horns often indicates hydrocephalus, and can be useful in
distinguishing ex vacuo dilation of the ventricles (ie, dilation due to parenchymal volume loss) from true hydrocephalus.20
hydrocephalus is detected, the next step is to determine whether the
hydrocephalus is communicating or noncommunicating. Noncommunicating
hydrocephalus results from a lesion in the ventricular system that
obstructs flow of CSF. Its presence is suggested by the coexistence of a
dilated proximal ventricular system and a decompressed distal
ventricular system. The point of transition between the dilated and
decompressed ventricles should be carefully scrutinized for the presence
of a mass. Because the anatomic “choke points” of the ventricular
system are located near the midline, one must be careful to scrutinize
midline structures, including the foramen of Monro, the aqueduct of
Sylvius, and the inferior fourth ventricle (Figure 7). Communicating
hydrocephalus, by contrast, shows dilation of the entire ventricular
system. In these cases,current or prior diseases affecting the CSF, such
as subarachnoid hemorrhage, meningitis, and CSF dissemination of tumor,
should be investigated.
infarction is relatively often accompanied by headache, particularly in
young patients or those with history of migraine.21 Detecting
arterial infarction on NCT depends on the duration and severity of
vascular occlusion. In general, most patients show ischemic changes on
NCT within 6 hours of symptom onset.22 Once arterial ischemia
is recognized, the etiology should be sought to help gauge the risk of
recurrence and determine optimal treatment.23 Fortunately,
the neurologic deficits associated with arterial ischemia generally help
distinguish these patients from patients with benign headaches.24
Posterior reversible encephalopathy syndrome
(PRES) is a neurological syndrome that manifests on imaging studies as
multifocal areas of edema usually involving the parieto-occipital white
matter, but often also involving other areas, including cortical and
subcortical watershed distributions, and occasionally the cerebellum,
basal ganglia, or brainstem (Figure 8).25 Headache is often a
clinical feature of PRES, although usually not the only presenting
feature. Typically, patients with PRES will also exhibit seizures,
visual disturbances, and alteration of consciousness.26 Furthermore,
PRES is most commonly observed in conjunction with particular disease
states, especially hypertension, eclampsia/preeclampsia,
immunosuppression, chemotherapy, and autoimmune diseases.27
The combination of suggestive imaging patterns, typical clinical
presentation, and predisposing conditions should suggest the diagnosis.
PRES abnormalities are best seen and characterized on MRI, they are
usually visible on CT. One study compared CT and MR detection of PRES,
and found that CT showed abnormalities in most cases of PRES (78%), but
that MRI provided diagnosis with greater specificity.28 In
that study, CT provided a specific diagnosis in only 45% of cases. In
situations where the clinical findings are suggestive but initial CT is
negative or equivocal, MRI should be performed for confirmation.
Headache is common in patients with brain tumors.29
As in the case of PRES, however, headache usually is not the only
presenting clinical feature in patients with newly diagnosed
intracranial tumors; patients usually have coexistent neurologic
deficits. In one study of 183 patients presenting with a brain tumor,
isolated headache was the clinical presentation in only 8% of patients.30 Because
primary headaches are much more common than tumors as a cause of
headache, brain tumors overall are not a common cause of acute headache,
with an incidence of < 1% among patients undergoing imaging for
headache.31 Although full characterization may require
further imaging, brain tumors of sufficient size to cause headache are
often readily visible on NCT.
presenting with severe headaches can present a diagnostic challenge due
to the wide variety of causes that can range from benign to
self-limiting to life threatening. Noncontrast CT plays a major role in
the initial work-up of these patients. Therefore, awareness of the
diseases that commonly show abnormalities on the initial CT is critical
to development of an active search pattern.
- Goldstein JN, Camargo CA, Jr.,
Pelletier AJ, Edlow JA. Headache in United States emergency
departments: Demographics, work-up and frequency of pathological
diagnoses. Cephalalgia. 2006;26:684-690.
- Detsky ME, McDonald DR, Baerlocher MO, et al. Does this patient with headache have a migraine or need neuroimaging? JAMA. 2006;296:1274-1283.
- Morgenstern LB, Huber JC, Luna-Gonzales H, et al. Headache in the emergency department. Headache. 2001;41:537-541.
JJ, Stiell IG, Sivilotti ML, et al. High risk clinical characteristics
for subarachnoid haemorrhage in patients with acute headache:
Prospective cohort study. BMJ.2010;341:c5204.
- Edlow JA, Caplan LR. Avoiding pitfalls in the diagnosis of subarachnoid hemorrhage. N Engl J Med. 2000;342:29-36.
- Al-Shahi R, White PM, Davenport RJ, Lindsay KW. Subarachnoid haemorrhage. BMJ. 2006;333:235-240.
- Vermeulen MJ, Schull MJ. Missed diagnosis of subarachnoid hemorrhage in the emergency department. Stroke. 2007;38:1216-1222.
S, Wallmann P. Towards evidence based emergency medicine: Best BETs
from the Manchester Royal Infirmary. Does a normal CT scan rule out a
subarachnoid haemorrhage? Emerg Med J. 2001;18:271-273.
- Van der Wee N, Rinkel GJ, Hasan D, van Gijn J. Detection of
subarachnoid haemorrhage on early CT: Is lumbar puncture still needed
after a negative scan? J Neurol Neurosurg Psychiatry. 1995;58:357-359.
- Vermeulen M, van Gijn J. The diagnosis of subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry. 1990;53:365-372.
Scotti G, Ethier R, Melancon D, et al. Computed tomography in the
evaluation of intracranial aneurysms and subarachnoid hemorrhage. Radiology. 1977;123:85-90.
- Van der Schaaf IC, Velthuis BK, Gouw A, Rinkel GJ. Venous drainage in perimesencephalic hemorrhage. Stroke. 2004;35:1614-1618.
GJ, Wijdicks EF, Vermeulen M, et al. Nonaneurysmal perimesencephalic
subarachnoid hemorrhage: CT and MR patterns that differ from aneurysmal
rupture. AJNR Am J Neuroradiol. 1991;12:829-834.
- Greebe P, Rinkel GJ. Life expectancy after perimesencephalic subarachnoid hemorrhage. Stroke. 2007;38:1222-1224.
- Rinkel GJ, van Gijn J, Wijdicks EF. Subarachnoid hemorrhage without detectable aneurysm. A review of the causes. Stroke. 1993;24:1403-1409.
CA, 2nd, Burdette JH, Elster AD, Williams DW 3rd. Pseudo-subarachnoid
hemorrhage: a potential imaging pitfall associated with diffuse cerebral
edema. AJNR Am J Neuroradiol. 2003;24:254-256.
- Fischbein NJ, Wijman CA. Nontraumatic intracranial hemorrhage. Neuroimaging Clin N Am. 2010;20:469-492.
- Dainer HM, Smirniotopoulos JG. Neuroimaging of hemorrhage and vascular malformations. Semin Neurol. 2008;28:533-547.
Almandoz JE, Schaefer PW, Forero NP, et al. Diagnostic accuracy and
yield of multidetector CT angiography in the evaluation of spontaneous
intraparenchymalcerebral hemorrhage. AJNR Am J Neuroradiol. 2009;30:1213-1221.
M, Hochberg FH. Ventricular differences between hydrostatic
hydrocephalus and hydrocephalus ex vacuo by computed tomography. Neuroradiology. 1979;17:191-195.
- Tentschert S, Wimmer R, Greisenegger S, et al. Headache at stroke
onset in 2196 patients with ischemic stroke or transient ischemic
attack. Stroke. 2005;36:e1-3.
- Tomura N, Uemura K, Inugami A, et al. Early CT finding in cerebral infarction: Obscuration of the lentiform nucleus. Radiology. 1988;168:463-467.
- Rovira A, Grive E, Alvarez-Sabin J. Distribution territories and causative mechanisms of ischemic stroke. Eur Radiol. 2005;15:416-426.
- Provenzale JM. Imaging evaluation of the patient with worst headache of life--it’s not all subarachnoid hemorrhage. Emerg Radiol. 2010;17: 403-412.
AM, Short J, Truwit CL, et al. Posterior reversible encephalopathy
syndrome: Incidence of atypical regions of involvement and imaging
findings. AJR Am J Roentgenol. 2007;189:904-912.
- Fugate JE, Claassen DO, Cloft HJ, et al. Posterior reversible
encephalopathy syndrome: Associated clinical and radiologic findings. Mayo Clin Proc. 2010;85:427-432.
- Bartynski WS. Posterior reversible encephalopathy syndrome, part 1: Fundamental imaging and clinical features. AJNR Am J Neuroradiol. 2008;29:1036-1042.
- Bartynski WS, Boardman JF. Distinct imaging patterns and lesion distribution in posterior reversible encephalopathy syndrome. AJNR Am J Neuroradiol. 2007;28:1320-1327.
- Loghin M, Levin VA. Headache related to brain tumors. Curr Treat Options Neurol. 2006;8:21-32.
A, Ibanez FJ, Herrera S, et al. Isolated headache as the presenting
clinical manifestation of intracranial tumors: A prospective study. Cephalalgia. 1994;14:270-272.
- Evans RW. Diagnostic testing for the evaluation of headaches. Neurol Clin. 1996;14:1-26.