1. health and fitness
2. disease

Imaging of Head Trauma
Tuong H. Le, MD, PhD, and Alisa D. Gean, MD
T
raumatic brain injury (TBI) is a leading cause of mortality
and morbidity in the world’s population, especially those
under age 44.1 In the United States alone, the cost of head
trauma has been estimated to be over 40 billion dollars annually.2 The National Center for Injury Prevention and Control has estimated that approximately 2% of the entire population of the USA currently lives with disabilities caused by
TBI.3 TBI accounts for more than 500,000 emergency department visits annually.4 The etiologies of head trauma are usually associated with the patient age. In the elderly population,
accidental falls are the most common causes. Motor vehicle
crashes are common culprits in young patients. In children,
abuse and neglect are common reasons.
TBI can be classified into primary and secondary injuries.
Primary lesions are the direct result of trauma to the head,
and secondary lesions arise as complications of primary lesions. Clinically, this classification is important because secondary injuries can be preventable, whereas primary injuries,
by definition, have already occurred by the time the patient
first presents for medical attention. TBI can be further divided according to location (intra- or extra-axial), mechanism (penetrating/open or blunt/closed), and clinical severity
(minor, mild, moderate, or severe). The severity of head injury is usually based on the Glasgow Coma Scale (minor:
GCS ⫽ 15; mild: GCS ⱖ 13; moderate: 9 ⱕ GCS ⱕ 12;
severe: 3 ⱕ GCS ⱕ 8).5 Primary extra-axial lesions include
epidural, subdural, subarachnoid, and intraventricular hemorrhage. Primary intra-axial lesions include cortical contusions, intracerebral hematomas, axonal shearing injuries,
gray matter injury, and vascular injury. Acute and subacute
secondary injuries include cerebral edema, ischemia, and
brain herniation. Chronic secondary lesions include hydrocephalus, the cerebrospinal fluid (CSF) leak, leptomeningeal
cyst, and encephalomalacia. These lesions will be discussed
in further detail following a discussion about different methods for imaging TBI.
University of California, San Francisco, San Francisco General Hospital, San
Francisco, CA.
Address reprint requests to Alisa D. Gean, MD, Professor of Radiology,
Neurology, and Neurological Surgery, University of California, San
Francisco, Chief of Neuroradiology, San Francisco General Hospital,
1001 Potrero Avenue, San Francisco, CA 94110. E-mail: agean@
sfghrad.ucsf.edu.
0037-198X/06/$-see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1053/j.ro.2006.04.003
Imaging Techniques
Skull films are poor predictors of intracranial pathology and
should not be performed to evaluate adult TBI.6-8 In the lowrisk patient, skull films rarely demonstrate significant findings. In the high-risk patients, the lack of abnormality on
skull films does not exclude major intracranial injury.9 Patients who are at high risk for acute intracranial injury must
be imaged by CT. The “scout view” that is obtained with all
CT exams can be used as a “pseudoskull film.” In cases of
suspected child abuse, a skull series may occasionally identify a fracture that is not identified on the concomitant CT
examination. While this finding may not have surgical importance, it, nevertheless, warrants removal of the child from
the hostile environment.
Computed tomography (CT), in the setting of acute trauma,
is indicated for severe TBI (GCS ⬍ 8), persistent neurologic
deficit, antegrade amnesia, unexplained asymmetric pupillary response, loss of consciousness more than 5 minutes,
depressed skull fracture, penetrating injury, or bleeding diathesis or anticoagulation therapy.10 The goal of imaging is to
identify treatable injuries to prevent secondary damage. In
the acute setting, CT is the modality of choice because it is
fast, widely available, and highly accurate in the detection of
skull fractures and intracranial hemorrhage. Life-support and
monitoring equipment can easily be accommodated in the
CT scanner suite. In addition, CT is usually superior to MR in
revealing skull fractures and radio-opaque foreign bodies.
With modern CT scanners, contiguous 3.75- or 5-mm sections from the skull base to the vertex can be obtained in less
than 10 minutes. Thinner 1- or 2.5-mm sections are used to
evaluate the orbits, maxillofacial structures, and skull base.
With recent advances in multidetector CT (MDCT), thin
cross-sectional slices can be performed which allowed for
high-quality three-dimensional reconstruction. Thinner
slices also improve the diagnostic accuracy of CT in the evaluation of maxillofacial, orbital, and temporal bone fractures.
Intravenous contrast administration should not be performed because it can both mask and mimic underlying hemorrhage.
CT images must be reviewed using multiple windows and
levels. With the widespread utilization of picture archiving
and communication system (PACS), most imaging interpretation is now performed on computer workstations, allowing
177
T.H. Le and A.D. Gean
178
Figure 1 Subgaleal, subdural, and epidural hematoma. Axial CT
image demonstrates a left parietal subgaleal hematoma (arrow).
Note that it is superficial to the temporalis muscle (*). Subjacent to
the scalp injury is a biconvex, hyperdense, extra-axial collection
compatible with an acute epidural hematoma (#). Anterior to the
epidural hematoma is a crescent-shaped hyperdense collection consistent with an acute subdural hematoma (arrowhead).
inhomogeneity caused by the presence of blood. Hemosiderin, a breakdown product of blood, is ferromagnetic. The
presence of hemosiderin alters the local magnetic susceptibility of tissue, resulting in areas of signal loss on gradientecho T2*-weighted images. Because hemosiderin can persist
indefinitely, its detection on gradient-echo T2*-weighted images allows for improved evaluation of remote injury. However, due to the inherent inhomogeneity adjacent to the paranasal sinuses and mastoid air cells, gradient-echo images are
limited in the evaluation of cortical contusions of the inferior
frontotemporal lobes. This limitation is even more problematic at high magnetic field strength unless parallel imaging is
used.16
Diffusion-weighted imaging (DWI), which measures the random motion of water molecules in brain tissue, has improved
the evaluation of TBI. Through its superior sensitivity to foci
of acute shearing injury, DWI has especially improved detection of DAI.16-18 DWI reveals more DAI lesions than fast spinecho T2-weighted or gradient-echo T2*-weighted images in
patients imaged within 48 hours of injury. The apparent
diffusion coefficient (ADC), which measures the magnitude
of water diffusion averaged over a three-dimensional space, is
often reduced in acute DAI. The fractional anisotropy (FA),
which measures the preferential motion of water molecules
along the white matter axons, is frequently reduced in
chronic DAI. Tissue anisotropy is exploited in the new technique of diffusion tensor imaging (DTI).
Imaging Findings
for rapid setting of windows and levels. A narrow window
width (W: 80, L: 40) is used to evaluate the brain. A slightly
wider window width (W: 150, L: 75) is used to exaggerate
contrast between extra-axial blood and the adjacent skull. An
even wider window (W: 2500, L: 500) is used to evaluate the
osseous structures (Figs. 2 and 3).
Magnetic resonance imaging (MRI) is recommended for patients with acute TBI when the neurologic findings are unexplained by CT. MR is also the modality of choice for subacute
or chronic injury. MR is usually comparable to CT in the
detection of an acute epidural and subdural hematoma.11,12
However, MR is more sensitive to subtle extra-axial collections, nonhemorrhagic lesions, brainstem injuries, and subarachnoid hemorrhage (SAH) when using fluid-attenuated
inversion recovery (FLAIR).13,14
Fluid-attenuated inversion recovery imaging improves the
conspicuity of focal gray matter abnormalities (eg, contusions), white matter abnormalities, shear injuries, and SAH
by eliminating (or “nulling”) the bright cerebrospinal fluid
signal. Sagittal and coronal FLAIR images are particularly
helpful in the detection of diffuse axonal injury (DAI) involving the corpus callosum and the fornix, two areas that are
difficult to evaluate on routine T2-weighted images. On a
more cautious note, abnormal high signal in the sulci and
cisterns of ventilated patients receiving a high inspired oxygen fraction (⬎0.60) have been observed and should not be
mistaken for hemorrhage.15
Gradient-echo imaging is highly sensitive to local magnetic
Scalp Injury
When reviewing CT scans for head trauma, begin by examining the extracranial structures for evidence of soft-tissue
injury and/or radio-opaque foreign bodies. Scalp injury is a
reliable indication of the site of impact. Scalp injury includes
soft-tissue lacerations, subgaleal hematoma, cephalohematoma, and residual foreign bodies. The subgaleal hematoma
is far and away the most common manifestation of scalp
injury. It can be recognized as focal soft-tissue swelling located beneath the subcutaneous fibrofatty tissue and above
the temporalis muscle and calvarium (Fig. 1).
Skull Fractures
Nondisplaced linear fractures of the calvarium can be difficult to detect on CT when the fracture plane is parallel to the
plane of section. Fortunately, isolated linear skull fractures
usually do not require treatment unless they are associated
with an epidural hematoma. Surgical treatment is usually
only indicated for depressed and compound skull fractures,
both of which are easily detectable on CT (Fig. 2). Depressed
skull fractures can be associated with an underlying contusion; therefore, attention to the subadjacent parenchyma is
essential (Fig. 14).
Thin-section (1-mm) CT using a bone algorithm is recommended for the evaluation of fractures in the skull base, orbit,
or facial bones. Thin sections are helpful for evaluating the
degree of comminution and depression of bone fragments.
Imaging of head trauma
179
Figure 2 Depressed skull fracture with an associated acute epidural hematoma (EDH). (A) Axial CT scan displayed in
“bone window” shows a right frontal depressed skull fracture (arrowhead). Skull fractures can be associated with an
underlying epidural hematoma (B, arrow) and/or contusion, especially depressed comminuted fractures.
MDCT with 3D reconstruction can elegantly display complex
fractures and they can be generated expeditiously with sophisticated algorithms.19
Temporal Bone Fractures
Pneumocephalus, opacification of the mastoid air cells, and
fluid in the middle ear cavity should always raise concern for
a temporal bone fracture. Thin-section (1 to 1.5 mm) axial
and direct coronal CT imaging with bone algorithm reconstruction is recommended for the evaluation of temporal
bone fractures. With MDCT, thinner section axial imaging
can be performed, and coronal reformats may be adequate for
interpretation without the need for direct coronal imaging,
which can be difficult in intubated or mentally altered patients, or in patients with suspected cervical spine injury.
Fractures are classified as longitudinal or transverse, depending on their orientation relative to the long axis of the
petrous bone. Longitudinal fractures parallel the long axis of
the petrous pyramid, and transverse fractures are perpendicular to the long axis of the petrous bone. Longitudinal temporal bone fractures (Fig. 3) usually result from direct impacts to the side of the head. They represent more than 70%
of temporal bone fractures.20 Complications include conductive hearing loss from dislocation or fracture of the ossicles,
Figure 3 Longitudinal temporal bone fracture. Axial CT scan shows
a longitudinal left temporal bone fracture (arrowheads) with opacification of the mastoid air cells. Diastasis of the left lambdoid suture
(open arrow) and fractures of the sphenoid sinus (curved arrow)
and left lateral orbital wall (arrow) are also present. (Reprinted with
permission from Gean AD: Imaging of head trauma. Philadelphia,
PA, William & Wilkins-Lippincott, 1994, p 63.)
T.H. Le and A.D. Gean
180
Figure 4 Diagram of the subdural hematoma (SDH) and EDH. The EDH is located above the outer dural layer (ie, the
periosteum), and the SDH is located beneath the inner (meningeal) dural layer. The EDH does not cross sutures. The
SDH does not cross the falx or the tentorium. (Reprinted with permission from Gean AD: Imaging of head trauma.
Philadelphia, PA, Williams & Wilkins-Lippincott, 1994, p 76.)
otorhinorrhea from CSF leak, and facial nerve palsy. Transverse temporal bone fractures typically derive from direct
impacts to the occiput or frontal region. Complications include sensorineural hearing loss, vascular injury, and perilymphatic fistula. Vertigo, nystagmus, and facial palsy are
also common complications. Mixed or complex temporal
bone fractures involve a combination of fracture planes. They
typically result from severe crushing blows to the skull. Patients with mixed temporal bone fractures have a high incidence of associated intracranial injury.
tached at sutural margins (Fig. 4). However, at the vertex,
where the periosteum that forms the outer wall of the sagittal
sinus is not tightly attached to the sagittal suture, the EDH
can cross midline. An important imaging finding that predicts rapid expansion of an arterial EDH is the presence of
low-attenuation areas within the hyperdense hematoma (the
Primary Head Injury
Extra-Axial Injury
The epidural hematoma (EDH) occurs in the potential space
located between the inner table of the skull and the dura (Fig.
2). The developing hematoma dissects the dura from the
inner table of the skull, forming an ovoid mass that displaces
the adjacent brain. They frequently result from a skull fracture, usually in the temporal squamosa, that disrupts the
middle meningeal artery.10 In children, they may occur from
stretching or tearing of meningeal arteries without an associated fracture. Venous EDHs are less common than arterial
EDHs and tend to occur at three classic locations: the posterior fossa from rupture of the torcula or transverse sinus; the
middle cranial fossa from disruption of the sphenoparietal
sinus; and the vertex from injury to the superior sagittal
sinus.21 Venous EDHs can be difficult to diagnose on axial
imaging but can be readily confirmed on coronal reformatted
images.
On CT, the acute EDH appears as a well-defined, hyperdense, biconvex, extra-axial collection (Fig. 2). It is usually
associated with an overlying skull fracture. Mass effect with
sulcal effacement and midline shift is frequently seen. Because the EDH is located in the potential space between the
dura and inner table of the skull, it rarely crosses cranial
sutures because the periosteal layer of the dura is firmly at-
Figure 5 EDH “Swirl sign.” Axial CT image shows a heterogenous,
mottling appearance (“swirl sign”), within a biconvex, extra-axial
fluid collection (black arrow). The heterogenous density within this
EDH is secondary to mixing of hyperacute (low attenuation) with
acute (high attenuation) blood. Left frontal subarachnoid hemorrhage (SAH) is also noted (arrowheads).
Imaging of head trauma
Figure 6 Rebleeding during evolution of a SDH (surgically confirmed). Axial CT image demonstrates a large, right, holohemispheric, extra-axial, fluid collection with mixed hyperdense (acute)
and isodense (hyperacute) blood. An acute-on-chronic SDH would
have a similar appearance.
so-called “swirl sign”), thought to represent active bleeding
(Fig. 5).22,23 It is an ominous sign that needs to be followed
closely.
Subdural hematomas (SDH) are generally venous in origin.
They usually arise from laceration of bridging cortical veins
during sudden head deceleration. Occasionally, they may
also result from disruption of penetrating branches of superficial cerebral arteries. Because the inner dural layer and
arachnoid are not firmly attached, SDHs are frequently seen
layering along the entire hemispheric convexity from the anterior falx to the posterior falx (Figs. 6 to 10). The SDH is
uncommonly seen in association with DAI. Because of the
prominent extra-axial space in the elderly resulting from cerebral atrophy, increased motion between the brain parenchyma and the calvarium is permitted, and an increased incidence of SDH in these patients has been observed. Other
causes of the SDH include rapid decompression of obstructive hydrocephalus, and injury to pial vessels and pacchionian granulations. In rapid decompression of the hydrocephalus, the brain surface recedes from the dura quicker than the
brain parenchyma can re-expand after being compressed by
the distended ventricles.
On CT, the acute SDH appears as hyperdense, homogenous, crescent-shaped, extra-axial collection (Fig. 1). Most
are supratentorial and located along the convexity. They are
also frequently identified along the falx and tentorium. Because the SDH is often associated with parenchymal injury,
181
the degree of mass effect seen frequently appears more severe
relative to the size of the collection.
The attenuation (density) of an acute SDH initially increases because of clot retraction. The acute SDH is hyperdense, measuring 50 to 60 Hounsfield Units (HU), relative to
normal brain, which measures 20 to 30 HU. The density will
progressively decrease as protein degradation occurs within
the hematoma. Rebleeding during evolution of a SDH appears as a heterogeneous mixture of fresh blood and partially
liquefied hematoma (Fig. 6). A sediment level or “hematocrit
effect” may be seen either from rebleeding or in patients with
clotting disorders (Fig. 7). The chronic SDH has a low-attenuation value similar to, but slightly higher than, CSF (Fig. 8).
It can be difficult to distinguish from prominent subarachnoid space in patients with cerebral atrophy. In these patients, intravenous contrast administration can be helpful by
demonstrating an enhancing capsule or displaced cortical
veins.
A small thin convexity SDH can be difficult to appreciate
adjacent to the hyperdense skull unless the images are viewed
with a “wide window.” With most viewings and interpretations now performed on computer workstations, the radiologist can adjust the window setting readily to avoid this potential pitfall (Fig. 9). During the transition from the acute to
the chronic SDH, an isodense phase occurs, usually between
Figure 7 SDH with hematocrit level. Axial CT image shows a large,
right, holohemispheric, extra-axial fluid collection. The collection
has a hyperdense “sediment” level (*). The attenuation gradient is
secondary to the presence of hemorrhage of different ages. There is
effacement of the frontoparietal sulci and mild midline shift resulting from the SDH mass effect.
182
Figure 8 Chronic SDH and hydrocephalus. Axial CT image demonstrates a left frontal extra-axial fluid collection (*) that has a lower
attenuation than the brain parenchyma, consistent with chronic
SDH. The left lateral ventricle is compressed due to the mass effect
(white arrowhead) and there is subfalcine herniation (black arrowhead). Mild dilation of the right lateral ventricle is compatible with
noncommunicating obstructive hydrocephalus.
Figure 9 Isodense subacute SDH. During the transition from acute to
chronic SDH, an isodense phase occurs. At this stage, the SDH
(arrowhead) can be difficult to discriminate from the adjacent parenchyma.
T.H. Le and A.D. Gean
Figure 10 Appearance of the chronic SDH on MRI. The chronic SDH
is always slightly higher in signal intensity than CSF on T1weighted, fluid attenuation inversion recovery (FLAIR) imaging and
proton-density MR images. It is hypointense to gray and white matter on T1-weighted images (A), and hyperintense to brain parenchyma on T2-weighted images (B-D).
1 to 3 weeks after the acute event, depending on the patient’s
hematocrit level, clotting capability, and presence or absence
of rebleeding. Recognition of indirect imaging findings, such
as effacement of sulci, displacement of gray matter with white
matter “buckling,” and midline shift on a noncontrast CT
scan can avoid this common pitfall.
The MR appearance of a SDH also evolves over time, depending on the biochemical state of hemoglobin. The acute
SDH is isointense to brain on T1-weighted and hypointense
on T2-weighted images. During the subacute phase, when
the subdural hematoma may be isodense or hypodense on
CT scans, T1-weighted images demonstrate high signal intensity due to the presence of methemoglobin in the subdural
collection. The chronic SDH appears hypointense on T1weighted and hyperintense on T2-weighted images relative
to normal brain (Fig. 10). The signal intensity of chronic SDH
is slightly higher than CSF signal intensity on T1-weighted,
FLAIR, and proton-density T2-weighted imaging. Because of
its multiplanar capability, MR is useful in identifying convexity and vertex hematomas that might not be detected on axial
CT scans because of the similar attenuation of the adjacent
bone.
Traumatic subarachnoid hemorrhage (SAH) can result from
(a) the disruption of small pial vessels; (b) extension into the
subarachnoid space by a contusion or hematoma; or (c) diffusion of intraventricular hemorrhage. SAH is very common
with TBI, but it rarely causes mass effect. On CT, SAH appears as linear or serpentine areas of high attenuation that
conform to the morphology of the cerebral sulci and cisterns
Imaging of head trauma
Figure 11 SAH and IVH secondary to aneurysm rupture. (A) On CT,
acute SAH appears as linear or serpentine areas of high attenuation
that conform to the morphology of the sulci and cisterns. Although
SAH is common with TBI, if the SAH is identified in certain locations
(Sylvian, interhemispheric, or basilar cisterns), an underlying ruptured aneurysm must also be considered (B, C, D). Contrast-enhanced CT angiography (CTA) is particularly helpful in this situation. In this case, CTA shows a right middle cerebral artery
trifurcation aneurysm (D). The aneurysm appears pointed, suggesting recent rupture. Intraventricular hemorrhage (IVH) is seen as a
high attenuation fluid level within the occipital horn of the left
lateral ventricle (B, white arrowhead).
(Fig. 11A-C). Common sites for SAH include the sylvian and
interpeduncular cisterns. The greatest accumulation of SAH
tends to occur contralateral to the site of impact (ie, contrecoup). Subarachnoid hemorrhage along the convexity or tentorium can be difficult to differentiate from a SDH. A useful
clue is the extension of the SAH into adjacent sulci. Occasionally, “effacement” of sulci due to the presence of intrasulcal SAH may be the only imaging clue of the presence of SAH.
In patients who are found unconscious after an unwitnessed
event, the detection of SAH in key cisterns (basilar, sylvian,
and circle of Willis) may indicate a ruptured aneurysm,
rather than trauma. In such cases, contrast-enhanced CT angiography is recommended as the next step in the patient
evaluation (Fig. 11D).
Acute subarachnoid hemorrhage may be more difficult to
detect on conventional MR than on CT because it can be
isointense to brain parenchyma on T1- and T2-weighted images. However, FLAIR has been shown to be more sensitive
than CT in detecting acute subarachnoid hemorrhage in animal model, especially when a high volume (1 to 2 mL) is
present.14 Subacute subarachnoid hemorrhage may be better
appreciated on MR because of its high signal intensity when
183
the blood is isointense to CSF on CT. Chronic hemorrhage on
MR scans may show hemosiderin staining in the subarachnoid space, which appears as areas of decreased signal intensity on T1- and T2-weighted sequences (“superficial hemosiderosis”). Hemosiderins are best detected on gradient-echo
T2*-weighted images. SAH may lead to communicating hydrocephalus by impeding CSF resorption at the level of
arachnoid villi.
Traumatic intraventricular hemorrhage (IVH) can also occur
by one of three mechanisms. First, it can result from rotationally induced tearing of subependymal veins along the surface
of the ventricles. Another mechanism is by direct extension of
a parenchymal hematoma into the ventricular system. Third,
IVH can result from retrograde flow of SAH into the ventricular system via the fourth ventricular outflow foramina (the
reverse mechanism in which IVH can extend into the subarachnoid space). Patients with IVH are at risk for developing
both communicating and noncommunicating hydrocephalus secondary to obstruction at the level of the arachnoid villi
or aqueduct, respectively. They are also at risk of ependymitis
from the irritant effects of the blood.
On CT, IVH usually appears as CSF-hyperdense fluid
level, layering within the ventricular system (Fig. 11B). Indeed, tiny collections of increased density layering in the
occipital horns may be the only clue to IVH. Occasionally, the
IVH may appear “tumefactive” as a cast within the ventricle.
Figure 12 DAI on CT and MR. In Grade II DAI, the corpus callosum
is injured. Note the low attenuation within the splenium of the
corpus callosum on CT (A). The FLAIR image demonstrates abnormal hyperintensity in the same region (B). The diffusion-weighted
image (DWI) shows a focus of bright signal intensity (C); the apparent diffusion coefficient (ADC) was correspondingly reduced (not
shown). The T2*-weighted image gradient-echo image shows scattered foci of susceptibility staining (signal loss) (D).
184
Figure 13 Cortical contusion near the skull base. (A) Sagittal T1weighted image demonstrates abnormal T1 shortening (increased
signal intensity) involving the right orbitofrontal lobe. There is associated decreased signal on the gradient-echo images due to the
presence of hemosiderin (B). The FLAIR (C) and T2-weighted (D)
images demonstrate mixed heterogenous signal due to a combination of methemoglobin, hemosiderin, and edema. Small bilateral
temporal contusions are also noted.
Intra-Axial Injury
Diffuse Axonal Injury (DAI) results from rotational acceleration and deceleration forces that produce shearing deformations of brain tissue. Clinically, DAI is characterized by loss or
severe impairment of consciousness beginning at the moment of direct impact. DAI in the chronic stage can result in
overwhelming cognitive and psychiatric problems. The affected areas of the brain are often distant from the site of
direct impact. DAI, which occurs in about half of all severe
head trauma cases, is of special interest because it tends to be
underdiagnosed by current imaging techniques.
MRI is clearly superior to CT for detecting DAI, and therefore, provides increased success at explaining neurologic deficits after trauma and predicting long-term outcome. Even
with MR, the incidence of DAI is thought to be underestimated. Newer imaging methods, such as DWI and DTI with
3D tractography, have shown potential in improving the detection of white matter injury in both acute and chronic
DAI.17,24,25
CT is mostly limited to depicting the small focal hemorrhages associated with axonal shearing injury. Common findings include small, petechial hemorrhages at the gray–white
junction of the cerebral hemispheres and corpus callosum.
Unfortunately, only about 20% of DAI lesions contain sufficient hemorrhage to be detectable on CT. Through its better
T.H. Le and A.D. Gean
sensitivity to blood products, gradient-echo T2*-weighted
MR demonstrates more DAI lesions than CT.17 However,
pathologic studies demonstrate that only a minority of DAI
lesions is associated with hemorrhage.26 Spin-echo T2weighted MR techniques, especially FLAIR imaging, can detect many nonhemorrhagic foci of DAI but still underestimate the true extent of traumatic white matter damage.27-29
On MR, nonhemorrhagic DAI lesions appear as multiple
small foci of increased signal on T2-weighted images (Fig.
12) and as decreased signal on T1-weighted images within
the white matter. Petechial hemorrhage causes a central hypointensity on T2-weighed images and hyperintensity on T1weighted images as a result of intracellular methemoglobin in
the subacute stage. The conspicuity of DAI on MR eventually
diminishes as the damaged axons degenerate and the edema
resolves. Residual findings may include nonspecific atrophy,
gliosis, or hemosiderin staining, which can persist indefinitely on gradient-echo T2*-weighted images (Fig. 12D).
The location of the of DAI tends to correlate with the
severity of the trauma. Grade I DAI involves only the peripheral gray–white junctions of the lobar white matter. The parasagittal regions of the frontal lobes and periventricular regions of the temporal lobes are commonly affected. Patients
with more severe trauma have DAI involving the lobar white
matter as well as the corpus callosum, particularly the posterior body and splenium (Fig. 12). The corpus callosum is
susceptible to DAI because the falx prevents displacement of
the cerebral hemispheres. In severe DAI, the dorsolateral
midbrain, in addition to the lobar white matter and corpus
callosum, is affected.
The cortical contusions is a focal brain lesion primarily involving superficial gray matter, with relative sparing of the underlying white matter. Regions of the brain that are in close contact
with the rough surface on the inner skull table are commonly
affected. These areas include the orbitofrontal and temporal
lobes, and less so, the parasagittal convexity. The temporal lobes
above the petrous bone or posterior to the greater sphenoid
Figure 14 Contusion at the margins of depressed skull fractures. (A)
Axial CT image displayed with “bone windowing” reveals a nondepressed right temporal skull fracture (arrowhead). (B) Using “brain
windowing,” an adjacent right frontal surface contusion is revealed
(arrowhead). A small left temporal SDH with associated pneumocephalus is also noted (arrows).
Imaging of head trauma
185
renchymal blood vessels. Because the bleeding occurs into areas
of relatively normal brain, intracerebral hematomas tend to have
less surrounding edema than cortical contusions. Most traumatic intracerebral hematomas are located in the frontotemporal white matter. Involvement of the basal ganglia has been described; however, hemorrhage in basal ganglia should alert the
radiologist that an underlying hypertensive bleed may be the
culprit (Fig. 16). Intracerebral hematomas are often associated
with skull fractures and other primary neuronal lesions, including contusions and DAI, especially in patients who are unconscious at the time of injury. Symptoms typically result from the
mass effect associated with an expanding hematoma. Indeed,
delayed hemorrhage is a common cause of clinical deterioration
during the first several days after head trauma.
Vascular Injury
Figure 15 Nonhemorrhagic contusion on CT. Axial CT image demonstrates an ill-defined area of low attenuation within the right
temporal lobe (arrow). The low attenuation is due to edema from
the nonhemorrhagic contusion. The inferior temporal lobe is commonly injured because of its proximity to the petrous ridge.
wing, and the frontal lobes above the cribriform plate, planum
sphenoidale, and lesser sphenoid wing are commonly affected
(Fig. 13). Contusions are associated with a better prognosis than
DAI, unless they are accompanied by brainstem injury or significant mass effect. Contusions can also occur at the margins of
depressed skull fractures (Fig. 14). The cerebellum is involved
less than 10% of the time.30
On CT, nonhemorrhagic contusions are difficult to detect
initially until the development of associated edema (Fig. 15).
Hemorrhagic contusions are more easily identified and appear as foci of high attenuation within superficial gray matter
(Fig. 14). They may be surrounded by larger areas of low
attenuation from the associated vasogenic edema. As the contusion evolves, the characteristic “salt and pepper” pattern of
mixed areas of hypodensity and hyperdensity becomes more
apparent. Contusions with severe mass effect may require
surgical decompression to prevent secondary injury.
On MRI, contusions appear as ill-defined areas of variable
signal intensity on both T1- and T2-weighted images, depending on the age of the lesions (Fig. 13). They are limited to the
surface and often have a “gyral” morphology. Hemosiderin from
an old contusion leads to decreased signal intensity on T2- and
especially T2*-weighted images, especially at higher field
strengths. The signal loss can persist indefinitely and serves as an
important marker of prior hemorrhage.
The intracerebral hematoma is the most common cause of
clinical deterioration in patients who have experienced a lucid
interval after the initial injury. In contrast to the common contusion described above, the intracerebral hematoma is due to
shear-induced hemorrhage from the rupture of small intrapa-
Vascular injuries are mentioned here because they are causes of
both intra- and extra-axial injuries, including the cause of hematomas and subarachnoid hemorrhages. Additional traumatic
vascular injuries include the arterial dissection, pseudoaneurysm, and arteriovenous fistula. Arterial injuries are usually related to skull base fractures. The most often injured artery is the
internal carotid artery, especially at sites of fixation, where it
enters the carotid canal at the base of the petrous bone, and at its
Figure 16 Intracerebral hematoma. (A) Axial CT shows a hyperdense
lesion within the right putamen. The surrounding low attenuation is
due to vasogenic edema. (B) On the T1-weighted image, the periphery of the hematoma is hyperintense due to the presence of methemoglobin. (C) The hematoma is also hyperintense on the T2weighted image due the presence of extracellular methemoglobin.
(D) Gradient-echo T2*-weighted image shows a surrounding hemosiderin rim, which is dark due to local magnetic susceptibility
inhomogeneity.
T.H. Le and A.D. Gean
186
Figure 17 Carotid dissection. T1-weighted MR image performed
with fat suppression shows a hyperintense crescent beneath the
adventitia of the left internal carotid artery. The high signal represents a subacute intramural hematoma.
exit from the cavernous sinus beneath the anterior clinoid process.
MR findings of vascular injury include (a) the presence of an
intramural hematoma, which is best seen on T1-weighted with
fat suppression (Fig. 17); (b) intimal flap with dissection; and (c)
absence of a normal vascular flow void secondary to slow flow or
occlusion. An associated parenchymal infarction supplied by
the injured vessel may also be seen. Conventional angiograms
remain the gold standard for confirmation and delineation of the
vascular dissection and may also show spasm or pseudoaneurysm formation. However, magnetic resonance angiography
and MDCT angiography serve as important screening tools in
the evaluation of patients with suspected vascular injury.
The acquired carotid cavernous fistula (CCF) typically results
from full-thickness arterial injury. The injury leads to communication between the cavernous portion of the internal carotid
artery and the surrounding venous plexus, resulting in venous
engorgement of the cavernous sinus (Fig. 18) and its draining
branches: the ipsilateral superior ophthalmic vein and inferior
petrosal sinus. Skull base fractures, especially those involving
the sphenoid bone, should alert the radiologist to search for
associated cavernous carotid injury. Another cause of CCF is the
rupture of a cavernous carotid aneurysm. On imaging, CCF can
present as an enlarged superior ophthalmic vein, cavernous sinus, and/or petrosal sinus. Other findings include proptosis,
preseptal soft-tissue swelling, and extraocular muscle enlargement. The findings may be bilateral because venous channels
connect the cavernous sinuses. Again, definitive diagnosis often
requires selective carotid angiography with rapid filming to
demonstrate the site of communication. Patients can present
with findings weeks or even months after the initial trauma.
Therefore, a CCF can be overlooked if a detailed clinical history
and ophthalmic examination is not performed.
Another traumatic vascular injury is the dural fistula, usually
caused by laceration of the middle meningeal artery with resultant meningeal artery to meningeal vein fistula formation. The
dural fistula generally drains via the meningeal veins; therefore,
they rarely lead to the formation of an EDH. Patients are often
asymptomatic or present with nonspecific complaints such as
tinnitus.
Secondary Head Injury
Acute
Diffuse cerebral swelling arises from an increase in cerebral
blood volume (hyperemia), vasogenic edema, or an increase
in tissue fluid (cerebral or cytotoxic edema). Imaging demonstrates effacement of the cerebral sulci and cisterns and
Figure 18 Acquired carotid cavernous
fistula (CCF). CTA demonstrates asymmetric enhancement of the right cavernous sinus secondary to abnormally
enlarged venous channels (black arrows).
Imaging of head trauma
Figure 19 Diffuse cerebral edema. Noncontrast CT scan in an infant
with diffuse cerebral edema following strangulation. There is a diffuse decrease in attenuation of the cerebral hemispheres with loss of
gray–white differentiation. Sparing of the brainstem and cerebellum
causes these structures to appear dense relative to the rest of the
brain. An acute SDH overlies the tentorium (arrows).
compression of the ventricles. Hyperemia is thought to be the
result of cerebral dys-autoregulation, and cerebral edema
usually occurs secondary to tissue hypoxia. In cerebral
edema, the gray–white differentiation is lost, which is in contrast to cerebral hyperemia where the gray–white differentiation is preserved. The cerebellum and brainstem are usually
spared in cerebral edema and may appear hyperintense relative to the affected cerebral hemispheres (Fig. 19).
Brain herniation occurs secondary to mass effect produced
by other causes. In subfalcine herniation, the most common
form of herniation, the cingulate gyrus is displaced across the
midline under the falx cerebri. Compression of the ipsilateral
ventricle due to mass effect and enlargement of the contralateral ventricle due to obstruction of the foramen of Monro can
be seen on imaging (Fig. 8). The anterior cerebral arteries
Figure 20 CSF leak. Coronal CT cisternography performed before (A) and after
(B) intrathecal contrast demonstrates a
meningocele and suspected bony defect
involving the left cribriform plate with
contrast leakage into the upper left nasal
vault (arrowhead).
187
(ACA) are displaced to the contralateral side, trapping the
callosomarginal branches of the ACA, and may lead to ACA
infarction. In uncal herniation, the medial temporal lobe is
displaced over the free margin of the tentorium. Effacement
of the ambient and lateral suprasellar cisterns is an important
clue of the presence of uncal herniation. In severe cases,
displacement of the brainstem can cause compression of the
contralateral cerebral peduncle against the tentorium (“Kernohan’s notch”), leading to peduncular infarction or hemorrhage. Occasionally, the third cranial nerve is compressed;
these patients present with a “blown pupil” and ipsilateral
hemiparesis. In transtentorial herniation, the brain herniates
either upward or downward. Upward herniation typically
occurs with large posterior fossa hematomas that displace
portions of the cerebellum and vermis through the tentorial
incisura. The mass effect of the hematoma can also cause
downward herniation of the cerebellar tonsils through the
foramen magnum. Downward herniation of the cerebrum
manifests as effacement of the suprasellar and perimesencephalic cisterns. Inferior displacement of the pineal calcification is an additional imaging clue for the presence of downward herniation.
Infarction or ischemia can complicate TBI as a result of increased intracranial pressure or mass effect on cerebral vasculature by herniation or hematoma. In addition to ACA
infarction secondary to the subfalcine herniation described
above, uncal herniation and tonsillar herniation can cause
ischemia in the territory of the posterior cerebral artery and
the posterior inferior cerebellar artery distributions, respectively.
Chronic
As mentioned above, traumatic hydrocephalus occurs secondary to impaired CSF reabsorption at the level of the arachnoid
villi or secondary to obstruction of the cerebral aqueduct and
4th ventricular outflow by SAH. Mass effect from cerebral
herniation or a hematoma can also cause noncommunicating
hydrocephalus via compression of the aqueduct and ventricular outflow foramina (Fig. 8).
Encephalomalacia is a common, but nonspecific, sequelae
of prior parenchyma injury. It may be clinically asymptomatic, but it can also be a potential seizure focus. On CT, the
imaging appearance consists of an area of low attenuation
T.H. Le and A.D. Gean
188
it is fast, is widely available, and can easily accommodate
life-support and monitoring equipment. CT can accurately
identify space-occupying lesions, acute hemorrhages, mass
effect, midline shift, hydrocephalus, ischemia, and herniation. MRI is indicated for patients with acute TBI when the
neurologic findings are unexplained by CT. MRI is also the
modality of choice for subacute or chronic injury. Advances
in MR methods, such as diffusion-weighted imaging, further
improve the neuroradiological evaluation of traumatic brain
injury and enhance our understanding of the pathophysiological manifestations of head trauma.
Acknowledgments
Figure 21 Leptomeningeal cyst. Lateral radiograph of the skull demonstrates a large lytic (lucent) lesion with scalloped margins located
within the left parietal bone (arrows). (Reprinted with permission
from Gean AD: Imaging of head trauma. Philadelphia, PA, Williams
& Wilkins-Lippincott, 1994, p 384.)
with volume loss. It typically follows CSF signal intensity on
both CT and MR, except for the areas of gliosis, which appears as low intensity on T1- and high intensity on T2weighted images. Encephalomalacia within the orbitofrontal
(especially the gyrus rectus) and anteroinferior temporal
lobes is characteristic of remote traumatic injury.
A CSF leak usually results from a dural tear secondary to a
skull base fracture. When communication between the subarachnoid space and middle ear occurs in association with a
ruptured tympanic membrane, CSF otorrhea may be noted.
CSF rhinorrhea occurs when there is communication between the subarachnoid space and the paranasal sinuses. CSF
leaks are often difficult to localize and can lead to recurrent
meningeal infection. Radionuclide cisternography is highly
sensitive for the presence of CSF extravasation. However, CT
scanning with intrathecal contrast is required for detailed
anatomic localization of the defect (Fig. 20).
The leptomeningeal cyst or “growing fracture” is a pediatric
lesion that is also caused by a tear in the dura in association
with a calvarial defect. The dural defect allows expansion of
the arachnoid at the site of the bony defect, presumably as a
result of CSF pulsations. Such expansion leads to progressive, slow widening of the skull defect or suture. The leptomeningeal cyst appears as a lytic skull defect on CT or plain
skull films, which can enlarge over time (Fig. 21).
Summary
The goal of imaging in the management of head trauma is to
identify treatable injuries to prevent secondary damage. CT
continues to be the modality of choice in the evaluation of
acute head injury. CT is preferred in the acute setting because
We thank the residents, fellows, and attendings from the
Neuroradiology Section of the Department of Radiology,
University of California, San Francisco for continuing effort
in submitting interesting cases to the teaching file server
(http://tfserver.ucsf.edu). Some of the cases presented in this
article were the products of their effort.
References
1. MacDorman MF, Minino AM, Strobino DM, et al: Annual summary of
vital statistics—2001. Pediatrics 110(6):1037-1052, 2002
2. Lewin J, Sumners D: Anorexia due to brain injury. Brain Inj 6(2):199201, 1992
3. National Center for Injury Prevention and Control. Injury Fact Book
2001-2002. Atlanta, GA: Centers for Disease Control and Prevention;
2001
4. Nirula R, Kaufman R, Tencer A: Traumatic brain injury and automotive
design: making motor vehicles safer. J Trauma 55(5):844-848, 2003
5. Teasdale G, Jennett B: Assessment of coma and impaired consciousness: a practical scale. Lancet 13:81-84, 1974
6. Bell RS, Loop JW: The utility and futility of radiographic skull examination for trauma. N Engl J Med 284(5):236-239, 1971
7. Hackney DB: Skull radiography in the evaluation of acute head trauma:
a survey of current practice. Radiology 181(3):711-714, 1991
8. Masters SJ: Evaluation of head trauma: efficacy of skull films. AJR Am J
Roentgenol 135(3):539-547, 1980
9. Adams J: Pathology of nonmissile head injury. Neuroimaging Clin
North Am 1:397-410, 1991
10. Zee CS, Go JL: CT of head trauma. Neuroimaging Clin North Am
8(3):525-539, 1998
11. Gentry LR, Godersky JC, Thompson B, et al: Prospective comparative
study of intermediate-field MR and CT in the evaluation of closed head
trauma. AJR Am J Roentgenol 150(3):673-682, 1988
12. Orrison WW, Gentry LR, Stimac GK, et al: Blinded comparison of
cranial CT and MR in closed head injury evaluation. AJNR Am J Neuroradiol 15(2):351-356, 1994
13. Noguchi K, Ogawa T, Seto H, et al: Subacute and chronic subarachnoid
hemorrhage: diagnosis with fluid-attenuated inversion-recovery MR
imaging. Radiology 203(1):257-262, 1997
14. Woodcock RJ Jr, Short J, Do HM, et al: Imaging of acute subarachnoid
hemorrhage with a fluid-attenuated inversion recovery sequence in an
animal model: comparison with non-contrast-enhanced CT. AJNR
Am J Neuroradiol 22(9):1698-1703, 2001
15. Frigon C, Jardine DS, Weinberger E, et al: Fraction of inspired oxygen
in relation to cerebrospinal fluid hyperintensity on FLAIR MR imaging
of the brain in children and young adults undergoing anesthesia. AJR
Am J Roentgenol 179(3):791-796, 2002
16. Le TH, Mukherjee P, Manley GT, et al: Automated detection of traumatic white matter injury using voxel-based morphometry of diffusion
tensor images: a 3T study with parallel imaging. American Society of
Neuroradiology 43rd Annual Meeting, May 23-27, 2005, Toronto, Ontario, Canada, Paper 374.
17. Huisman TA, Sorensen AG, Hergan K, et al: Diffusion-weighted imag-
Imaging of head trauma
18.
19.
20.
21.
22.
23.
24.
ing for the evaluation of diffuse axonal injury in closed head injury.
J Comput Assist Tomogr 27(1):5-11, 2003
Liu AY, Maldjian JA, Bagley LJ, et al: Traumatic brain injury: diffusionweighted MR imaging findings. AJNR Am J Neuroradiol 20(9):16361641, 1999
Soto JA, Lucey BC, Stuhlfaut JW, et al: Use of 3D imaging in CT of the
acute trauma patient: impact of a PACS-based software package. Emerg
Radiol 11(3):173-176, 2005
Gentry LR: Temporal bone trauma. Neuroimaging Clin North Am
1:319-340, 1991
Gean AD, Kates RS, Lee S: Neuroimaging in head injury. New Horiz
3(3):549-561, 1995
Al-Nakshabandi NA: The swirl sign. Radiology 218(2):433, 2001
Greenberg J, Cohen WA, Cooper PR: The “hyperacute” extraaxial intracranial hematoma: computed tomographic findings and clinical significance. Neurosurgery 17(1):48-56, 1985
Arfanakis K, Haughton VM, Carew JD, et al: Diffusion tensor MR imaging in diffuse axonal injury. AJNR Am J Neuroradiol 23(5):794-802,
2002
189
25. Le TH, Mukherjee P, Henry RG, et al: Diffusion tensor imaging with
three-dimensional fiber tractography of traumatic axonal shearing injury: an imaging correlate for the posterior callosal “disconnection”
syndrome: case report. Neurosurgery 56(1):189, 2005
26. Adams JH, Graham DI, Gennarelli TA, et al: Diffuse axonal injury in
non-missile head injury. J Neurol Neurosurg Psychiatry 54(6):481483, 1991
27. Ashikaga R, Araki Y, Ishida O: MRI of head injury using FLAIR. Neuroradiology 39:239-242, 1997
28. Hesselink JR, Dowd CF, Healy ME, et al: MR imaging of brain contusion: a comparative study with CT. AJNR Am J Neuroradiol 150:11331142, 1988
29. Mittl RL, Grossman RI, Hiehle JF, et al: Prevalence of MR evidence of
diffuse axonal injury in patients with mild head injury and normal head
CT findings. AJNR Am J Neuroradiol 15:1583-1589, 1994
30. Gentry LR: Head Trauma, in Atlas SW (ed): Magnetic resonance imaging of the brain and spine (ed 2). Philadelphia, PA, Lippincott-Raven,
1996, pp 611-647