This article will review the most recent developments in contrast-enhanced neuroimaging of CNS pathology. It has been approved for two (2) catergory I CME credits.
CME CATEGORY 1
This CME program is sponsored by an educational grant from
Bracco Diagnostics Inc.
The Institute for Advanced Medical Education designates this
continuing medical education activity for a maximum of 2 Category 1
credits toward the AMA Physician's Recognition Award. Each
physician should claim only those credits that he/she actually
spent in the activity.
To go directly to the quiz please
This CME article consists of text and related images appearing
in this journal article and online at
You should read the articles and accompanying images, refer to the
references, and complete the self-evaluation quiz available online
There is no charge for participating in this program.
Estimated time for completion:
Date of release:
After completing this activity, the reader will:
* Be familiar with the newest technical advances in MRI as they
affect the CNS
* Know when contrast agents are useful in evaluation of CNS
disease by clinical MRI
* Be familiar with the latest applications for MRI in the
The central nervous system (CNS) was the first field of
application in which the utility of magnetic resonance imaging
(MRI) was demonstrated.
In the early 1980s it was shown that most lesions appeared bright
on T2-weighted images and dark on T1-weighted images on the basis
of their higher water content. The inherent contrast resolution
observed was much greater than that available with computed
tomography (CT). However, the lack of an available contrast agent
during the early days of MRI meant that the specificity for lesion
characterization was relatively low and that a contrast-enhanced CT
examination was still the standard imaging examination.
With the introduction of the first gadolinium (Gd) contrast
agents in the mid-1980s, the balance of power shifted toward MRI
for the evaluation of lesions in the CNS.
The subsequent introduction in the mid-1990s of new sequences, such
as fluid-attenuated inversion recovery (FLAIR), further increased
the potential applications of MRI to the extent that MRI can now
replace CT for most, if not all, diagnostic evaluations, including
On the other hand, CT is frequently still the primary imaging
modality used for emergency patients and for patients with
contraindications for MRI due to the presence of a pacemaker or a
ferromagnetic aneurysm clip.
Today, MRI--specifically contrast-enhanced MRI--is standard for
a full diagnostic evaluation of many CNS pathologies. Compared with
other diagnostic techniques, MRI permits better visualization of
anatomy and vessels. In consequence, MRI is of high value for the
detection and diagnosis of pathology. While the greater speed of CT
combined with technological advances (angiography, perfusion,
three-dimensional [3D] postprocessing) make this still very much a
valid technique, ongoing developments in functional MR imaging
(fMRI) and interventional neuroradiology are likely to lead to
further applications for contrast-enhanced MRI in neuroimaging.
This article will review the most recent developments in
contrast-enhanced neuroimaging of CNS pathology.
In the early days of MR neuroimaging, spin-echo and
gradient-echo sequences permitted the acquisition of excellent MR
images based on differences in soft-tissue contrast. Subsequent
developments in both hardware and software have led to significant
progress in MR neuroimaging, such that today a wide variety of
acquisition sequences is available for MR imaging of brain and
From the very beginning, the fundamental approach to MR imaging
of the CNS has centered upon the acquisition of T1- and T2-weighted
These sequences are less sensitive to susceptibility effects than
are gradient-echo techniques. Fast-spin-echo sequences, in
particular, allow the rapid acquisition of high-resolution
T2-weighted and proton-density-weighted images and can be combined
with inversion prepulses, as in the FLAIR sequences. The
T1-weighted spin-echo sequences provide crucial information for the
differential diagnosis of lesions, especially when they are used
following an injection of Gd contrast agents to reveal breakdown of
the blood-brain barrier (BBB). Proton-density
(intermediate-weighted sequences) images are frequently of value
for the evaluation of thrombus and for the delimitation of
perpendicular flow during the evaluation of aneurysms or
arteriovenous malformations (AVMs) before and after
The advent of gradient-echo sequences with short repetition time
(TR) allowed the acquisition of the first 3D MR images.
These sequences today are fundamental to numerous MR angiographic
techniques and remain the first choice for high-resolution 3D
T1-weighted imaging. The use of contrast-enhanced 3D imaging is
applicable in many situations, such as for the optimal detection of
brain metastases. Three-dimensional gradient-echo sequences are
also employed in the context of neuro-navigation surgery for which
3D MR stereotactic imaging is mandatory.
Echoplanar imaging (EPI) involves the use of a fast train of
gradient echoes and is often employed in T2* perfusion imaging for
the detection of hemorrhage
and in fMRI using blood-oxygen-leveldependent (BOLD) techniques.
Echoplanar imaging is essentially a readout module, similar to
spin-echo or fast-spin-echo imaging.
The readout module is preceded by a combination of 90° and 180°
radiofrequency (RF) pulses, which determine the root contrast, eg,
gradient echo, spin echo, or inversion recovery.
EPI sequences have been used for neuroimaging; however, single-shot
EPI is by far more prevalent. Until recently, one of the principal
drawbacks of EPI sequences has been a relatively high sensitivity
to susceptibility artifacts.
This is partially alleviated by multishot EPI, which results in
much less geometric distortion and less blurring, particularly in
the critical region of the brain stem. With single-shot EPI,
however, acquisition times on the order of 0.1 sec are now
achievable, which is sufficient to eliminate all motion artifacts.
Starting from a spin-echo EPI sequence with additional gradient
pulses, diffusion-weighted images (DWI) can be obtained.
Further developments include the use of new parallel imaging
techniques such as sensitivity encoding (SENSE)
to reduce the EPI single-shot echo train length by a factor of
about 2 to 3 (see below).
ECHOPLANAR DIFFUSION WEIGHTED IMAGING
Echoplanar diffusion imaging is classically employed for the
detection of acute ischemia based on the presence of cytotoxic
While this is the best known cause of high signal on diffusion
images, high signal is also seen in the case of acute
demyelination, in tumors with high nuclear cytoplasmic ratios, in
abscesses, and in encephalitis. Another potential mimicker of
ischemia is a phenomenon known as T2 shine through where the high
signal on the T2-weighted reference image (b
image) shines through to the diffusion image. In order to avoid
misdiagnosing restricted diffusion in such cases, apparent
diffusion coefficient (ADC) maps can be calculated.
One recent development (which may have applications for the
evaluation of white matter diseases,
and spinal cord pathology
) is diffusion tensor imaging (DTI). Since water diffuses faster
parallel to the white matter than perpendicular to it, normal white
matter has higher diffusion anisotropy than diseased white matter
or normal gray matter. This difference in anisotropy can be
observed on both DWI and ADC maps. Anisotropy is largest when the
white matter in the voxel is totally intact and is decreased when
the voxel involves damaged white matter or gray matter, which is
less anisotropic. Scalar parameters, such as fractional anisotropy
or relative anisotropy, can also be calculated using the three
diffusion coefficients (Dx, Dy, Dz).
These can be used to construct anisotropy maps of the entire
In DTI, images sensitive to diffusion in at least six directions
and one reference (b
) image are acquired. The postprocessing of these images amounts to
a reorientation of the x, y, and z axes within the voxel so that
the new z-axis for each voxel (called the principal eigenvector)
corresponds to the principal direction of diffusion within that
voxel. Since this direction is generally parallel to the main white
matter bundle in the voxel, the final images render a tractogram
demonstrating the actual location of the main white matter tracts
in the brain.
White matter tractography may be used in conjunction with fMRI
activation to demonstrate physical connectivity as well as temporal
correlation between areas of activation.
ECHOPLANAR PERFUSION IMAGING
Cerebral perfusion is defined as the delivery of nutrients and
oxygen via blood to brain parenchyma per unit mass (reported in
mL/100 g of tissue/min). Perfusion can be assessed using arterial
spin labeling (ASL) techniques.
With such techniques a slice of the brain is tagged with an RF
pulse and then, after a variable period of time, a slice downstream
is interrogated. Spins within a certain velocity range will
demonstrate changes in intensity that permit a relative
quantification of perfusion. With further refinement, ASL
techniques may be used for fMRI and for the evaluation of stroke
patients. However, while ASL methods permit a measure of the
cerebral blood flow (CBF), at 1.5 T the signal-to-noise ratio (SNR)
of this method is relatively poor.
Contrast-enhanced methods based on T2* bolus techniques or T1
permeability sequences are usually preferred, particularly for the
evaluation of stroke patients. With contrast-enhanced perfusion MR
imaging, hemodynamic parameters such as CBF, cerebral blood volume
(CBV), mean transit time (MTT), flow heterogeneity, extraction
fraction, and permeability surface area product can be derived from
imaging data involving the passage of an MR contrast agent through
the cerebrovascular system.
In echoplanar perfusion imaging a compact bolus of Gd is
injected and sequential imaging is performed of the entire brain
using T2*-weighted gradient-echo EPI or T2-weighted spin-echo EPI
techniques. Whereas the former technique usually requires a
standard 0.1 mmol/kg dose of a conventional Gd contrast agent and
reveals both capillaries and veins, the latter technique frequently
requires a double dose and reveals the capillaries only.
As the contrast agent bolus traverses the brain, the paramagnetic
Gd causes T2* shortening, which results in decreased signal within
a voxel. The signal intensity versus time curve initially dips and
then returns to the baseline. The area under the curve is
proportional to the relative cerebral blood volume (rCBV), which
can be determined by computer integration. It is also possible to
determine the midpoint of the curve in time (MTT) or the time to
peak, both of which are useful indicators of the time it takes the
blood to get to the specific voxel. By dividing the rCBV by the
MTT, the relative cerebral blood flow (rCBF) can be determined. The
modifier relative is necessary since a given voxel is being
compared with all the other voxels in the image. In order to
acquire an absolute measure of rCBF, an accurate arterial input
function is needed. By placing a cursor over the middle cerebral
artery, investigators have recently been able to achieve
which facilitates comparison with positron emission tomography
(PET) and xenon CT.
The rCBV is a measure of capillary density in the brain and in
brain tumors has been shown to increase with angiogenesis. Thus,
the rCBV is an appropriate parameter to distinguish between grade
II, grade III, and grade IV gliomas. Unfortunately, low-grade
(grade I) gliomas cannot be distinguished from normal brain at
MR spectroscopy, on the other hand, can distinguish low-grade
gliomas from normal brain on the basis of elevated choline and
decreased N-acetylaspartate (NAA).
With increasing malignancy, the choline/NAA ratio generally
increases. With treatment, eg, of lymphomas, it has been shown to
revert to normal.
At 3.0 T, it is quite likely that additional peaks will be
demonstrated that will be capable of additional characterization of
malignancy and response to treatment.
Functional MRI (fMRI) is based on the increase in blood flow to
a portion of the brain when it is activated by a specific task
and is not a technique for which contrast enhancement is required.
In fMRI, the amount of blood passing to a portion of activated
brain increases, increasing the relative concentrations of
oxyhemoglobin (oxygenated blood) to deoxyhemoglobin (deoxygenated
blood). Since deoxy-hemoglobin is paramagnetic and has a short T2,
it is dark on a T2- or T2*-weighted image. Thus, with increasing
oxyhemoglobin, there is increased signal intensity. By comparing
images within regions of activated brain versus nonactivated brain,
focal areas of increased signal can be detected, which correspond
to increased blood flow. The contrast resulting from the
deoxygenation of oxyhemoglobin to deoxyhemoglobin can be utilized
by means of BOLD sequences.
Several MR angiographic (MRA) techniques are available for
imaging of the intra- and extra-cranial vasculature. The inflow
(time-of-flight; TOF MRA) method is widely used and can be combined
with magnetization transfer pulses and, optionally, with Gd
injection to give high resolution images of the cerebral
Phase-contrast MRA, on the other hand, permits the acquisition of
directional information and flow quantification.
Both methods have benefited from developments in fast gradient
hardware, which permits short echo times (TE) and thus fewer inflow
artifacts and turbulence-related signal losses.
A rapidly developing approach to MR angiography of the intra-
and extra-cranial vasculature is contrast-enhanced MRA (CE-MRA).
Generally, CE-MRA is performed using gradient-echo sequences with
short TR/TE (<6/<2 ms) and a large flip angle (>30°). The
overall effect is one of increased vascular signal due to tissue
signal saturation and a Gd-induced reduction in blood T1.
Gadolinium-based contrast agents are generally injected as a bolus,
preferably with a power injector. Depending on the injected dose
(usually between 0.1 and 0.3 mmol/kg bodyweight), the blood T1 can
be decreased by between 50 and 150 ms. These very short T1 values
give rise to markedly increased signal intensities and inherently
high SNRs on heavily T1-weighted sequences. While the use of a
short TR ensures a short acquisition time, the use of a short TE
ensures that CE-MRA images are less prone to turbulence-related
signal loss and intra-voxel dephasing effects than TOF or
phase-contrast sequences. Another important feature of CE-MRA is
that it is almost independent of the direction of flow and,
therefore, in-plane flow can be imaged over a very large
field-of-view (for example, in carotid studies conducted in the
coronal plane). In 3D CE-MRA, a large imaging matrix over a large
field-of-view is usually employed, which permits the acquisition of
images with high spatial resolution.
Alternatively, higher temporal resolution can be achieved by
means of two-dimensional (2D) or 3D MR digital subtraction
angiography (MRDSA). This technique consists of dynamic
acquisitions with high temporal resolution (2D projection images at
a rate of 2 frames/sec or a 3D volume every 3 to 6 sec) and
involves real-time subtraction of a mask image, itself acquired
before the Gd bolus has reached the arteries of interest.
Thus, for the first time, 2D MRDSA has the potential to provide
similar dynamic information to that which is routinely obtained
with conventional digital subtraction angiography (DSA). The
hemodynamic information gained on MRDSA has already proven to be
clinically relevant for the evaluation of arteriovenous
malformations and fistulas.
In general, an intravenous injection of 7 mL Gd at a flow rate
of 3.5 to 5 mL/sec provides satisfactory images for brain AVM
studies. Because the technique involves a subtraction, the
injection can be repeated, and different imaging orientations can
be obtained. Moreover, by acquiring alternate images in two
orthogonal planes, the need for re-injection can be eliminated.
Although the SNR drops with increasing time resolution, the use of
high R1 contrast agents, such as gadobenate dimeglumine, may
represent a way to obviate the problem.
Unlike 2D MRDSA, 3D CE-MR arteriography is individually timed
after the injection of 20 to 40 mL Gd at a rate of 2 to 4 mL/sec,
with an acquisition duration of 15 to 35 sec.
In order to trigger the 3D acquisition in the arterial phase, 2D
MRDSA can be implemented as Gd tracking method, enabling the MR
operator to follow the progress of the contrast agent in the
vasculature up to a reference artery: at this point, the 3D
sequence is started. In this arrangement the 3D k-space has to be
sampled from the center to the periphery, following the so-called
elliptical centric scheme.
Three-dimensional CE-MR arteriography is an appropriate technique
for numerous purposes. Used in conjunction with a head-neck coil,
it is an effective technique to study the carotid and vertebral
arteries comprehensively in the coronal plane both from their
origin up to the intracranial arteries. It has also been shown to
be valuable for the detection of intracranial aneurysms and for the
evaluation of AVMs.
Parallel imaging techniques are especially useful in CE-MRA.
Sensitivity encoding shortens the acquisition time, which in turns
allows multiple 3D arterial phases to be acquired. The reduction of
raw data collection with SENSE can be reinvested in the form of an
increased matrix size or an increased number of slices. Moreover,
knowledge of coil sensitivity profiles can be exploited in order to
correct signal intensity inhomogeneities resulting from the use of
coupled surface coils. A reduced SNR is inherent to the SENSE
approach. However, this does not represent a critical problem in MR
arteriography, since the arterial SNR is high due to the high
arterial Gd concentration.
At high field (3.0 T), images with higher spatial resolution are
attainable, with voxel sizes of about 0.6 mm
While this is adequate to demonstrate arterial occlusion due to an
embolus or intracranial atherosclerosis, it is not usually
sufficient to evaluate vasculitis or arterial spasm from
subarachnoid hemorrhage. For this, 250 µm spatial resolution is
required (1024 ¥ 1024), which is the same as that available in
catheter angiography. In order to achieve this high resolution, it
is likely that imaging at 3.0T should be combined with the use of
improved Gd chelates.
CLINICAL APPLICATIONS OF CONTRAST-ENHANCED NEURO
Pathologies affecting the CNS are numerous and diverse, and
there are marked differences between diseases in the need for
contrast enhancement to achieve accurate diagnosis. The following
section will review the current value of contrast enhancement for
the diagnostic evaluation of CNS pathology in the brain and spine.
The relative value of recent contrast-enhanced techniques for the
imaging of these pathologies is also discussed.
The vast majority of congenital abnormalities do not require
contrast enhancement to be visualized satisfactorily. Exceptions,
however, include some vascular malformations and tumors associated
with congenital abnormalities as seen, for example, in
The intracranial types of vascular malformation are AVMs,
capillary telangiectases, cavernous angiomas, and venous
malformations. In many of such cases, conventional MRI is a superb
technique because it provides exquisite detail of the parenchyma
while simultaneously demonstrating the vascular anomalies.
Cerebral AVMs are classically described as congenital
malformations consisting of arteriovenous shunts located in the
subpial space, with arteries that drain directly into veins in the
absence of a normal capillary system. Patients with AVMs can
present with various symptoms: cerebral and subarachnoid
hemorrhage, seizures, neurologic deficits, or headache. The
treatment of AVMs potentially involves three modalities:
microsurgery, endovascular embolisation, and radiosurgery. The most
appropriate choice of modality depends on several factors, such as
angioarchitectural features, the size of the nidus, and the
location in the brain. If comparatively small in size and located
near the convexity, surgical removal is frequently the approach of
choice. Larger AVMs, on the other hand, are best treated by
endovascular therapy, often during multiple sessions. When the AVM
volume is relatively small, high-dose radiosurgery induces complete
obliteration. Endovascular embolization causes huge hemodynamic
changes and provides partial obliteration, or, less frequently,
MRI and MRA provide valuable information concerning the anatomic
location, the morphology of the AVM, and the surrounding brain
parenchyma. Additionally, MR can assess the presence of risks
factors, including aneurysm, venous thrombosis, central venous
drainage, periventricular or deep location and the occurrence of a
previous hemorrhage. An improved delineation of feeding arteries
and draining veins is obtained when 3D phase-contrast acquisitions
are acquired after administration of a Gd contrast agent. However,
3D CE-MRA techniques have the advantage of a shorter acquisition
time and less sensitivity to turbulence-related signal losses. The
dynamic information required by clinicians is traditionally
obtained by means of conventional X-ray angiography DSA, but recent
developments in MRDSA acquired after a short Gd bolus have shown
that some hemodynamic features can be demonstrated.
In this context, the advent of high R1 contrast agents, ie, those
with a greater T1 shortening per unit dose, is extremely promising
Dural AVMs and dural fistulae are often more difficult to detect
than parenchymal AVMs as they are located close to the bony
structures of the skull or skull base, and typically also involve
the cavernous sinus. A cortical dilated vein or a sinus stenosis
may be the only indication of this vascular anomaly. Traditionally,
conventional angiography, rather than plain contrast-enhanced MRI,
has been used most frequently for the evaluation of these lesions.
However, advances in CE-MRA have not only meant that depiction of
the dural malformation is now achievable non-
invasively, but also that the interventional neuroradiologist
performing the embolization can be guided more easily.
Capillary telangiectases are vascular malformations that were,
in the past, most frequently identified at postmortem examinations.
Today, they are often detected incidentally on contrast-enhanced
studies as poorly delineated foci of increased signal intensity,
most frequently in the mid pons.
Cavernous angiomas are typically not detected on DSA but have a
very characteristic appearance on unenhanced T1- and T2-weighted MR
images and a variable appearance on enhanced images. Venous
angiomas associated with cavernous angiomas are extremely important
to detect, as they may be responsible for hemorrhage. These lesions
are best seen on contrast-enhanced MRI.
Venous malformations are equally important to detect, and their
presence may have important clinical implications. The use of
late-phase CE-MRA images can be recommended as an easy, rapid, and
efficient way to display these anomalies.
Changes in the signal behavior of tumors after contrast
enhancement are often of significant diagnostic importance. The
signal behavior of tumors on contrast-enhanced T1-weighted MRI
frequently permits better tumor characterization and tumor
delineation (Figures 2 through 7). Increased tumor vascularity,
however, is not synonymous with malignancy. Several recent
developments have improved the approach to the diagnosis of brain
tumors. As discussed above, echo-planar perfusion imaging can be
used to show the region of highest rCBV, which usually corresponds
to the most malignant portion of the tumor. This information guides
the choice of the biopsy site, as tumors are often heterogeneous
and the most malignant areas should be identified for proper
grading in order to enable the optimal therapeutic choice. Relative
cerebral blood volume measurements have been shown to correlate
with the vascularity of various intracranial mass lesions. MR CBV
maps can be used to assess the degree of neovascularization in the
brain. In addition to aiding tumor grading, the evaluation of
malignancy, and the identification of other types of lesions
without neovascularization such as radiation necrosis and cerebral
abscess, perfusion MRI may be useful for the follow-up of treatment
by providing a noninvasive assessment of changes in tumor rCBV.
Furthermore, MR measurements of rCBV have been shown to
correlate with both conventional angiographic assessments of tumor
vascular density and histologic measurements of tumor
neovascularization. The results of perfusion-sensitive MR imaging
using a gradient echoplanar technique have been shown to correlate
with both histologic and angiographic vascularities.
In fibrillary low-grade astrocytomas, rCBV measurements appear
useful for the noninvasive assessment of angiogenesis. Data suggest
that high pre-therapeutic angiogenic activity in low-grade
astrocytomas indicates a subgroup of tumors at higher risk for
early local recurrence or malignant transformation after
fractionated stereotactic radiotherapy.
Perfusion MR data has been shown to correlate with clinical
response in patients undergoing antiangiogenic therapy.
Recent work aimed at correlating tumor grade with abnormalities in
the recirculation of a contrast agent bolus may provide independent
information on microcirculation and angiogenesis. Abnormalities of
contrast agent recirculation may be of value as surrogate markers
in trials of antiangiogenic therapy.
MR spectroscopy (MRS) can also be used to show the most
malignant portion of the tumor, based on the ratio of choline to
N-acetyl aspartate (NAA). However, MRS suffers from lower spatial
resolution than perfusion imaging. Although MRS voxel sizes have
improved from 2 cm (single voxel) to 1 cm (single slice,
multivoxel) and 7 mm (multislice-multivoxel or 3D), perfusion
imaging has an in-plane resolution of 2 mm.
Contrast-enhanced FLAIR is more useful than unenhanced FLAIR and
enhanced T1-weighted imaging for subtle cortical processes, such as
leptomeningeal carcinomatosis and early meningoencephalitis. This
is probably due to the fact that less Gd is required to leak into
the adjacent sulcus to change the signal intensity of the
cerebrospinal fluid (CSF) and prevent nulling on FLAIR that is
required to cause increased signal on T1-weighted imaging.
Unenhanced and contrast-enhanced T1-weighted imaging is still
required to distinguish enhancement from T2 prolongation as a cause
of hyperintensity on enhanced FLAIR.
Contrast-enhanced MRI is mandatory for better evaluation of most
infectious types of lesions such as meningitis, ventriculitis,
abscess formation, and encephalitis. Unfortunately, the contrast
enhancement patterns are often nonspecific. Other MR sequences,
such as MR diffusion and MR spectroscopy, allow the differentiation
of tumor from abscess in cases of ring-like enhancing lesions
(Figure 8). In patients with HIV, micro-abscesses may occur that
can be revealed only with Gd-enhanced T1-weighted sequences.
However, in such cases, diffusion sequences and MRS may again serve
as valuable additional tools to the contrast-enhanced sequences.
Subdural empyema is a severe condition often overlooked on CT that
can be readily appreciated on Gd-enhanced T1-weighted MRI.
Although multiple sclerosis (MS) was the earliest application
demonstrated for MRI in the CNS more than 20 years ago, new MR
techniques continue to improve the diagnosis of this and other
demyelinating diseases, such as acute disseminated
encephalomyelitis (ADEM). Using 2-mm sagittal FLAIR images,
subcallosal striations can be seen as a stack of 1-mm thick coins
aligned perpendicular to the ependymal stripe. Detection of such
lesions using thin-slice, sagittal FLAIR has been shown to be
highly correlated (
<0.001) with MS.
Acute inflammatory lesions show early disruption of the BBB,
which is best demonstrated on contrast-enhanced T1-weighted images
(Figure 9). Gadolinium enhancement correlates in time with the
acute inflammatory phase; without treatment this enhancement
subsides after approximately 1 month.
Acute demyelination can also be seen with diffusion imaging,
which may be useful in making the diagnosis. Moreover, early
studies also suggested that proton spectroscopy could lead to
better categorization of MS lesions than contrast enhancement.
However, with technological developments and the advent of novel
contrast agents, this may no longer be the case.
Degenerative Diseases and Dementia
Currently, the use of contrast enhancement for imaging of
degenerative diseases and dementia is limited. Perfusion studies
can be helpful for the evaluation of Alzheimer's disease in which a
temporoparietal hypoperfusion can be found, or in frontotemporal
dementias, in which, again, areas of hypoperfusion are observed
even in early stages of the disease.
Although MR perfusion studies provide similar information as PET,
absolute quantification of the measurements is difficult with
However, in current clinical practice, PET studies are usually
preferred to perfusion MRI studies.
MRI is not only a valuable complementary imaging procedure to CT
for the evaluation of acute head trauma but is frequently
considered the best imaging procedure for the evaluation of most
traumatic lesions. Patients with corpus callosal injuries, in whom
the incidence of diffuse axonal injury is comparatively high,
may have normal CT findings but demonstrate evidence of the injury
on MRI. These lesions might be responsible for some aspects of the
postconcussive syndrome. Although contrast injection is rarely
required for the evaluation of posttraumatic changes, it can be of
benefit when infectious complications occur, such as in cases of
subdural empyema or intracranial abscesses linked to open fractures
of the skull and skull base involving the sinuses.
In cases of traumatic dissection of the carotid or vertebral
arteries, contrast-enhanced T1-weighted bolus MR angiography is the
preferred procedure since the technique is fast and reliable.
Moreover, the technique can be performed in conjunction with a
postcontrast T1-weighted brain examination to demonstrate the
extent of the BBB disruption. At present, in routine practice,
diffusion imaging is the preferred approach to evaluating secondary
brain ischemic lesions linked to traumatic arterial
Acute Ischemia and Stroke
Fifteen percent of all strokes in the Western world are due to
hemorrhage rather than ischemic events. CT is currently performed
in the acute setting to exclude hemorrhage in patients with
evolving symptoms who are being considered for heparinization.
However, the ability of MRI to detect hyperacute hemorrhage either
in the brain parenchyma or in the subarachnoid space may enable MRI
to become the primary imaging tool for the evaluation of acute
stroke. Since routine treatment of stroke with thrombolytic agents
generally does not allow either the time or the expense of
performing both a CT and an MR examination, it is likely MRI will
prove to be the most effective technique to detect hemorrhage, to
further characterize the lesions, and to determine whether
thrombolysis or neuroprotection is indicated.
The evaluation of stroke patients is commonly performed by means
of echoplanar perfusion and diffusion imaging. Whenever the area of
the perfusion abnormality is greater than the diffusion
abnormality, additional brain is at risk for extension of
infarction. This area of potentially salvageable brain is known as
the ischemic penumbra
and is the target of current thrombolytic agents and future
so-called neuroprotective agents.
While the core of the infarct is determined easily by diffusion
imaging, the ischemic penumbra can occupy a volume three times
greater than the core. Therefore, patients with an ischemic
penumbra should be identified immediately and treated within 3
hours (for intravenous thrombolysis) or 6 hours (for intra-arterial
thrombolysis) under current protocols. With echoplanar diffusion
and perfusion techniques, frequently decisions to perform
thrombolysis can be individualized. For example, a recent study
revealed that when the rCBF in the infarcted hemisphere was <35%
of the normal hemisphere, bleeding occurred with intra-arterial
thrombolysis even when it was performed within 2 hours of symptom
onset. On the other hand, if the ratio of rCBFs exceeded 55%, no
bleeding occurred, even if intra-arterial thrombolysis was
administered as late as 12 hours after symptom onset.
Ideally, all potential candidates for thrombolytic therapy would
undergo echoplanar diffusion and perfusion imaging, since as many
as 50% of current stroke patients receive thrombolytics
inappropriately, sometimes resulting in unnecessary cerebral
hemorrhage. For example, an estimated 20% of patients who present
with stroke in fact have a process other than ischemia.
These include hemiplegic and hemisensory migraines, Todd's
paralysis following a seizure, and certain inflammatory conditions.
Thus, positive diffusion imaging will confirm the presence of true
ischemia, eliminating patients who do not have ischemia and who
should not be given thrombolytics. In addition, approximately 10%
of all patients have had silent infarcts prior to the clinical
While these older areas of infarction may not be evident on CT,
they are clearly seen on echoplanar diffusion imaging and would
constitute an increased risk of hemorrhage if thrombolytics were
administered, since they exceed the 3- to 6-hour time window.
Approximately 20% of all patients presenting with stroke have
matched diffusion and perfusion defects, ie, they have no ischemic
Such patients are not candidates for thrombolytic therapy since the
core of the infarct cannot be salvaged with current therapies.
Not all strokes are detected with echoplanar diffusion imaging.
Indeed, positive identification is not usually made until the rCBF
is <20% of normal.
Most MR stroke evaluation today is accompanied by MR angiography of
the carotid arteries and the intracranial circulation.
SPINE AND SPINAL CORD
In the spine, MRI has rapidly replaced myelography and myelo-CT
for the evaluation of myelopathy and is today considered the
principal modality for the evaluation of spinal cord lesions.
Contrast-enhanced MRI is also of major use for the evaluation of
bone lesions and degenerative disease in the spine. Currently, CT
occupies a secondary, adjunctive role in the evaluation of the
spine and spinal cord, generally for the evaluation of trauma and
for the detection of small bone fragments impinging upon the
Spinal Cord Tumors
Tumors of the spinal cord are rare. Overall, only approximately
5% of spinal tumors are intramedullary, while 40% are intradural
extramedullary and 55% are extradural.
Contrast-enhanced MRI is mandatory for the work-up and
preoperative evaluation of spinal cord tumors (Figure 10). However,
it is important to bear in mind that not all tumors enhance, and
that not all enhancing intramedullary masses expanding the cord are
Most intra-medullary tumors are one of the following types:
astrocytomas, ependymomas, or hemangioblastomas. However, a
differential diagnosis is important as surgical planning and
prognosis will be very different for the different lesion types
(Figure 11). Hence, it is important that contrast enhancement
patterns are described carefully: ependymomas usually enhance more
intensely and homogeneously, while astrocytomas enhance moderately
and are usually inhomogeneous in appearance. Hemangioblastomas are
benign, richly vascular spinal cord tumors, often located
superficially, in the subpial region. Tumor nodules are usually
small and are typically associated with extensive
hydrosyringomyelia. After Gd injection, intense and homogeneous
contrast uptake is seen. Contrast administration is especially
useful in the case of small, often multiple nodules such as seen in
von Hipple Lindau disease.
Many intramedullary tumors exhibit cystic components that are
either part of the tumor or are associated satellite cysts that are
located cephalad or caudal to the solid nodule with walls that do
not enhance after Gd injection. These cysts develop as an
intramedullary reaction to the presence of the tumor and the liquid
has a similar behavior to CSF. They collapse spontaneously after
surgery. On the other hand, tumoral cysts have enhancing walls and
should be removed at surgery.
Even with MRI, it may be difficult or impossible to
differentiate spinal tumors from intramedullary non-neoplastic
In a study of 212 patients undergoing surgery for intramedullary
spinal cord tumors, Lee et al
reported finding non-neoplastic lesions in 4% of patients;
histology showed demyelinating disease, sarcoidosis, amyloid
angiopathy, and a mass of non-neoplastic inflammatory cells of
Intradural Extramedullary Tumors:
These lesions are predominantly benign tumors and include
nervesheath tumors and meningiomas. Contrast injection should be
administered systematically in order to better delineate the
borders of the tumor and to help in tumor characterization (Figure
Most spinal epidural tumors are malignant tumors, primarily
metastases. However, lymphoma, leukemia, and multiple myeloma must
also be included in the differential diagnosis. Contrast injection
is useful to better delineate the intracanalar extension but should
not be given from the outset as bony infiltration can be masked
after contrast injection. Demonstration of the possible association
of neoplastic bony infiltration is best achieved on unenhanced
MRI-guided biopsy is an area of growing clinical interest due in
part to the superior tissue contrast of MRI relative to CT or
ultrasound. The development of MR-compatible biopsy devices, as
well as the diffusion of techniques involving image co-registration
and the use of stereotactic frames or pellets, has made possible
the era of guided microsurgery. Image-guided techniques have proven
useful for directing biopies and resections, and for determining
radiosurgery targets when using gamma knives or accelerator beams.
Newer image co-registration techniques (MR-MR or MR-PET) permit
improved follow-up studies of residual AVM niduses or small
cerebral-pontine tumors after resection or irradiation.
A number of centers have now placed MR scanners in neurosurgical
operating rooms. While this may appear somewhat extravagant, it is
important to remember that current high-end operating microscopes
cost close to $1 million which is similar to the cost of an open
magnet with a coupled optical tracking system. Such systems are
clearly useful for biopsy; however, they have shown even greater
utility in guiding the resection of brain tumors. Several years
ago, the neurosurgeons at Brigham and Women's Hospital (Boston, MA)
and the neuroradiologists at Long Beach Memorial Medical Center
(Long Beach, CA) presented data at U.S. national meetings showing
that in 80% of cases in which the neurosurgeons thought they had
achieved a gross total resection, MR-visible tumor could still be
While this is unlikely to have a major impact on outcome for
high-grade gliomas, true gross total resection of low-grade gliomas
effectively achieves a cure. Leaving even a small remnant of
low-grade glioma behind increases the chance of subsequent
degeneration into a glioblastoma multiforme, which is almost always
fatal. Thus, whenever possible, low-grade gliomas should be
resected under MR guidance.
Recently, at the Erasme Hospital (Brussels, Belgium), a
low-field open intraoperative magnet was installed to guide and
monitor the removal of both pituitary and brain tumors. Although
the image field-of-view is limited and image resolution is
relatively low, image fusion between the low-field images obtained
in the operating theater and the preoperative high-resolution
images has markedly improved MR guidance (Figure 13). In this
setting, contrast injection is required to enhance lesion border
definition: the use of gadobenate dimeglumine (Gd-BOPTA;
MultiHance, Bracco Diagnostics Inc., Princeton, NJ) is presently
undergoing evaluation, since better contrast enhancement can be
expected due to the higher relaxivity of this contrast agent.
In addition to the presence of MR scanners in neurosurgical
operating rooms, MR imaging systems are now combined with
angiography suites. This could be useful for the evaluation of
perfusion and diffusion in stroke patients while intra-arterial
thrombolytic therapy is administered.
In the future, transcranial focused ultrasound may be capable of
achieving brain tumor ablations without breaking the skin. One
advantage of performing ablations under MR guidance is the
sensitivity of MR to temperature changes (MR thermometry).
With large arrays of focused ultrasound transducers,
power can be delivered to a point in the brain with minimal heating
of the surrounding normal tissue. Such techniques would be
particularly useful for deep-seated tumors where exposure is a
problem and gaining access to the tumor frequently causes as much
damage as the tumor itself. Focused ultrasound may also be used in
the future to deliver gene therapy to focal areas of the brain, eg,
tumors. Liposomes loaded with genetic material can be selectively
lysed using focused ultrasound at a specific target. Since neither
surgery, radiotherapy, nor chemotherapy has yet demonstrated
significant impact on survival in patients with glioblastoma
multiforme, it is hoped that gene therapy/molecular medicine will
achieve this goal in the future.
TECHNICAL DEVELOPMENTS IN MRI
As MRI gradually became a dominant imaging modality in medicine,
efforts by physicists further accelerated its development. As a
result of multiple refinements in software and hardware, the future
of CNS MRI will be both better and faster, with applications not
even dreamed of a few years ago. Major hardware improvements will
be found in every major MR subsystem: main magnet, gradients, RF,
and computer. Moreover, improvements in imaging technique have
already greatly benefited clinical practice and will continue to do
so. Finally, the development of novel contrast agents with unique
properties promises also to further extend the horizons for MRI of
The field strength of the main magnet, which had plateaued at
1.5 T for almost two decades, has now jumped to 3.0 T. Systems with
even higher field strengths are now being used for research
Increasingly, the new 3.0 T MR systems are being used in
academic settings and busy outpatient clinical environments.
Specific applications that are facilitated at high field include
fMRI and MR spectroscopy (MRS). Functional MRI has improved due to
an increased BOLD effect at higher field caused by increased
susceptibility effects. MR spectroscopy has improved as a result of
increasing peak separation at higher fields. Arterial spin-labeling
techniques should also perform better at 3.0 T. MR angiography at
3.0 T may also permit marked improvements in spatial resolution.
There are currently three classes of gradients, which are
characterized by strength and slew rate (strength/rise time). These
are labeled standard (10 mT/m strength, 10 mT/m/msec slew rate),
EPI (20 to 30 mT/m strength, 70 to 120 slew), and cardiovascular
(40+ mT/m strength, 150 to 200 slew).
Standard gradients were available when the first high-field
systems were introduced in 1984. EPI gradients became available in
1994 and cardiovascular gradients became available in 1999.
Refinements in the gradient subsystem have resulted in the
ability to acquire echoplanar images in 100 msec, ie, the same time
as electron beam CT. Although the speed of EPI was the initial
promise of these strong, fast gradients, the real application has
come with the introduction of echoplanar diffusion and perfusion
imaging, particularly for evaluation of ischemic stroke. The
ability of EPI gradients also facilitated the introduction of
contrast-enhanced MR angiography (CE-MRA), which is now the
preferred imaging technique not only for the carotid arteries, but
also for every other artery in the body.
With the newest class of cardiovascular gradients, these early
EPI applications are being extended, eg, to diffusion tensor
imaging and even higher resolution CE-MRA. The great sensitivity of
T2*-weighted gradient echo EPI techniques to susceptibility effects
has allowed fMRI in the brain. These gradients also allow
single-shot fast-spin-echo (HASTE) imaging in several hundred msec
as well as high-resolution gradient-echo techniques (eg, True FISP,
FIESTA, and CISS) with acquisition times of under one second.
Advances in the radiofrequencycoil subsystem have the potential
to impact the speed of imaging, much as fast-spin echo did a decade
ago. Using phased arrays, parallel imaging (eg, SENSE, SMASH)
becomes possible, allowing time-savings on the order of a factor of
four or higher. Sensitivity encoding takes advantage of the local
sensitivity of the individual coils in the phased array to decrease
image acquisition times or to improve the image quality. This has
been demonstrated to have major potential impact in MRA and fMRI.
Since this is accomplished by decreasing the number of actual
acquisitions, halving the acquisition time results in the usual
drop in signal-to-noise (S/N) of 40%. Thus, parallel imaging
techniques will have their greatest utility in applications that
are not limited by signal-to-noise issues, eg, CE-MRA and 3.0 T
Improvements in the computer subsystem have resulted in the
ability to handle increasingly large data files in increasingly
shorter times. Bandwidth continues to increase for the analog to
digital converter and the entire imaging data chain. Backend (ie,
workstation) applications continue to improve (eg, multiplanar
reformations, volume rendering, and maximum
The Gd-based contrast agents in current widespread use for MRI
of the CNS can be considered first generation agents and
* Magnevist (gadopentetate dimeglumine [Gd-DTPA]; Schering AG,
Berlin, Germany/Berlex Imaging, Wayne NJ);
* ProHance (gadoteridol [Gd-HP-DO3A]; Bracco Diagnostics Inc.,
* Omniscan (gadodiamide [Gd-DTPA-BMA]; Amersham Health,
* Optimark (gadoversetamide [Gd-DTPA-BMEA]; Mallinckrodt, St.
* Dotarem (gadoterate meglumine [Gd-DOTA]; Guerbet,
Aulnay-Sous-Bois, France); and
* Gadovist (gadobutrol [Gd-DO3A-butrol]; Schering AG).
These agents have a purely extracellular distribution and cross
the BBB only when breakdown occurs. Despite different chemical
structures and properties, these agents have similar R1 relaxivity
values of approximately 5.0 mM
and behave in a similar manner when used for contrast-enhanced MRI
of the CNS.
Gadobenate dimeglumine (Gd-BOPTA, MultiHance) is a
second-generation Gd agent. This agent resembles the above agents
in terms of its pharmacokinetic profile and physicochemical
properties but differs in that it possesses a two-fold greater R1
relaxivity (9.7 mM
) deriving from a capacity for weak and transient interaction with
Intra-individual crossover studies with first-generation chelates
have shown that this greater relaxivity significantly improves both
and delineation of intra-axial enhancing lesions.
Parrallel group studies comparing
Gd-BOPTA with conventional Gd agents have also demonstrated
potential benefits of Gd-BOPTA in contrast enhancement.
Future generations of MR contrast agents may include agents
targeted to specific pathologies through the covalent labeling of
monoclonal antibodies with Gd or other paramagnetic agents. Other
innovative possibilities include agents that, upon attaching to
receptors, open to expose Gd to water and cause T1 shortening. Such
targeted agents would be expected to cause enhancement only when
joined with the target.
MRI continues to evolve with increasing field strength, stronger
and faster gradients, and more sophisticated RF subsystems and
computers. This allows faster, higher resolution imaging and
evaluation of processes at the molecular level, eg, diffusion and
perfusion. Contrast agents continue to evolve from the first
generation of extracellular agents to the latest Gd agents that are
able to interact with proteins to achieve greater T1 relaxivity.
Future contrast agents are likely to be targeted even more
specifically to pathology. MR continues to be used to guide
interventions from biopsies to brain tumor resections. In the
future, MR will likely be used to guide stroke therapy, gene
therapy, and minimally invasive brain tumor ablations using
techniques, such as focused ultrasound, that do not even require
the skin to be broken.
This activity has been planned and implemented in accordance
with the Essential Areas and Policies of the Accreditation Council
for Continuing Medical Education (ACCME) through the joint
sponsorship of the Institute for Advanced Medical Education and
Anderson Publishing, Ltd. The Institute for Advanced Medical
Education is accredited by the Accreditation Council for Continuing
Medical Education (ACCME) to sponsor continuing medical education
The Institute for Advanced Medical Education designates this
continuing medical education activity for a maximum of 2 Category 1
credits toward the AMA Physician's Recognition Award. Each
physician should claim only those credits that he/she actually
spent in the activity.
See page 42 for instructions on how to participate in the
As of January 1, 2003, courses approved for AMA Category 1 CME
credit that are relevant to the radiologic sciences are accepted
for Category B CE credit on a one-to-one basis by the American
Registry of Radiologic Technologists (ARRT).
In compliance with the Essentials and Standards of the ACCME,
the authors of this CME tutorial are required to disclose any
significant financial or other relationships they may have with the
manufacturer(s) of any commercial product(s) or provider(s) of any
commercial service(s) discussed in this program.
Dr. Baleriaux and Dr. Metens report no such relationships exist.
Dr. Bradley discloses relationships with Bracco Diagnostics and
Berlex Imaging through their speakers' bureaus and with GE Medical
Systems and Siemens Medical Systems through research support.
To go directly to the quiz please