Perfusion imaging capabilities are now widely available on high-performance magnetic resonance and helical computed tomography systems. Perfusion data, which can be complementary or sometimes critical, is coming into widespread use in the evaluation and surveillance of patients with neoplasm, brain ischemia, neuropsychiatric disorders, and epilepsy. In this article, the authors review perfusion techniques in the settings of stroke, neuropsychiatry, neuro-oncology, and others.
Dr. Tanenbaum is Section Chief of Neuroradiology, MRI, and
CT; and Dr. Hariharan is Director of Neuro-oncology and
Neurology Residency Program Director at the NJ Neuroscience
Institute, Edison, NJ.
Perfusion imaging techniques are now routinely available on
high-performance magnetic resonance (MR) and helical computed
tomography (CT) systems. MR and CT perfusion first-pass studies
provide information analogous to that provided by Technicium
D1-hemamethypropylene amine oxime (Tc-HMPAO) single-photon computed
tomography (SPECT), and xenon CT, and, in many circumstances,
mirror F-18 fluorodeoxyglucose (FDG) positron-emission tomography
(PET) as part of a comprehensive brain examination. Perfusion
information can be complementary or critical in the evaluation of
patients with brain disease. These techniques are coming into
widespread use in the evaluation and surveillance of patients with
neoplasm, brain ischemia, neuropsychiatric disorders, and
High-performance, gradient-en-hanced MR scanners employ
extremely fast single-shot echo planar imaging (EPI) techniques to
image multiple physiologic phases at multiple (z-axis) brain
locations during the first pass of a gadolinium chelate,
paramagnetic contrast agent (figures 1 and 2). CT scanners can
repetitively scan one or multiple (on multichannel systems) slice
locations during the first pass of a standard nonionic iodinated
contrast agent to obtain maps of brain perfusion (figure 3).
Perfusion CT provides information equivalent to xenon CT.
Since xenon CT technique is more costly, personnel intensive, and
more dependent on patient cooperation, it has achieved limited
acceptance in the community setting.
Commercially available software is used to process first-pass
data into maps of brain perfusion. Algorithms for the processing of
regional cerebral blood volume (rCBV) maps are widely available for
MR and CT. Calculation of mean transit time (rMTT) requires the
assumption of an instantaneous contrast agent arrival in the target
organ, physically impossible with first-pass techniques. Very high
injection rates (10+ cc/sec) have been used to simulate
instantaneous tracer arrival, but concerns about patient tolerance
has limited utilization of this approach in the clinical community.
Fortunately, complex deconvolution algorithms that compensate for
the actual, finite arrival time of the injected first-pass contrast
agent are now commercially available for CT data. Regional cerebral
blood flow (rCBF) is calculated from the ratio of rCBV and rMTT.
Accurate rCBF data that correlate well with xenon CT studies can
therefore be obtained readily with first-pass CT.
Quantitation of MR data is complicated by a number of factors,
however. Assumptions about the relationship of signal intensity
(MR)/density change (CT) with tracer concentration, well documented
with iodine based contrast agents and
x-ray/CT, may not hold for the susceptibility based signal changes
associated with gadolinium-based MR agents. This complicates
establishment of an accurate input function for the MR first-pass
data. A commercially available quantitative perfusion software
package is not yet available.
Perfusion techniques offer the most sensitive measure of the
extent of brain tissue under ischemic conditions in patients with
symptoms suggesting acute and subacute stroke. The deficit on a
perfusion study is often greater than that seen on
diffusion-weighted imaging (DWI) (and CT) in the acute setting. A
reduction of cerebral blood flow (CBF) is a typical accompaniment
of acute stroke. Depending on severity, this will manifest as a
compensatory increase or a resultant decrease in cerebral blood
volume (CBV) (figure 4), as well as a regional prolongation of MTT.
The combination of MR angiography, DWI, and perfusion MR
(figure 5)or more commonly CT, CT angiography,
and (recently) perfusion CT--are used in the triage of patients in
whom thrombolytic intervention is contemplated (figures 6-8).
Quantitative assessment of CBF yields information about brain
tissue viability and hemorrhagic risk, which is critical for
thrombolytic therapy decision-making. Subtracting the volume of
brain with restricted diffusion from the perfusion-indicated volume
of tissue under ischemic conditions yields the commonly accepted MR
paradigm for tissue at risk for extension of infarction. The CT
analogue of DWI is not clear and is under active study.
Outside of the setting of acute stroke, perfusion imaging can
yield useful information about the functionality of collateral
circulation and, thus, the significance of vascular occlusive
disease (figures 9 and 10).
Perfusion techniques are in routine use for the evaluation of
neuropsychiatric disorders. Traumatic brain injuries typically
manifest as regions of de-creased CBV that correlate better than
structural studies with the results of neuropsychologic testing
(figure 11). Positive results increase treating physicians'
diagnostic confidence, often leading to more aggressive therapy.
Dementia of the Alzheimer's type typically presents with decreased
temporoparietal perfusion, mirroring the findings seen with SPECT
and PET (figure 12).
Characteristic patterns are evident with perfusion imaging of
Certain lesions, such as meningiomas, exhibit a striking increase
in relative CBV (figure 13), while schwannomas and lower grade
gliomas demonstrate less impressive alterations. Generally,
high-grade gliomas are very heterogeneous on perfusion studies with
areas of high and low CBV. Therefore, CBV studies are useful in the
characterization of focal brain lesions, particularly
The extent of a typically non-enhancing, low-grade primary brain
neoplasm is better assessed by perfusion studies that measure
capillary density, rather than disruption of the
blood-brain-barrier (BBB), which typically is intact. Perhaps the
most important role for perfusion MRI in the workup of tumors is in
lesion surveillance. Breakdown in the BBB, seen as contrast
enhancement on routine imaging sequences, is non-specific and could
represent postoperative change, necrosis, or tumor. Residual or
recurrent tumor will manifest an increase in CBV in contrast to
gliosis or radiation necrosis, which has a low CBV (figure 14).
Perfusion studies achieve a similar sensitivity to metabolic
assessment with FDG-PET, with greater specificity.
Taking advantage of the first-pass of the agent injected for the
routine contrast-enhanced study, perfusion imaging poses minimal
incremental cost other than the short time it takes to acquire the
images. Dedicated reimbursement may be available through the use of
image post-processing Current Procedural Terminology (CPT)
Perfusion imaging is a powerful, clinically practical technique
that provides critical information in the evaluation of patients
with ischemic, neuropsychiatric, and neoplastic brain disease.