Secondary Logo

Journal Logo

Advanced Search
Stroke

Diffusion- and Perfusion-Weighted Magnetic Resonance Imaging in Human Acute Ischemic Stroke: Technical Considerations

Ozsunar, Yelda M.D.; Sorensen, A. Gregory M.D.

Author Information

From the Department of Radiology, Massachusetts General Hospital and Harvard Medical School, MGH NMR Center, Boston, Massachusetts, U.S.A.

Address correspondence and reprint requests to Dr. Yelda Ozsunar, MGH-NMR Center, Mail code CNY 149-2301, Boston MA 02129, U.S.A., or [email protected]

Topics in Magnetic Resonance Imaging: October 2000 - Volume 11 - Issue 5 - p 259-272
  • Free

Abstract

In the past 10 years, diffusion-weighted imaging (DWI) and perfusion-weighted imaging (PWI) have become routine parts of diagnostic work-up of many pathological conditions, but perhaps none so widely as acute ischemic stroke imaging. Conventional magnetic resonance imaging (MRI) techniques, such as T1, T2, proton density, and fluid attenuated inversion recovery (FLAIR), are limited in their ability to depict the early extent of brain tissue injury. These sequences typically visualize the ischemic tissue 8–12 hours after ischemic onset, which many investigators fear may be too late for decision making about treatment with thrombolysis, neuroprotective drugs, or other interventions (1–5). DWI and perfusion magnetic resonance not only provide anatomic information, but they also evaluate cerebral physiopathology by observing water mobility and microvascular hemodynamics. This article reviews the physiological and physical aspects of DWI and PWI in tissues and summarizes the clinical applications of these techniques in acute ischemic stroke.

DIFFUSION-WEIGHTED IMAGING

Diffusion, also called Brownian motion, is a term used to describe the random movement of molecules due to thermal energy. Molecules diffuse in all three dimensions, and physicists describe the diffusability of a substance using the diffusion coefficient (D). The average path length (L) that diffusing water protons travel through the microenvironment of the brain is related to the duration of observation (t) and the diffusion coefficient (D). Einstein described this relationship as: L = 2 · t · D. For clinical purposes, the substance whose diffusion we are most interested in measuring is water, or water diffusion through water (sometimes termed water self-diffusion).

In tissues, measurement of diffusion becomes more complex because the water molecules run into other substances. In particular, water diffusion properties in tissue are affected by the structure of the cytosol, cellular elements, and barriers such as membranes. As a result, diffusion is said to be restricted. This restriction takes on two different aspects. First, because the majority of water is in tissue, almost all the water has less diffusion than if it were in a beaker or test tube. (A major exception to this is water in the cerebrospinal fluid in the ventricles.) Second, this restriction can take on directionality. If the restriction is less prominent, such as in cerebrospinal fluid inside the ventricles, molecules move in all directions equally; this is termed isotropic diffusion (Fig. 1A). When diffusion along one direction is faster than in another, as in myelin fibers, the diffusion is not isotropic; this is termed anisotropic diffusion (Fig. 1B).

F1-3
FIG. 1.:
Nature of the two types of diffusion. A: Isotropic diffusion. Molecular motion is unrestricted and occurs in all directions (free water, cerebrospinal fluid, etc.). B: Anisotropic diffusion. Molecule moves preferentially along one direction (fiber tracts, white matter, etc.).

This microscopic diffusion (D) can be measured by using DWI. The signals observed by DWI, however, might not be pure. The signals may contain either microscopic diffusion or other energy-dependent translational movement, such as bulk flow, gross motion, or even tissue perfusion. Hence, quantification of the diffusion observed by MRI is called “apparent diffusion coefficient (ADC)” rather than D in recognition of this. For normal brain tissue, however, we often assume that D and ADC are the same. Using the Einstein relationship, we can calculate that if the ADC value of normal brain tissue is measured from 0.72 to 1.2 × 10−3 mm2/s in gray matter and from 0.62 to 0.83 × 10−3 mm2/s in white matter (6–8), water molecules move roughly 8 μm within the 50 ms of diffusion encoding time.

The major application of diffusion imaging currently is for detection of the alterations in diffusion that occur with acute cerebral ischemia. This has been well documented (1–3,6,8) and is an extremely reliable finding. Despite more than 10 years of study, however, the exact mechanism for the reduction in water mobility is not completely understood. One popular hypothesis is that this restriction might be due to shifts in water location. Under physiological conditions, approximately 80% of the brain water is distributed in the intracellular compartment and 20% is in the extracellular compartment (9). Prolongation of cerebral ischemia results in failure of the Na-K pump, which eventually leads to increased intracellular Na+. The extracellular water then shifts into the intracellular space, and the water content of the extracellular space decreases up to 50%(9,10). Many have hypothesized that diffusion of intracellular water might be more restricted compared to water in the extracellular compartment. This shift might explain why the net ADC value of tissue decreases (11,12), but this is still quite controversial. Other mechanisms that have been proposed to explain the etiology of decreased diffusion in acute ischemia include increased water viscosity and extracellular tortuosity, changes in cytoplasmic circulation, decreased cellular membrane permeability, and water streaming (12–16).

Measuring diffusion with MRI

The physical principle of using nuclear magnetic resonance to measure diffusion was first described by Stejskal and Tanner (17) in 1965, well before imaging had been applied to nuclear magnetic resonance techniques. Two symmetric gradient pulses are added to typical spin-echo (SE) echo-planar sequences to acquire DWI (Fig. 2). The first rephasing diffusion gradient is applied between the 90° pulse and the 180° cause dephasing. The second diffusion gradient after the 180° pulse causes rephasing of stationary molecules. However, mobile molecules—even microscopically moving molecules—accumulate phase shift in their transverse magnetization and cannot rephase perfectly. These molecules cause signal loss depending how far they move. Hence, normal tissue loses signal on DWI relative to the initial T2-weighted imaging. However, in areas of decreased diffusion or low ADC values, such as in ischemic tissue, there is less signal loss. This results in a high signal relative to normal brain tissue on DWI (Fig. 3). The amount of signal attenuation (SA) in a given voxel is related to the degree of diffusivity in that voxel (D) and the diffusion sensitivity of the MRI pulse sequence. The diffusion sensitivity is quantitated as the “b value” of the image:EQUATION 1

F2-3
FIG. 2.:
Diffusion-weighted magnetic resonance sequence. Two symmetric diffusion sensitizing gradients are added into a typical spin-echo pulse sequence. Diffusion sensitivity depends on amplitude (G), duration (δ), and separation (Δ) of diffusion gradients.
F3-3
FIG. 3.:
Conventional versus diffusion-weighted MRI in acute infarct. A 41-year-old woman was examined 4 hours after onset of symptoms. T2-weighted and FLAIR images show no obvious abnormality. However, diffusion-weighted imaging (DWI) and ADC map clearly depict restricted diffusion in left middle cerebral artery territory. The acute infarct is seen as a bright lesion on isotropic DWI but as a dark area on ADC map.

Stejskal and Tanner calculated that this diffusion sensitivity or b value is affected by the duration (δ), amplitude (G), and interval (Δ) of the applied diffusion gradients proportionally. They provided the formula for calculating the b value of a given pulse sequence:EQUATION 2

Because the diffusion approach is based on signal loss, the signal-to-noise ratio often is limited in diffusion-weighted images. Hence, it is common to acquire multiple averages or do other postprocessing on the initial diffusion images. In addition, because the diffusion encoding gradients typically are applied in only one direction at a time, anisotropy artifacts could arise if only a single direction of diffusion encoding were to be used. As a result, what the scanner acquires as “raw diffusion weighted images” and what is eventually used for interpretation can be somewhat different, just as with magnetic resonance angiography. The individual directions of diffusion encoding often can have different signal intensities than the “average” or trace-weighted diffusion-weighted image (Fig. 4). Thus, postprocessing is an important part of DWI, although it usually occurs automatically on the scanner without user intervention. Because of this postprocessing, the raw data can be generated into different types of synthetic maps. A few of these maps have particular potential merit in the clinical setting.

F4-3
FIG. 4.:
Raw diffusion-weighted images of a healthy volunteer (A) and a patient with embolic infarct (arrow) in the left centrum semiovale (B). In full tensor mapping, diffusion gradients are applied in six directions. These directions are combinations of the x, y, and z matrix. White matter tracts where fibers are perpendicular to the applied diffusion gradient present with lower diffusion and therefore higher signal. In contrast, white matter tracts parallel to the gradient appear darker due to faster diffusion.
F4B-3
Figure 4:
Continued

Isotropic diffusion-weighted images

Isotropic diffusion-weighted images (trace DWI) are the most commonly displayed in current publications, on current scanners, and in most practices. As noted earlier, because molecular diffusion is three dimensional and can be anisotropic, such as in white matter, at least three orthogonal directions must be sampled. Otherwise, the strong directional preponderance of diffusion can causes artifacts, which can lead to misinterpretation. For example, white matter tracts that are perpendicular to the direction of diffusion gradient inherently generate high signal and can be confused with stroke. Diffusion gradients applied in orthogonal directions are averaged to obtain isotropic diffusion-weighted images, which are relatively insensitive to directional preponderance (Fig. 3) (7,18–20).

ADC maps

The ADC map, also called trace ADC, is generated to allow assessment of ADC itself without T2 effects. This is somewhat similar to the older approach of calculating pure T1 or T2 images rather than using T1-or T2-weighted images, but there is an important distinction. Because of the inverse exponential relationship between signal intensity on DWI and ADC, the images have reversed contrast in comparison to one another, even though they fundamentally contain very similar information. The ADC map is calculated according to Equation 1 and, in practice, is based on the logarithm of two images with different b values, such as Bhigh and Blow [ADC = -log Bhigh/Blow]. In this kind of image, restricted diffusion is demonstrated as low signal intensity. Unlike isotropic DWI or raw DWI, the ADC map represents only the diffusional properties of water but not T2 effect (Fig. 5) (21). Of course, an ADC map could be generated for each of the many directions that DWI data are acquired, but in practice an isotropic or trace-weighted ADC map is used.

F5-3
FIG. 5.:
T2 shine-through effect. MRI obtained 1 day after onset. Isotropic diffusion-weighted imaging (DWI) shows a hyperintense lesion consistent with acute cerebral infarct (white arrow). Lesion is mildly hyperintense on T2-weighted image. Infarcted tissue is seen on ADC map and exponential images as dark and bright signals, respectively. Signal abnormality seen on ADC map and exponential images are not as prominent as on DWI due to shine-through effect. Both ADC map and exponential images represent only diffusional properties but not T2 effect.

Exponential images

Exponential images have both proponents and detractors. Its clinical advantage appears to depend on having very high signal-to-noise ratio images, such as can be obtained by either performing multiple averages or using low-resolution imaging. To calculate these images, the diffusion-weighted image is divided by the signal intensity of T2 in order to take out the T2 effects. Exponential images give similar information as ADC maps, but on a reversed scale. Hence, acute infarct is seen as hyperintense signal without T2 effect (Figs. 5 and 6) (22).

F6-3
FIG. 6.:
Acute infarct appearances on different types of diffusion-weighted imaging (DWI) maps. A 78-year-old man was admitted with right-sided hemiplegia and change in mental status. Magnetic resonance scanning was performed 3 hours after ictus. Isotropic DWI shows a large acute infarct involving the territory of the left middle cerebral artery. Restricted diffusion, i.e., low ADC, is visualized as a bright lesion on isotropic DWI, exponential images, and fractional anisotropy, but it is seen as dark on ADC map.

Fractional anisotropy

Three-direction (orthogonal) sampling may be sufficient for routine evaluation and calculation of trace-weighted DWI and ADC. However, anisotropically restricted diffusion and information about myelin fiber orientation also may be important. If the full “tensor” or diffusion description is desired, then more than three directions of measurement are necessary to fully characterize diffusion. Typically, diffusion in a voxel is assumed to be Gaussian, which means that at least six directions should be measured. Once this diffusion tensor is acquired, scalar metrics of anisotropy can be computed. One of the most popular of these is fractional anisotropy (FA). FA maps show white matter tracts as having high anisotropy compared to gray matter (Fig. 6) (8,19,23,24).

Clinical application of DWI in ischemic stroke

As noted earlier, for reasons that are incompletely understood, water mobility decreases acutely with acute cerebral ischemia. Abnormally restricted diffusion is seen as high signal on diffusion-weighted images, and other articles in this issue of the journal highlight how this is useful in clinical practice.

In acute ischemic stroke, the hyperintensity seen on DWI is generated by two factors: decreased molecular diffusion and T2 relaxation effect (20–23,25). A few hours after stroke onset, T2 effects become prominent. Because diffusion-weighted images typically have long TE and long TR values, they have some T2 weighting as well as sensitivity to diffusion; hence, some of the “brightness” seen on diffusion-weighted images occasionally can be due to T2 effects in combination with (or, rarely, separate from) abnormally restricted diffusion. This phenomenon, also called the “T2 shine-through” effect, can cause difficulty in interpreting diffusion-weighted images. When this is an issue, the T2 effect on DWI can be mitigated by viewing images with no T2 weighting, such as on ADC maps and/or exponential images (22,26–29).

Although DWI can be applied successfully to many pathological processes, cerebral infarcts are the most common and important application fields of the technique. In addition to significant advantages over conventional MR, DWI has a high detection rate for ischemic stroke, ranging from 97% to 100%(30–34). As discussed elsewhere in this issue, the high sensitivity and specificity arise from the technique's ability to discriminate what is presumed to be cytotoxic edema from what is presumed to be vasogenic edema (35,36). Because cytotoxic edema is the hallmark of acute ischemia, DWI can show infarct within minutes (37).

DWI can differentiate acute from chronic infarcts (Figs. 7 and 8). ADC values decrease by approximately 40–50% within 24 hours after stroke onset, remain decreased for 1–2 weeks (16,38), but then rise to values similar to those of normal tissue (termed “pseudonormalization”) and then to values above normal. These values remain elevated in chronic infarcts (19). During the period of pseudonormalization, DWI alone may look normal, although the signal intensity of the infarct on T2 will be obviously elevated (Table 1).

T1-3
TABLE 1:
Temporal evaluation of abnormal signal of acute, subacute, and chronic infarct based on DWI, ADC map, FLAIR, and T2-weighted images
F7-3
FIG. 7.:
Acute versus chronic infarcts. A 69-year-old man had a history of multiple prior cerebrovascular events and right-sided weakness. Diffusion-weighted imaging (DWI) MRI obtained 2.5 hours after onset revealed acute infarct in the left posterior temporal lobe (white arrow), which was not seen apparently on FLAIR images. The patient also had small old infarcts on right periventricular white matter (black arrow). The chronic infarct is seen as a bright lesion on isotropic DWI and ADC map due to T2 shine-through effect and ADC elevation, respectively.
F8-3
FIG. 8.:
Change of ADC value in acute and chronic infarct on different maps. A 52-year-old man with a history of right-sided weakness and transient neurological disturbance was scanned 4 hours and then 3 weeks after onset. Early diffusion-weighted imaging (DWI) showed a hyperintense lesion in the left medial temporal lobe, which is consistent with acute infarct of left posterior cerebral artery. The abnormality seen on DWI is not significant on FLAIR image. The hyperintense lesion seen on early isotropic DWI changed to hypointense lesion on follow-up DWI due to ADC elevation.

On rare occasions, acute ischemic lesions as seen on DWI reverse and disappear partially or completely (39–42). This appears to occur only if the ischemia is reversed early in the course of infarction. This implies that, in the absence of treatment or spontaneous reperfusion, a decreased ADC value in stroke indicates that the tissue will go on to infarction. As described in the accompanying articles in this issue of the journal, in many patients, tissue that is not yet abnormal on diffusion imaging at the time of acute examination can proceed to infarction. Such tissue typically is around the diffusion “core” lesion and, in many instances, appears to have reduced blood flow or perfusion. This mismatch between the diffusion abnormality and the perfusion abnormality may provide a target for therapy and is the subject of intense investigation.

PERFUSION-WEIGHTED IMAGING

The progression from ischemia to infarct is characterized by changes in cerebral perfusion, which represents the volume of blood delivered to a unit mass of brain tissue per unit of time. Positron emission tomography (PET) and single photon emission computed tomography (SPECT) traditionally have been used to evaluate tissue perfusion. These techniques are based on a diffusible radioactive tracer and, with appropriate assumptions, can allow near-quantitative information of cerebral perfusion. However, the relatively poor spatial resolution, high cost, and ionizing radiation of these techniques have led some to investigate other approaches. Perfusion MRI, also referred to as hemodynamically weighted MRI, was developed to circumvent these limitations. It is a minimally invasive, widely available method that can have higher spatial resolution and better anatomic visualization compared to radionuclide-based techniques.

Physical and technical background of PWI

Two approaches have been used to assess tissue perfusion with MRI: arterial spin labeling technique and contrast agent bolus tracking technique.

Arterial spin labeling techniques are based on the same principle as time-of-flight angiography. Typically, the blood spins are labeled with an inversion recovery pulse or other magnetic tag. Then, the slice of interest is sampled in a way to cause saturation; thus, the incoming unsaturated spins cause signal contrast relative to the saturated background tissue on the distal slice. Accordingly, the labeled blood spins entering from the capillary bed to extracellular space produce high signal in the region (43). This effect is used to construct perfusion maps by image substraction or temporal evaluation. Spin labeling techniques have an outstanding advantage in that they do not require exogenous magnetic resonance contrast agents. Although the longer imaging time (10–15 min) and lower spatial resolution are current limitations of the technique, it appears to allow quantification of normal-range cerebral blood flow (which is challenging with the susceptibility-based technique) (44,45) and may be able to quantify changes in diseased blood flow states as well (46–48).

Contrast agent bolus tracking techniques, also called dynamic contrast susceptibility-weighted perfusion imaging, is the most readily available and commonly used method at present. The technique is based on the susceptibility effect of paramagnetic contrast agents (nondiffusible tracers), such as gadolinium chelate or dysprosium. Magnetic resonance contrast agents are used most often in the brain to detect disruption in the blood–brain barrier (BBB) or in the macrovasculature using the T1 effects of the contrast agent. However, intravascular spins represent <5% of total MRI spins from tissue; hence, T1-based perfusion techniques have an inherent limitation. Susceptibility-weighted perfusion imaging technique can overcome this limitation by using T2 or T2* relaxation effect of the paramagnetic agent. Although the contrast agent is confined to the intravascular compartment, its susceptibility effect extends beyond the vessel walls. T2 and T2* relaxation effects disturb the local magnetic inhomogeneity of uniform magnetic fields and cause signal loss. Therefore, T2-and T2*-weighted imaging sequences are used in perfusion MRI. According to both simulation studies and measurements, gradient-echo (GRE) sequence has been found to be sensitive to susceptibility changes in any size vessel, whereas SE sequence is preferentially sensitive to vessels <30 μm in diameter (49). Hence, when the capillary bed is of interest, the SE sequence is preferred routinely rather than GRE. On the other hand, gadolinium-based contrast causes greater signal changes when used with GRE techniques, so GRE can be useful if only a limited dose of contrast can be administered or if predominantly large vessel flow is of interest.

In practice, a magnetic resonance contrast agent is administered as an intravenous bolus, preferably by a power injector at 5 cc/s (50). Because the signal change is proportional to contrast agent concentration, a double dose (0.2 mmol/kg) provides a better signal-to-noise ratio. After injection, the first-pass bolus of contrast agent through the cerebral vascular bed is monitored by the fast MRI technique. The time interval between images should be approximately 1–2 seconds, because the time course of passage of blood to the capillary bed occurs within a few seconds. Echo-planar imaging typically is used because of its good temporal resolution and multislice imaging facility, which allows 10–13 slices even with a shot-to-shot time of 1.5 seconds. If echo-planar imaging is not available, one can perform a more limited perfusion imaging with lower slice numbers and poorer resolution using standard GRE readout technique.

The raw images obtained by repeated imaging are used to generate time versus magnetic resonance signal intensity curves (Fig. 9A). This curve is converted to a time versus concentration curve (ΔR2 curve) by using the equation ΔR2(t) = -ln[S(0)/S(t)]/TE, in which S(0) and S(t) represent signal intensities at baseline and time t, and TE represents the echo time. Typically, the ΔR2 curve is analyzed further at each pixel to obtain synthetic relative perfusion maps, such as cerebral blood flow, cerebral blood volume, mean transit time, and time to peak (51–54). As in diffusion imaging, postprocessing is necessary to construct these maps. As with DWI, some of this can be automated, although not to the same degree due to the need to specify certain input parameters in order to generate the highest quality images. Some of the maps that can be created include the following.

F9-3
FIG. 9.:
Postprocessing steps after raw perfusion data are obtained. A: MR signal intensity versus time curve is generated for each voxel by measuring region of interest. B: Concentration versus time curve (ΔR2) is obtained using susceptibility physics and tracer kinetics principles. These data are used to construct the relative cerebral blood volume (rCBV) map, which is related to relative cerebral blood flow (rCBF) and mean transit time (MTT) by the central volume equation.

Cerebral blood flow (CBF) represents the capillary flow in the tissue. Normal absolute CBF is sustained around 50–55 ml/100 g/min. The decrease of CBF value below 10–12 ml/100 g/min leads to membrane pump failure and cell death (55). Any change between these ranges activates cerebral autoregulation, in which vasodilation and increased oxygen extraction play major roles.

Cerebral blood volume (CBV) indicates the volume of cerebral capillaries and venules. Absolute CBV typically is measured in milliliters per 100 g. The CBV map can be obtained by the mathematical integration of the area under the concentration versus time curve (Fig. 9B). Any incident that increases the blood volume in the microvascular bed, such as vasodilation, can increase CBV. CBV also appears to decrease with decreasing CBF (56,57).

Mean transit time (MTT) is defined as the measurement of the length of the time that the blood spends in the cerebral capillary bed. Absolute MTT is expressed in units of time, typically seconds. An MTT map can be generated by using the central volume equation (MTT = rCBV/rCBF).

Time to peak (TTP) represents the length of the time from the start of injection until the contrast agent reaches the highest signal change (and presumably the highest concentration). It typically highlights inflow delays in the circulation.

Technical limitations

Routinely, the maps obtained by perfusion MRI provide qualitative and semiquantitative information on cerebral circulation. Moreover, individual factors, such as cardiac output variations, vascular occlusions, or variations in collateral circulation, can make the technique even more complicated to interpret. For instance, a major arterial occlusion can cause broadening of the concentration versus time curve due to collateral circulation and lead to underestimation of CBV and CBF. In practice, absolute quantitation of CBV, CBF, and MTT is difficult. Traditional diffusible tracer techniques, such as PET and xenon CT, can only provide quantitation when certain assumptions are met, and these assumptions typically are not present in pathological states such as acute ischemic stroke. Although there are efforts to quantitate CBF using bolus techniques with MRI (58,59), this is still somewhat controversial. Therefore, the term of relative r often is used to remind users of the relative nature of the perfusion in MRI maps (rCBV, rCBF, etc.).

CBF maps are a relatively recent advance that typically require use of an arterial input function, the name given to measurement from a region of interest over a major intracranial vessel (60–63). This allows deconvolution of the tissue-time concentration curve by application of indicator dilution theory (64). CBF maps and the resulting ability to create more accurate maps of MTT appear to be a substantial improvement of perfusion MRI.

One of the limitations of susceptibility-based techniques is the susceptibility artifact due to temporal bone and paranasal sinuses. The technique also requires a good intravenous access, preferably a 18-gauge catheter for rapid bolus injection. Another complicating factor of contrast agent bolus tracking technique is a “leaking” problem, which arises from severe BBB breakdown. Because of high permeability, the contrast agent leaks into the extravascular space and produces T1 enhancement. These signals counteract the T2 signal attenuation effect of contrast agent and can result in underestimation of CBV values. Presaturation of leaky regions by a small amount of contrast media or postprocessing correction algorithms may help to achieve corrected rCBV maps (65). Other solutions to make the sequence less T1 sensitive are increasing TR (time of repetition) and using nongadolinium-based contrast agents such as dysprosium. This typically is not needed in acute ischemic stroke, but rather in diseases with more pronounced BBB breakdown, such as brain tumors.

Clinical usefulness of PWI

The precise relationship among rCBV, rCBF, and MTT is at once mathematically simple and biologically complex (66). Each factor has a complementary and proportional role in delineation of cerebral hemodynamic autoregulation and stroke evaluation. The degree and duration of perfusion disturbance determine the severity and extent of tissue damage. Under physiological conditions, there is coupling and a linear relationship between CBV and CBF (67–69) Hemodynamic failure causes uncoupling of these parameters. According to PET studies, the infarct core represents severely reduced CBF (10–12 ml/100 g/min) and decreased metabolic activity. However, peri-infarcted tissue can show low CBF (12–20 ml/100 g/min) and high metabolic activity, depending on the stage of infarct (70,71). The metabolically challenged tissue, often thought of as an ischemic penumbra, may represent ischemic, dysfunctional, but potentially salvageable tissue. Some have further postulated that this penumbra is circumscribed by a mildly ischemic but functioning “oligemic” brain parenchyma. The exact extent and duration of these various zones of tissue are controversial. For example, viable tissue has been shown within a type of penumbra up to 48 hours after stroke (72,73). These investigators and others argue that this 48-hour period represents the therapeutic window for thrombolytic therapy. On the other hand, numerous clinical trials (74–76) demonstrated that administering intravenous recombinant tissue plasminogen activator (rt-PA) after 3 hours is not efficacious, at least in large groups of patients.

Some investigators believe the resolution to this paradox is patient selection. In particular, some hope that the combination of diffusion and perfusion MRI will help in evaluation of the ischemic lesion and selection of patients who may benefit from treatment. In this manner, the benefit might be maximized and the complication rate minimized. Reduced rCBV and rCBF, as well as prolonged MTT of blood, typically are seen in ischemic stroke (Figs. 10 and 11). The large perfusion and a small DWI abnormality (called diffusion-perfusion mismatch) have been reported in some small series to be present in two thirds of the stroke lesions in the hyperacute stage (77–79). Although larger series are required, patients with a DWI/PWI mismatch may be the most likely candidates to benefit from thrombolytic therapy (80). A larger diffusion-perfusion mismatch may represent a higher possibility for infarct enlargement (Fig. 12) (81). Although DWI on average underestimates the ultimate infarct volume, and rCBF and MTT on average overestimate it, these early results almost certainly are generalizations; the exact classification of patients into treatment groups awaits larger studies. Nevertheless, even with the current data, there are clearly some benefits to PWI. For example, rCBV has been found to be more accurate than DWI in predicting final infarct size (79). PWI is well correlated with early and chronic neurological outcome; therefore, it may play an important role in predicting a persisting neurological deficit (82).

F10-3
FIG. 10.:
Appearance of acute infarct on perfusion imaging. These images are from the same patient as shown on Fig. 8. Perfusion maps document reduced relative cerebral blood volume (rCBV), relative cerebral blood flow (rCBF), and prolonged mean transit time (MTT; white arrow). There is no significant mismatch between diffusion-weighted imaging and perfusion-weighted imaging. Computed tomographic angiography also supports occlusion of the left proximal posterior cerebral artery (black arrow).
F11-3
FIG. 11.:
Additional example of perfusion abnormality. A 68-year-old man with a history of heart valve disease was examined 3 days after onset. Diffusion-weighted imaging and ADC map show punctate foci of signal abnormalities on left corona radiata suggesting multiple acute embolic infarcts (large arrow). Magnetic resonance angiography demonstrates a stenosis of the superior and inferior division of the left middle cerebral artery. Perfusion-weighted imaging shows reduced relative cerebral blood volume and relative cerebral blood flow, and prolonged MTT abnormalities in the same territory. Perfusion abnormality is larger than diffusion abnormality.
F12-3
FIG. 12.:
Perfusion-diffusion mismatch and infarct enlargement. A 59-year-old man presented with a sudden onset of left hemiplegia and speech disturbance. Diffusion-weighted imaging shows an acute infarct in right middle cerebral artery distribution. Perfusion images demonstrate moderately reduced relative cerebral blood volume, significantly reduced relative cerebral blood flow, and extensively prolonged mean transit time. The area of diffusion and perfusion mismatch is more prominent at the posterior aspect of the lesion, which represents tissue at risk.

In conclusion, DWI and PWI can provide useful information that not only shows anatomical localization and characterization of the ischemic lesion, but also helps in our understanding of underlying physiopathology. These techniques may be important adjunctive methods in patient selection and follow-up for newly highlighted therapies.

REFERENCES

1. Bryan RN, Levy LM, Whitlow WD, et al. Diagnosis of acute cerebral infarction: Comparison of CT and MR imaging. Am J Neuroradiol 1991; 12:611–20.
2. Yuh WT, Crain MR, Loes DJ, et al. MR imaging of cerebral ischemia: Findings in the first 24 hours. Am J Neuroradiol 1991; 12:621–9.
3. Fisher M, Bogousslavsky J. Evolving toward effective therapy for acute ischemic stroke. JAMA 1993; 270:360–4.
4. Minematsu K, Fisher M, Li L, et al. Effects of a novel NMDA antagonist on experimental stroke rapidly and quantitatively assessed by diffusion-weighted MRI. Neurology 1993; 43:397–403.
5. Silver B, Weber J, Fisher M. Medical therapy for ischemic stroke. Clin Neuropharmacol 1996; 19:101–28.
6. Chien D, Kwong KK, Gress DR, et al. MR diffusion imaging of cerebral infarction in humans. Am J Neuroradiol 1992; 13:1097–102.
7. Ulug AM, Beauchamp Jr, N Bryan RN, et al. Absolute quantitation of diffusion constants in human stroke. Stroke 1997; 28:483–90.
8. Sorensen AG, Wu O, Copen WA, et al. Human acute cerebral ischemia: Detection of changes in water diffusion anisotropy by using MR imaging. Radiology 1999; 212:785–92.
9. Matsuoka Y, Hossmann KA. Cortical impedance and extracellular volume changes following middle cerebral artery occlusion in cats. J Cereb Blood Flow Metab 1982; 2:466–74.
10. Branston NM, Strong AJ, Symon L. Impedance related to local blood flow in cerebral cortex. J Physiol (Lond) 1978; 275:81–2.
11. Nicholson C, Sykova E. Extracellular space structure revealed by diffusion analysis. Trends Neurosci 1998; 21:207–15.
12. Sykova E, Svoboda J, Polak J, et al. Extracellular volume fraction and diffusion characteristics during progressive ischemia and terminal anoxia in the spinal cord of the rat. J Cereb Blood Flow Metab 1994: 14:301–11.
13. Armitage PA, Bastin ME, Marshall I, et al. Diffusion anisotropy measurements in ischaemic stroke of the human brain. MAGMA 1998; 6:28–36.
14. Wick M, Nagatomo Y, Prielmeier F, et al. Alteration of intracellular metabolite diffusion in rat brain in vivo during ischemia and reperfusion. Stroke 1995; 26:1930–3.
15. Duong TQ, Ackerman JJ, Ying HS, et al. Evaluation of extra-and intracellular apparent diffusion in normal and globally ischemic rat brain via 19F NMR. Magn Reson Med 1998; 40:1–13.
16. Yang Q, Tress BM, Barber PA, et al. Serial study of apparent diffusion coefficient and anisotropy in patients with acute stroke. Stroke 1999; 30:2382–90.
17. Stejskal EO, Tanner JE. Spin diffusion measurements: Spin echoes in the presence of a time-dependent field gradient. J Chem Phys 1965; 42:288–92.
18. Warach S, Gaa J, Siewert B, et al. Acute human stroke studied by whole brain echo planar diffusion-weighted magnetic resonance imaging. Ann Neurol 1995; 37:231–41.
19. Warach S, Boska M, Welch KM. Pitfalls and potential of clinical diffusion-weighted MR imaging in acute stroke. Stroke 1997; 28:481–2.
20. Le Bihan D. Diffusion/perfusion MR imaging of the brain: From structure to function. Radiology 1990; 177:328–9.
21. van Gelderen P, de Vleeschouwer MH, DesPres D, et al. Water diffusion and acute stroke. Magn Reson Med 1994; 31:154–63.
22. Provenzale JM, Engelter ST, Petrella JR, et al. Use of MR exponential diffusion-weighted images to eradicate T2 “shine-through” effect. Am J Roentgenol 1999; 172:537–9.
23. Pierpaoli C, Jezzard P, Basser PJ, et al. Diffusion tensor MR imaging of the human brain. Radiology 1996; 201:637–48.
24. Basser PJ, Pierpaoli C. A simplified method to measure the diffusion tensor from seven MR images. Magn Reson Med 1998; 39:928–34.
25. Basser PJ, Pierpaoli C. Microstructural and physiological features of tissues elucidated by quantitative-diffusion-tensor MRI. J Magn Reson B 1996; 111:209–19.
26. Le Bihan D, Breton E, Lallemand D, et al. MR imaging of intravoxel incoherent motions: Application to diffusion and perfusion in neurologic disorders. Radiology 1986; 161:401–7.
27. Fischer M, Bockhorst K, Hoehn-Berlage M, et al. Imaging of the apparent diffusion coefficient for the evaluation of cerebral metabolic recovery after cardiac arrest. Magn Reson Imaging 1995; 13:781–90.
28. Kuroiwa T, Nagaoka T, Ueki M, et al. Different apparent diffusion coefficient: Water content correlations of gray and white matter during early ischemia. Stroke 1998; 29:859–65.
29. Pierpaoli C, Alger JR, Righini A, et al. High temporal resolution diffusion MRI of global cerebral ischemia and reperfusion. J Cereb Blood Flow Metab 1996; 16:892–905.
30. Jones SC, Perez-Trepichio AD, Xue M, et al. Magnetic resonance diffusion-weighted imaging: Sensitivity and apparent diffusion constant in stroke. Acta Neurochir Suppl (Wien) 1994; 60:207–10.
31. van Everdingen KJ, van der Grond J, Kappelle LJ, et al. Diffusion-weighted magnetic resonance imaging in acute stroke. Stroke 1998; 29:1783–90.
32. Lovblad KO, Laubach HJ, Baird AE, et al. Clinical experience with diffusion-weighted MR in patients with acute stroke. Am J Neuroradiol 1998; 19:1061–6.
33. Gonzalez RG, Schaefer PW, Buonanno FS, et al. Diffusion-weighted MR imaging: Diagnostic accuracy in patients imaged within 6 hours of stroke symptom onset. Radiology 1999; 210:155–62.
34. Ricci PE, Burdette JH, Elster AD, et al. A comparison of fast spin-echo, fluid-attenuated inversion-recovery, and diffusion-weighted MR imaging in the first 10 days after cerebral infarction. Am J Neuroradiol 1999; 20:1535–42.
35. Le Bihan D, Turner R, Douek P, et al. Diffusion MR imaging: Clinical applications. Am J Roentgenol 1992; 159:591–9.
36. Provenzale JM, Sorensen AG. Diffusion-weighted MR imaging in acute stroke: Theoretic considerations and clinical applications. Am J Roentgenol 1999; 173:1459–67.
37. Warach S, Chien D, Li W, et al. Fast magnetic resonance diffusion-weighted imaging of acute human stroke. Neurology 1992; 42:1717–23.
38. Marks MP, de Crespigny A, Lentz D, et al. Acute and chronic stroke: Navigated spin-echo diffusion-weighted MR imaging. Radiology 1996; 199:403–8.
39. Woolfenden AR, O'Brien MW, Schwartzberg RE, et al. Diffusion-weighted MRI in transient global amnesia precipitated by cerebral angiography. Stroke 1997; 28:2311–4.
40. Kidwell CS, Alger JR, Di Salle F, et al. Diffusion MRI in patients with transient ischemic attacks. Stroke 1999; 30:1174–80.
41. Doege CA, Kerskens CM, Romero BI,et al. MRI of small human stroke shows reversible diffusion changes in subcortical gray matter. Neuroreport 2000; 11:2021–4.
42. Ay H, Buonanno FS, Rordorf G, Schaefer PW, et al. Normal diffusion-weighted MRI during stroke-like deficits. Neurology 1999; 52:1784–92.
43. Edelman RR, Siewert B, Darby DG, et al. Qualitative mapping of cerebral blood flow and functional localization with echo-planar MR imaging and signal targeting with alternating radio frequency. Radiology 1994; 192:513–20.
44. Kwong KK, Chesler DA, Weisskoff RM, et al. MR perfusion studies with T1-weighted echo planar imaging. Magn Reson Med 1995; 34:878–87.
45. Alsop DC, Detre JA. Multisection cerebral blood flow MR imaging with continuous arterial spin labeling. Radiology 1998; 208:410–6.
46. Chalela JA, Alsop DC, Gonzalez-Atavales JB, et al. Magnetic resonance perfusion imaging in acute ischemic stroke using continuous arterial spin labeling. Stroke 2000; 31:680–7.
47. Wong EC, Buxton RB, Frank LR. Quantitative perfusion imaging using arterial spin labeling. Neuroimaging Clin North Am 1999; 9:333–42.
48. Detre JA, Alsop DC. Perfusion magnetic resonance imaging with continuous arterial spin labeling: Methods and clinical applications in the central nervous system. Eur J Radiol 1999; 30:115–24.
49. Fisel CR, Ackerman JL, Buxton RB, et al. MR contrast due to microscopically heterogeneous magnetic susceptibility: Numerical simulations and applications to cerebral physiology. Magn Reson Med 1991; 17:336–47.
50. Saini S, Fretz CJ, Fisel CR, et al. In vitro evaluation of a mechanical injector for infusion of magnetic resonance contrast media. Invest Radiol 1991; 26:748–51.
51. Rosen BR, Belliveau JW, Vevea JM, et al. Perfusion imaging with contrast agents. Magn Reson Med 1990; 14:249–65.
52. Rosen BR, Belliveau JW, Buchbinder BR, et al. Contrast agents and cerebral hemodynamics. Magn Reson Med 1991; 19:285–92.
53. Rosen BR, Aronen HJ, Kwong KK, et al. Advances in clinical neuroimaging: Functional MR imaging techniques. Radiographics 1993; 13:889–96.
54. Aronen HJ, Gazit IE, Louis DN, et al. Cerebral blood volume maps of gliomas: Comparison with tumor grade and histologic findings. Radiology 1994; 191:41–51.
55. Jones TH, Morawetz RB, Crowell RM, et al. Thresholds of focal cerebral ischemia in awake monkeys. J Neurosurg 1981; 54:773–82.
56. Caramia F, Huang Z, Hamberg LM, et al. Mismatch between cerebral blood volume and flow index during transient focal ischemia studied with MRI and GD-BOPTA. Magn Reson Imaging 1998; 16:97–103.
57. Zaharchuk G, Mandeville JB, Bogdanov Jr, AA et al. Cerebrovascular dynamics of autoregulation and hypoperfusion. An MRI study of CBF and changes in total and microvascular cerebral blood volume during hemorrhagic hypotension. Stroke 1999; 30:2197–205.
58. Lythgoe DJ, Ostergaard L, William SC, et al. Quantitative perfusion imaging in carotid artery stenosis using dynamic susceptibility contrast-enhanced magnetic resonance imaging. Magn Reson Imaging 2000; 18:1–11.
59. Simonsen CZ, Ostergaard L, Vestergaard-Poulsen P, et al. CBF and CBV measurements by USPIO bolus tracking: Reproducibility and comparison with Gd-based values. J Magn Reson Imaging 1999; 9:342–7.
60. Ostergaard L, Weisskoff RM, Chesler DA, et al. High resolution measurement of cerebral blood flow using intravascular tracer bolus passages. Part I: Mathematical approach and statistical analysis. Magn Reson Med 1996; 36:715–25.
61. Ostergaard L, Sorensen AG, Kwong KK, et al. High resolution measurement of cerebral blood flow using intravascular tracer bolus passages. Part II: Experimental comparison and preliminary results. Magn Reson Med 1996; 36:726–36.
62. Wenz F, Rempp K, Hess T, et al. Effect of radiation on blood volume in low-grade astrocytomas and normal brain tissue: Quantification with dynamic susceptibility contrast MR imaging. Am J Roentgenol 1996; 166:187–93.
63. Sorensen AG, Copen WA, Ostergaard L, et al. Hyperacute stroke: Simultaneous measurement of relative cerebral blood volume, relative cerebral blood flow, and mean tissue transit time. Radiology 1999; 210:519–27.
64. Rempp KA, Brix G, Wenz F, et al. Quantification of regional cerebral blood flow and volume with dynamic susceptibility contrast-enhanced MR imaging. Radiology 1994; 193:637–41.
65. Lev MH, Rosen BR. Clinical applications of intracranial perfusion MR imaging. Neuroimaging Clin North Am 1999; 9:309–31.
66. Weisskoff RM, Chesler D, Boxerman JL, et al. Pitfalls in MR measurement of tissue blood flow with intravascular tracers: Which mean transit time? Magn Reson Med 1993; 29:553–8.
67. Baron JC, Rougemont D, Soussaline F, et al. Local interrelationships of cerebral oxygen consumption and glucose utilization in normal subjects and in ischemic stroke patients: A positron tomography study. J Cereb Blood Flow Metab 1984; 4:140–149.
68. Sette G, Baron JC, Mazoyer B, et al. Local brain haemodynamics and oxygen metabolism in cerebrovascular disease. Positron emission tomography. Brain 1989; 112(pt 4):931–51
69. Gibbs JM, Wise RJ, Leenders KL, et al. Evaluation of cerebral perfusion reserve in patients with carotid-artery occlusion. Lancet 1984; 11:310–4.
70. Heiss WD, Huber M, Fink GR, et al. Progressive derangement of periinfarct viable tissue in ischemic stroke. J Cereb Blood Flow Metab 1992; 12:193–203.
71. Hakim AM, Evans AC, Berger L, et al. The effect of nimodipine on the evolution of human cerebral infarction studied by PET. J Cereb Blood Flow Metab 1989; 9:523–34.
72. Marchal G, Beaudouin V, Rioux P, et al. Prolonged persistence of substantial volumes of potentially viable brain tissue after stroke: A correlative PET-CT study with voxel-based data analysis. Stroke 1996; 27:599–606.
73. Furlan M, Marchal G, Viader F, et al. Spontaneous neurological recovery after stroke and the fate of the ischemic penumbra. Ann Neurol 1996; 40:216–26.
74. Steiner T, Bluhmki E, Kaste M. The ECASS 3-hour cohort. Secondary analysis of ECASS data by time stratification. ECASS Study Group. European Cooperative Acute Stroke Study. Cerebrovasc Dis 1998; 8:198–203.
75. Osborn TM, LaMonte MP, Gaasch WR. Intravenous thrombolytic therapy for stroke: A review of recent studies and controversies. Ann Emerg Med 1999; 34:244–55.
76. Hacke W, Brott T, Caplan L, et al. Thrombolysis in acute ischemic stroke: Controlled trials and clinical experience. Neurology 1999; 53(7 suppl 4):S3–14.
77. Barber PA, Darby DG, Desmond PM, et al. Prediction of stroke outcome with echoplanar perfusion-and diffusion-weighted MRI. Neurology 1998; 51:418–26.
78. Albers GW. Expanding the window for thrombolytic therapy in acute stroke. The potential role of acute MRI for patient selection. Stroke 1999; 30:2230–7.
79. Rordorf G, Koroshetz WJ, Copen WA, et al. Regional ischemia and ischemic injury in patients with acute middle cerebral artery stroke as defined by early diffusion-weighted and perfusion-weighted MRI. Stroke 1998; 29:939–43.
80. Jansen O, Schellinger P, Fiebach J, et al. Early recanalisation in acute ischaemic stroke saves tissue at risk defined by MRI. Lancet 1999; 12:2036–7.
81. Baird AE, Benfield A, Schlaug G, et al. Enlargement of human cerebral ischemic lesion volumes measured by diffusion-weighted magnetic resonance imaging. Ann Neurol 1997; 41:581–9.
82. Marks MP, Tong DC, Beaulieu C, et al. Evaluation of early reperfusion and i.v. tPA therapy using diffusion-and perfusion-weighted MRI. Neurology 1999; 52:1792–8.

Section Description

Issue Editor: A. Gregory Sorensen

Keywords:

Magnetic resonance imaging; Diffusion imaging; Perfusion imaging; Ischemic stroke; Cerebrovascular disorder

© 2000 Lippincott Williams & Wilkins, Inc.