Dynamic Contrast-enhanced Susceptibility-weighted Echo-planar ...

State of the Art Soonmee Cha, MD Edmond A. Knopp, MD Glyn Johnson, PhD Stephan G. Wetzel, Dr med Andrew W. Litt, MD David Zagzag, MD, PhD

Index terms: Brain, MR, 13.121411, 13.121412, 13.121416, 13.12143, 13.12144 Brain neoplasms, 13.343, 13.363, 13.38 Lymphoma, 13.343 Magnetic resonance (MR), perfusion study, 13.121416, 13.12143, 13.12144 State of the Art Published online before print 10.1148/radiol.2231010594 Radiology 2002; 223:11–29 Abbreviations: CBV ⫽ cerebral blood volume GBM ⫽ glioblastoma multiforme PCL ⫽ primary cerebral lymphoma rCBV ⫽ relative CBV ROI ⫽ region of interest ⌬R2* ⫽ change in relaxation rate TDL ⫽ tumefactive demyelinating lesion

Intracranial Mass Lesions: Dynamic Contrast-enhanced Susceptibility-weighted Echo-planar Perfusion MR Imaging1 Dynamic contrast agent– enhanced perfusion magnetic resonance (MR) imaging provides physiologic information that complements the anatomic information available with conventional MR imaging. Analysis of dynamic data from perfusion MR imaging, based on tracer kinetic theory, yields quantitative estimates of cerebral blood volume that reflect the underlying microvasculature and angiogenesis. Perfusion MR imaging is a fast and robust imaging technique that is increasingly used as a research tool to help evaluate and understand intracranial disease processes and as a clinical tool to help diagnose, manage, and understand intracranial mass lesions. With the increasing number of applications of perfusion MR imaging, it is important to understand the principles underlying the technique. In this review, the essential underlying physics and methods of dynamic contrast-enhanced susceptibility-weighted echo-planar perfusion MR imaging are described. The clinical applications of cerebral blood volume maps obtained with perfusion MR imaging in the differential diagnosis of intracranial mass lesions, as well as the pitfalls and limitations of the technique, are discussed. Emphasis is on the clinical role of perfusion MR imaging in providing insight into the underlying pathophysiology of cerebral microcirculation. ©

RSNA, 2002

1

From the Departments of Radiology (S.C., E.A.K., G.J., S.G.W., A.W.L.), Neurosurgery (E.A.K., D.Z.), and Pathology (D.Z.), New York University Medical Center, 530 First Ave, HCCBasement, MRI Center, New York, NY 10016. Received March 12, 2001; revision requested April 27; revision received June 25; accepted July 16. S.C. supported in part by an RSNA Seed Grant 3; S.G.W. supported in part by Swiss National Science Foundation/ Karger Stiftung and by Novartis Stiftung. Address correspondence to S.C. (e-mail: [email protected]).

©

RSNA, 2002

Recent advances in neuroimaging allow the acquisition of physiologic maps that complement the anatomic information provided by traditional techniques. In particular, the development of fast magnetic resonance (MR) imaging methods provides a means of obtaining high-spatial-resolution images of water self-diffusion, axonal organization, and blood oxygenation. Of particular clinical use are the maps of cerebral microcirculation that can be constructed from dynamic contrast agent– enhanced perfusion MR imaging. Several parameters (blood volume, transit time, clearance, extraction fraction, blood flow, and permeability–surface area product) can, in principle, be derived by following the passage of a tracer or an indicator such as an MR contrast agent through the cerebrovascular system (1). Strictly, cerebral perfusion is defined as the steady-state delivery of nutrients and oxygen via blood to brain tissue parenchyma per unit volume or mass and is typically measured in milliliters per 100 g of tissue per minute (2). In perfusion MR imaging, however, the term “perfusion” is broadly applied to include several tissue microcirculatory hemodynamic parameters (cerebral blood volume [CBV], cerebral blood flow, and mean transit time) that can be estimated from the dynamic data. In the evaluation of intracranial mass lesions, however, CBV appears to be the most useful parameter. Although a detailed description of the fundamentals of tracer kinetics used in dynamic imaging is beyond the scope of this review (readers can refer to several excellent reviews on this topic [1,3,4]), the essential principles will be discussed. Also, we will focus on clinical applications of perfusion MR imaging and how the technique can be helpful in the diagnosis, management, and understanding of various intracranial mass lesions. Limitations and potential pitfalls will be discussed along with appropriate strategies to overcome some of the weaknesses of the technique. When available, imaging findings will be 11

correlated with histopathologic findings, focusing on the vascular morphology and status of angiogenesis of various intracranial mass lesions.

TECHNICAL CONSIDERATIONS Contrast Agents for Perfusion MR Imaging Perfusion MR imaging methods exploit signal changes that accompany the passage of tracer through the cerebrovascular system. The tracer can be either endogenous (arterial water) or exogenous (deuterium oxide, gadopentetate dimeglumine) and either freely diffusible (arterial water, deuterium oxide) or nondiffusible (gadopentetate dimeglumine). In each case, the analysis is based on the indicator-dilution methods originally developed for the radioisotope measurements of blood flow and volume (5). Endogenous or arterial spin-labeling techniques such as echo-planar MR imaging and signal targeting with alternating radio frequency, or EPISTAR, and flow-sensitive alternating inversion-recovery, or FAIR, magnetically label blood spins upstream from the imaging section with either inverting or saturating radiofrequency pulses. In this approach, the regional changes in signal intensity are determined by means of an interaction between blood flow and T1 relaxation. Comparison between images acquired with and those acquired without labeling allow calculation of tissue perfusion (6 – 10). Arterial spin labeling does not require injection of an exogenous contrast agent, can be performed without additional hardware, and, in principle, yields fully quantitative measurements of cerebral blood flow. However, the necessarily long transit delay between the application of the labeling pulse and the arrival of labeled blood in the imaging section allows T1 recovery, which decreases perfusion contrast. Even with signal averaging (which increases imaging time), arterial spin-labeling cerebral blood flow maps suffer from a poor signal-to-noise ratio and have not been widely used clinically. Techniques in which freely diffusible exogenous tracers such as deuterium oxide (11), 17H2O (12) or trifluoromethane (13,14) are used have also been proposed. However, these are either expensive or suffer from poor signal-to-noise ratio and, again, have not been used widely in clinical practice. One of the oldest methods of contrastenhanced perfusion MR imaging, originally developed by Villringer, Rosen, and 12

䡠

Radiology

䡠

April 2002

colleagues (15,16), uses conventional MR imaging contrast agents such as gadopentetate dimeglumine. The change in relaxation times during passage of a contrast agent bolus can be converted to tissue contrast agent concentration. The theory of tracer kinetics is then used to derive CBV values from the concentration-time curve. This method has the advantage that CBV can be measured without knowledge of the arterial input function and provides a better signal-tonoise ratio than do spin-labeling techniques. Although quantitative estimates of CBV are relative rather than absolute, CBV measurements have shown to correlate with the vascularity of various intracranial mass lesions. Contrast-enhanced perfusion MR imaging will be further discussed later in this section. An entirely different approach based on pharmacokinetic modeling was developed by Tofts and Kermode (17) and others (18,19). These techniques were developed primarily to measure vascular permeability and may in the future provide valuable information on brain tumors in which angiogenesis leads to the formation of immature and permeable blood vessels. To date, however, the methods have not been widely used.

Sequence Considerations For CBV measurements, a series of images are acquired at approximately 1-second intervals before, during, and after the bolus injection of contrast agent. Longer intervals may be feasible, but the signal intensity–time curve would then be measured less accurately. Rapid gradient-echo imaging is capable of generating approximately two T1-weighted sections per second, which is not generally sufficient to cover a large heterogeneous tumor. Echo-planar imaging, however, is capable of generating approximately 10 MR sections every second and is ideal for rapid dynamic imaging. The passage of gadopentetate dimeglumine causes changes in both T2 and T2* so that both spin-echo and gradient-echo echo-planar sequences provide robust measurements of CBV. Gradient-echo sequences are, however, much more sensitive. When a paramagnetic contrast agent such as gadopentetate dimeglumine passes through the cerebral vascular system, it induces differences in local magnetic susceptibility between vessels and the surrounding tissue. Although the vascular space is a small fraction (4%– 5%) of the total tissue blood volume, this compartmentalization of the contrast

agent causes targeted paramagnetism within the intravascular spins, as well as in the surrounding spins within a given voxel. Thus, both intra- and extravascular spins undergo a reduction of T2* that leads to a large transient signal loss of approximately 25% in normal white matter with a standard dose of contrast agent (0.1 mmol per kilogram of body weight). T2-weighted spin-echo images are less sensitive and require a double or even quadruple dose of contrast agent in order to yield substantial signal changes during the bolus passage. On the other hand, gradient-echo sequences are more prone to magnetic susceptibility artifacts. Asymmetric spin-echo echo-planar sequences provide a potentially useful compromise between gradient-echo and spinecho echo-planar sequences. In asymmetric spin-echo echo-planar sequences, the echo center is displaced from the Hahn echo time, resulting in a mixture of T2 and T2* weighting. The degree of asymmetry can be adjusted to trade off sensitivity against susceptibility to artifacts (20,21). Thus, when imaging lesions near brain-bone-air interfaces, such as the temporal or inferior frontal lobes where these artifacts are more pronounced, spin-echo sequences may be preferable. However, artifacts on gradient-echo images can be overcome, to a large extent, by reducing section thickness (22). Although this reduces the signal-to-noise ratio, we have found that this technique still provides diagnostic images. A second advantage of spin-echo sequences is that results with simulations and phantoms suggest that spin-echo images will be sensitive only to contrast agent within the capillaries, whereas gradient-echo sequences will be sensitive to contrast agent in both capillaries and larger vessels (23). Although contamination by venous signals with gradient-echo sequences will potentially cause overestimates of CBV, it is relatively easy to identify the location of veins and perform measurements of CBV in regions of interest (ROIs) chosen to avoid them.

Data Analysis: Calculating CBV from Image Intensities It has been shown by Zierler (5) that in the absence of recirculation and contrast agent leakage, CBV is proportional to the area under the contrast agent concentration–time curve. The gadopentetate dimeglumine concentration is proportional to the change in relaxation rate (ie, change in the reciprocal of T2* [⌬R2*]), which can be calculated from the signal by usCha et al

Figure 1. Diagrams illustrate calculation of rCBV. A, Signal intensity decrease during passage of contrast agent bolus is measured from a series of gradient-echo echo-planar MR images. B, Change in the relaxation rate (⌬R2*) is calculated from signal intensity, and a baseline subtraction method is applied to measured data. C, Corrected ⌬R2* curve. D, rCBV is proportional to the area under curve (shaded area).

ing the following equation (16): ⌬R2* ⫽ [⫺ln(SIt/SI0)/TE], where SIt is the pixel signal intensity at time t, SI0 is the precontrast signal intensity, and TE is the echo time. This equation is valid only if the T1 enhancement associated with blood-brain barrier disruption has a negligible effect on signal intensity, which is ensured by using either a long repetition time, a small flip angle, or both to reduce saturation. In general, the assumptions of negligible recirculation and contrast agent leakage are violated. The effects of this can be reduced by fitting a gamma-variate function to the measured ⌬R2* curve (3). The gamma-variate function approximates the curve that would have been obtained without recirculation or leakage. CBV can then be estimated from the area under this fitted curve rather than from the original data. In our experience, however, the gamma variate fit is unstable: Small variations in the initial parameter estimates yield wide variations in the results. This instability occurs even with data averaged over ROIs in areas of high perfusion and appears to be inherent to the procedure. In practice, satisfactory fits can often be found only by repeating the procedure with multiple different initial estimates until a set is found that causes the fitting algorithm to converge. This approach can be applied to fit concentration-time curves from multiple ROIs, since each fit takes very little time, but it is not suited for pixel-by-pixel calVolume 223

䡠

Number 1

culations of CBV maps. To obtain CBV maps, alternative corrections for leakage are preferable. The simplest method is to estimate the end of the bolus and calculate the area within the bolus alone. This will, however, result in a systematic overestimate of CBV in areas where the blood-brain barrier is damaged. Alternatively, having estimated the beginning and end of the bolus, a baseline can be subtracted from under the curve. The area under the corrected contrast agent concentration–time curve is proportional to the CBV and does not yield an absolute measurement. It is necessary, therefore, to express the measurement relative to a standard reference, usually contralateral white matter. We refer to this as relative CBV (rCBV). The steps in the data analysis are, therefore, as follows (Fig 1): (a) Obtain curves of signal intensity changes over time (Fig 1, A). (b) Estimate the mean precontrast signal intensity from 10 data points acquired before arrival of the bolus; it is important to exclude the first one to two images, which are acquired during the time the steady-state MR signal is established. (c) Calculate ⌬R2* and apply baseline subtraction to the ⌬R2* curve (Fig 1, B). (d) Calculate the area under the fitted curve (Fig 1, C). (e) Calculate rCBV relative to normal white matter (Fig 1, D). At our institution, the echo-planar images are transferred to a workstation (Ultra 10; Sun Microsystems, Palo Alto,

Calif) for postprocessing. Programs developed in house with the C and IDL programming languages (Research Systems, Boulder, Colo) are used. The source images are first inspected for overall image quality and motion artifact. A single ROI is placed over the contralateral unaffected centrum semiovale white matter, and rCBV maps are generated to serve as a road map. CBV maps can then be calculated on a pixel-by-pixel basis and displayed as a gray-scale image. Figure 2a–2d presents raw-data images with changes in tissue signal intensity before, during, and after the passage of intravascular gadopentetate dimeglumine. However, small but important variations in CBV are not always apparent from these maps. An alternative is to use a color overlay displayed on the raw image (Fig 2d), in which the abnormal CBV values are often more apparent. Application of a threshold for the color overlay at a CBV of approximately 50% greater than that of the unaffected (normal) white matter allows visualization of underlying anatomy that can be helpful in interpretation. Figure 2e shows a transverse contrast-enhanced T1-weighted image in the same patient. The rCBV map (Fig 2d) reflects overall lesion vascularity, whereas the contrast-enhanced T1-weighted image (Fig 2e) demonstrates disruption or complete lack of the blood-brain barrier, as in this case of a highly vascular extraaxial tumor (choroid plexus carcinoma). Although rCBV maps are good indicators of hypervascular regions within intracranial mass lesions, values calculated on a pixel-by-pixel basis suffer from a poor signal-to-noise ratio. It is therefore preferable to calculate rCBV values from ROIs within the region. In our practice, the rCBV map is primarily used as a road map and a qualitative assessment of lesion vascularity. To obtain ROI rCBV values, multiple ROIs are drawn to obtain a qualitative impression of blood volume. The sizes of ROIs are varied, depending on the size of lesion, but usually 1–5 mmdiameter ROIs are used. The degree of signal intensity change correlates well with rCBV (24), so the region with the highest rCBV is easily identified. Furthermore, in lesions that are located close to major vascular structures, careful inspection of the signal intensity–time curve clearly indicates vessels that produce very large signal intensity changes. CBV measurements of lesions are generally calculated relative to ROI values in contralateral normal white matter. If the lesion is predominantly cortical, the reference ROI is placed over the contralat-

Intracranial Mass Lesions: Echo-planar Perfusion MR Imaging

䡠

13

Figure 2. (a– d) Transverse gradient-echo echo-planar (repetition time msec/echo time msec, 1,000/54) MR images in a 33-year-old woman with a choroid plexus carcinoma of the left lateral ventricle. (a) Image obtained before intravenous injection of gadopentetate dimeglumine. (b) Image obtained at peak arrival time (in this case, 10 seconds after injection of contrast agent) demonstrates marked signal intensity decrease in the tumor (arrow). (c) Image obtained 50 seconds after injection shows return of signal intensity to baseline except in areas of disruption or absence of the blood-brain barrier (arrows), where leakage has occurred. (d) Color overlay of rCBV map shows the intense hypervascularity (red) of the tumor. (e) Transverse contrast-enhanced T1-weighted (600/14) MR image demonstrates marked enhancement of the tumor (arrow) due to lack of a blood-brain barrier.

eral cortex rather than over the white matter. Similarly, with deep gray matter lesions in the basal ganglia or thalamus, a contralateral, uninvolved, similar location should preferably be selected as a reference point.

Imaging Protocol Three critical components are necessary for dynamic contrast-enhanced perfusion MR imaging. First, an imaging technique, such as echo-planar imaging, capable of acquiring images of multiple sections with a temporal resolution of 1 or 2 seconds is needed. Second, a power injector to ensure rapid uniform administration of contrast agent and, therefore, a well-defined and narrow bolus should 14

䡠

Radiology

䡠

April 2002

be used. Third, a software program that will allow data analysis and rCBV measurements is required. At our institution, all patients undergo placement of an 18- or 20-gauge intravenous catheter prior to imaging, usually in the antecubital area, for the purpose of contrast agent administration. Since perfusion MR imaging complements rather than replaces conventional MR imaging, our imaging protocol includes nonenhanced transverse T2-weighted (3,400/119) or fluidattenuated inversion recovery (9,000/ 110/2,500 [repetition time msec/echo time msec/inversion time msec]) and contrast-enhanced T1-weighted (600/14) sequences. The perfusion imaging protocol uses a gradient-echo susceptibility-

weighted blipped echo-planar imaging sequence. Unlike the original echo-planar imaging sequence (25), where the phase-encoding gradient is applied continuously, blipped echo-planar imaging involves a brief phase-encoding gradient “blip” between readout intervals (26). With blipped echo-planar imaging, the k-space trajectory is much more amenable to Fourier transformation. A complete discussion of echo-planar imaging and the required hardware modifications is beyond the scope of this report. Interested readers are referred to the literature (27,28). The specific imaging parameters are 1,000/54; field of view, 230 ⫻ 230 mm; section thickness, 3–7 mm; matrix, 128 ⫻ 128; in-plane voxel size, 1.8 ⫻ 1.8 mm; intersection gap, 0%–30%; flip angle, 30°; signal bandwidth, 1,470 Hz/ pixel. Five to seven sections are obtained to cover the entire lesion volume as idenCha et al

Figure 3. Transverse gradient-echo echo-planar (1,000/54) MR images show reduction of susceptibility artifacts when a smaller section thickness is used in a 34-year-old man with left temporal lobe radiation necrosis. (a) Images obtained with 6-mm section thickness and 20% intersection gap show extensive susceptibility artifacts (arrows) in the inferior frontal and temporal lobes. (b) Images obtained at same levels as in a but with 3-mm section thickness and 20% gap show the presence of a cavity (arrowheads) in the left middle cranial fossa. Susceptibility artifacts are markedly reduced.

tified on T2-weighted images. A series of 60 multisection acquisitions are acquired at 1-second intervals. The combination of a 1,000-msec repetition time and a 30° flip angle ensures that T1 effects are minimized. The first 10 acquisitions are performed prior to the contrast agent injection to establish a precontrast baseline. At the 10th acquisition, gadopentetate dimeglumine (0.1 mmol/kg) is injected with the power injector at a flow rate of 3–5 mL/sec through the intravenous catheter, the contrast agent injection is immediately followed by a bolus injection of saline (total of 20 mL at the same rate). Volume 223

䡠

Number 1

Technical Pitfalls and Limitations There are several important limitations associated with contrast-enhanced perfusion MR imaging. First, because the technique is susceptibility weighted, it is exquisitely sensitive to structures or lesions that induce strong magnetic field inhomogeneity, such as blood products, calcium, melanin, metals, or lesions near a brain-bone-air interface (eg, middle and anterior cranial fossa). One simple way to reduce inhomogeneity and susceptibility artifacts is to decrease section thickness, as shown in Figure 3, although this will also reduce the signal-to-noise ratio and

section coverage. If section coverage is insufficient to cover the tumor, the intersection gap can be increased. This does carry the risk of missing small vascular regions, but even with thicker sections such small regions may be missed due to volume averaging. Second, the cost of imaging hardware can be high, since perfusion MR imaging requires high-performance gradients and very fast echo-planar imaging sequences. Third, the calculation of rCBV can be grossly inaccurate in lesions such as glioblastoma multiforme (GBM) or meningioma, where there is a severe breakdown or absence of the blood-brain barrier (29).

Intracranial Mass Lesions: Echo-planar Perfusion MR Imaging

䡠

15

Figure 4. Left parietal lobe GBM in a 46-year-old man. (a) Transverse T2-weighted (3,400/119) MR image shows left parietal lobe mass with extensive surrounding edema (arrow). (b) Transverse contrast-enhanced T1-weighted (600/14) MR image shows enhancement of the tumor (arrow). (c) rCBV map demonstrates marked hypervascularity (red) of the tumor.

It should also be emphasized that rCBV measurements are a relative rather than an absolute quantification of blood volume.

CLINICAL APPLICATIONS Perfusion MR imaging is increasingly being used as a diagnostic and research tool that provides maps of the regional variations in cerebral microvasculature of normal and diseased brains (30 –34). With relatively short imaging and data processing times and the use of a standard dose of contrast agent, perfusion MR imaging is a promising tool that can easily be incorporated as part of the routine clinical evaluation of intracranial mass lesions. Although still investigational, MR imaging CBV measurements can be used as an adjunct to conventional imaging to help assess the degree of neovascularization in brain tumors, evaluate tumor grading and malignancy, identify tumor-mimicking lesions (such as radiation necrosis, cerebral abscess, and tumefactive demyelinating lesion [TDL]) by demonstrating their lack of angiogenesis, and assess the status of viable tissue surrounding an acute infarct. It must be emphasized, however, that perfusion MR imaging is a relatively new and promising imaging tool rather than a standard proven technique for tumor grading and staging. In the future, perfusion MR imaging may become useful in the monitoring of treatment, and its results may also potentially serve as an arbiter when de16

䡠

Radiology

䡠

April 2002

Figure 5. Signal intensity–time curve shows marked signal intensity decrease during peak arrival of the contrast agent, followed by partial recovery of the signal intensity loss (arrowheads). A second decrease in signal intensity (arrow) is due to recirculation. Recovery of signal intensity is incomplete due to leakage of contrast agent through a disrupted blood-brain barrier.

termining the efficacy of novel therapeutic agents, especially antiangiogenic therapy.

Intracranial Neoplasms Vascular morphology and the degree of angiogenesis are important elements in the evaluation of different tumor types and the determination of the biologic aggressiveness of intracranial neoplasms (35). Perfusion MR imaging provides in vivo maps of CBV that depict the overall tumor vascularity, which allows indirect assessment of tumor angiogenesis (30). MR measurements of rCBV have been

shown to correlate with both conventional angiographic assessments of tumor vascular density and histologic measurements of tumor neovascularization (30,31,36). Increased tumor vascularity, however, is not synonymous with malignancy. There are several intracranial neoplasms, especially those that are extraaxial in location (eg, meningioma, choroid plexus papilloma), that can be highly vascular but rather benign in terms of biologic behavior. With the recent interest in and development of antiangiogenic cancer therapies that directly attack tumor vessels Cha et al

Figure 6. Photomicrographs show histopathologic findings of the vascular hyperplasia of GBM. Top left: Early vascular hyperplasia remote from the glioblastoma “border.” Note endothelial cell plumping (arrows). (Hematoxylin-eosin stain; original magnification, ⫻200.) Bottom left: Angiogenesis at the edge of an infiltrating glioblastoma. Invading glioma cells (short arrows) can be seen between the hyperplastic vascular complexes (long arrows). (Hematoxylin-eosin stain; original magnification, ⫻100.) Top right: Intratumoral neovascularization showing glomeruloid vessels (long arrows). Mitotic figures (short arrows) in vascular cells are evident. (Hematoxylineosin stain; original magnification, ⫻100.) Bottom right: Azocarmine staining underscores the exuberant vascular proliferation (arrow) that is characteristic of GBM. (Original magnification, ⫻100.)

(37– 41), perfusion MR imaging can be used for a noninvasive assessment of changes in tumor rCBV during treatment and thus for monitoring the effectiveness of therapy. Conventional MR imaging is limited by its nonspecificity and inability to differentiate between tumor recurrence and therapy-related necrosis. Perfusion MR imaging, on the other hand, has been shown to correlate better with clinical response in patients undergoing antiangiogenic therapy (32). Proton (hydrogen 1 [1H]) MR spectroscopy may allow further characterization of intracranial mass lesions by providing information on metabolic changes such as those associated with cellular turnover or proliferative activity (choline), decrease in neuronal density (N-acetylaspartate), and presence of necrosis (lipids, lactate) (42). Although 1H MR spectroscopy also remains investigational in its role for accurate preoperative grading of gliomas, the presence of lipids and lactate strongly suggests high-grade tumors (43– 46). In a recent 1H MR spectroscopy study of gliomas (47), choline metabolites were positively correlated with histopathologic findings of proliferative acVolume 223

䡠

Number 1

tivity as determined with the Ki-67 labeling index. 1H MR spectroscopy is a useful and promising diagnostic modality for the metabolic characterization of brain tumors, as well as of other mass lesions of the brain. A detailed description of the clinical application of 1H MR spectroscopy is beyond the scope of this article, and interested readers are referred to the literature (48 –52).

Primary Gliomas Malignant gliomas, the most common primary brain tumors, are highly invasive and extremely vascular tumors. Gliomas are histologically heterogeneous, with components that include varying degrees of cellular and nuclear pleomorphism, mitotic activity, vascular proliferation, and necrosis (53,54). GBM, the most malignant glioma, is the most common primary brain tumor, accounting for 12%– 15% of all intracranial neoplasms (35). Despite decades of research, GBM remains resistant to therapies that have proved successful in treating other solid tumors. GBM remains one of the most

malignant and fatal neoplasms known to humanity. An important factor in the malignancy of GBM is the ability of the tumor to recruit and synthesize vascular networks for further growth and proliferation. The degree of vascular proliferation is one of the most critical elements in the determination of tumor grade and prognosis for several reasons. First, the degree of vascular proliferation, or angiogenesis, is one of the most important histologic criteria (along with cellularity, mitosis, pleomorphism, necrosis) for determination of the degree of malignancy and grade of the glioma. Second, vascular networks are not only the principal route for delivery of oxygen and nutrients to the neoplastic cells and removal of metabolic waste products, such networks also serve as a path for tumor infiltration along the perivascular spaces. Third, the cerebral capillary endothelium (site of the bloodbrain barrier)— composed of a continuous homogeneous basement membrane, numerous astrocytic processes, and tight junctions and an important host defense mechanism responsible for the regulation of movement of molecules—is frequently destroyed by malignant tumor cells. Fourth, a destroyed or altered bloodbrain barrier serves as a diagnostic tool at both computed tomography (CT) and MR imaging by allowing for contrast agent extravasation and, therefore, identification of the tumor (55,56). Preoperative assessment of glioma vascularity can be helpful in determining malignant potential and, therefore, potentially in alteration of therapeutic management. Contrast-enhanced T1-weighted MR images of the brain depict areas of a disrupted or absent blood-brain barrier (57,58). Perfusion MR imaging, on the other hand, demonstrates variations in regional CBV that reflect alterations in tumor vascularity. Figure 4 demonstrates an example of an rCBV map of a highly vascular GBM with corresponding T2-weighted and contrast-enhanced T1-weighted images. Figure 5 displays a signal intensity–time curve of the GBM in Figure 4. Figure 6 illustrates histopathologic findings of the same GBM, demonstrating striking angiogenesis and glomeruloid capillaries. In the past, catheter-based angiography has been the definitive imaging modality to depict tumor vascularity or tumor-related vasculopathies such as arteriovenous shunting or early draining veins. MR imaging rCBV maps of glioma demonstrate not only the overall vascularity of the tumor but also the inherent heterogeneity and geographic differences within a

Intracranial Mass Lesions: Echo-planar Perfusion MR Imaging

䡠

17

Figure 7. Nonenhancing right frontal lobe tumor in a 32-year-old woman. (a) Transverse T2-weighted (3,400/119) MR image shows a hyperintense mass (arrow) with minimal surrounding edema (arrowheads). (b) Transverse contrast-enhanced T1-weighted (600/14) MR image shows a homogeneously hypointense right frontal lobe tumor (arrow). (c) rCBV map demonstrates an area of increased vascularity (arrow), and results of a biopsy directed to that area showed histologic vascular proliferation. The patient subsequently underwent surgical resection, and the diagnosis after pathologic examination was anaplastic astrocytoma.

single tumor. These CBV maps, in conjunction with conventional MR images, can be used to grade gliomas preoperatively, guide stereotactic biopsy, evaluate different tumor types, differentiate between recurrent tumor and delayed radiation necrosis, and monitor tumor response to therapy. Glioma grading.—As stated previously, in primary high-grade glioma vascular morphology is a critical parameter in helping determine malignant potential and survival. Glioma grading is important for determination of both prognosis and therapy. Authors of several recent studies (30,31,36) have found statistically significant correlations between tumor rCBV and glioma grade, as well as with conventional catheter-based angiographic tumor vascularities. In our experience with 51 preoperative cases of pathologically proved GBM, the maximum rCBV of the tumor varied from 1.95 to 28.59 (mean ⫾ SD, 5.5 ⫾ 4.5) (unpublished data, 2001). This wide variation is expected, since GBM is one of the most heterogeneous tumors in terms of histologic composition, with frequent areas of necrosis. Anaplastic astrocytoma (19 cases in our experience), which is differentiated from GBM by the lack of necrosis, demonstrated less variation in maximum rCBV measurements, with a range of 1.29 –10.26 (mean, 4.03 ⫾ 2.43). Lowgrade glioma (13 cases) has a maximum rCBV range of 0.89 –3.73 (mean, 1.86 ⫾ 18

䡠

Radiology

䡠

April 2002

0.77). Figures 7 and 8 demonstrate two cases with similar-appearing nonenhancing gliomas and very different CBV maps and histopathologic tumor grades. The tumor with high rCBV (Fig 7) is an anaplastic astrocytoma, whereas the tumor with low rCBV (Fig 8) is a low-grade fibrillary astrocytoma. It is important to recognize that there is an overlap of rCBV measurements between different glioma grades. Moreover, owing to the inherent extreme histologic heterogeneity of gliomas, rCBV measurements can and do vary considerably. Therefore, rCBV maps of gliomas should not be interpreted without concomitant evaluation of conventional MR images, which can provide other valuable information such as blood-brain barrier integrity or degree and characteristics of T2 abnormality. Stereotactic biopsy guidance.—Biopsy remains the definitive method to help determine tumor type and grade. The sampling error in biopsies of high-grade gliomas is well known and is partly attributable to extreme geographic heterogeneity within a single tumor (59,60). Ideally, the grading of gliomas should be based on histologic evaluation of specimens from the most malignant area of the tumor. Determination of this region can be difficult, however. Most biopsies are guided with contrast-enhanced T1weighted MR or CT images (61), which depict areas of blood-brain barrier break-

down that may not indicate the most malignant or vascular portion of the tumor. CBV maps can depict regions of increased vascularity that can serve as additional targets for stereotactic biopsy. At our institution, rCBV maps are routinely used to select biopsy sites for both enhancing and nonenhancing tumors and help reduce sampling error and the number of nondiagnostic biopsies. As shown in Figure 7, the rCBV map is particularly useful in nonenhancing tumors because it can be used to direct the biopsy to the “hot” area or presumed site of increased tumor vascularity on the map. On the basis of the contrast-enhanced T1-weighted image alone, it may be challenging to select a biopsy target. Assessment of response to therapy.—Recent developments in antiangiogenic drugs that target actively dividing tumor vasculature have added to the armamentarium of chemotherapeutic agents for brain tumors. Although at an early stage of development, antiangiogenic therapy shows promise as a means to destroy tumor vessels, the main route of nutrient and oxygen for tumor cells, without altering the preexisting normal vasculature. Imaging of gliomas after surgery, irradiation, and/or chemotherapy is often performed to assess tumor activity. However, conventional contrast-enhanced MR imaging cannot be used to reliably identify tumor progression, nor can it be used to monitor changes in tumor vascuCha et al

Figure 8. Nonenhancing right frontal lobe tumor in a 29-year-old man. (a) Transverse T2-weighted (3,400/119) MR image shows a mild degree of surrounding edema (arrows). (b) Transverse contrast-enhanced T1-weighted (600/14) MR image shows a nonenhancing right frontal lobe tumor (arrow). (c) rCBV map demonstrates no evidence of increased tumor vascularity. The patient underwent surgical resection, and the final pathologic results revealed low-grade fibrillary astrocytoma.

lature. In a study with patients receiving thalidomide (32), an antiangiogenic drug recently approved by the U.S. Food and Drug Administration, combined with carboplatin, we found that serial rCBV measurements during therapy predicted clinical response in the patients. Serial rCBV measurements correlated better with changes in patients’ clinical status than did conventional contrast-enhanced T1weighted MR images.

Differentiation of Therapy-induced Necrosis and Recurrent Tumor Differentiation between radiation necrosis and recurrent tumor carries obvious therapeutic implications. Recurrent tumors may benefit from repeat surgery with adjuvant chemotherapy or targeted high-dose radiation therapy, whereas radiation necrosis may be treated conservatively with steroids. Currently, biopsy or resection is the only definitive means of differentiating between recurrent tumor and radiation necrosis. However, surgical manipulation of radiation necrosis can cause further damage to adjacent brain parenchyma. Delayed radiation necrosis is usually indistinguishable from recurrent tumor on imaging or clinical grounds. Clinically, both entities can manifest with progressive focal neurologic deficits and signs of increased intracranial pressure. Volume 223

䡠

Number 1

At imaging, both can manifest as a mass lesion with surrounding edema (62– 66). Conventional contrast-enhanced CT or MR imaging cannot be used to reliably distinguish radiation necrosis from recurrent tumor. Both disease processes can involve varying degrees of bloodbrain barrier disruption that result in abnormal contrast enhancement. In addition, the diffuse neoplastic infiltration of recurrent tumor and the demyelinating process of radiation necrosis can both cause extensive white matter abnormalities, as well as extensive vasogenic edema. Pathologically, recurrent tumor and radiation necrosis are markedly dissimilar. Although the exact pathogenesis of delayed radiation necrosis remains obscure, a constant pathologic feature is extensive vascular injury and tissue hypoxia, in contrast to the features of recurrent tumor, which is characterized by neovascularization (67,68). Among imaging modalities, fluorodeoxyglucose positron emission tomography (PET) can be useful in differentiating the two entities (69,70). However, authors of a few recent studies (71,72) have suggested that PET may have low sensitivity and specificity. Moreover, PET scanners, unlike MR machines, are not yet widely available, thus limiting the use of PET in routine diagnosis and

evaluation of brain tumors. Differentiation between recurrent tumor and radiation necrosis by means of SPECT can also be difficult (73–75). In addition, the inherent low spatial resolution of both PET and SPECT is a particular issue when recurrent tumor and radiation necrosis coexist, which unfortunately is a common occurrence. MR imaging CBV maps can demonstrate the pathologic differences in vascularity between therapy-induced necrosis and recurrent tumor, and our preliminary results have shown them to be capable of facilitating differentiation between the two entities (76). Figures 9 and 10 demonstrate surgically proved recurrent GBM and radiation necrosis, respectively, where the conventional MR imaging features were nonspecific in both instances but perfusion MR imaging showed increased rCBV in recurrent tumor and low rCBV in radiation necrosis, thus allowing differentiation between the two.

Gliomatosis Cerebri Gliomatosis cerebri is a rare neoplasm of the brain characterized by moderately pleomorphic glial cells that infiltrate along the preexisting structures without causing demonstrable destruction (35). One of the hallmarks of gliomatosis is increased cellularity

Intracranial Mass Lesions: Echo-planar Perfusion MR Imaging

䡠

19

Figure 9. Recurrent tumor in a 72-year-old man with previously resected and irradiated right frontal lobe GBM. (a) Transverse T2-weighted (3,400/119) MR image displays a large area of hyperintensity abnormality involving the right frontal lobe. (b) Transverse contrast-enhanced T1-weighted (600/14) MR image shows nonenhancing right frontal lobe lesion (arrowheads) that extends into the genu of the corpus callosum (large arrow). Also note thick dural enhancement (small arrow) adjacent to the craniotomy site. (c) rCBV map demonstrates marked increase in vascularity in the right frontal lobe and genu of the corpus callosum, suggesting recurrent tumor. At repeat surgery, recurrence of GBM was confirmed.

Figure 10. Delayed radiation necrosis in an 82-year-old woman. (a) Transverse fluid-attenuated inversion recovery (9,000/110/2,500) MR image shows extensive hyperintensity abnormality that extends into the genu of the corpus callosum (large arrow). (b) Transverse contrast-enhanced T1-weighted (600/14) MR image shows heterogeneous, irregularly enhancing left frontal lobe mass (arrows). (c) rCBV map demonstrates no evidence of increased vascularity, suggesting radiation necrosis rather than recurrent tumor. The patient underwent surgical resection, and radiation necrosis without evidence of tumor was found.

without frank destruction of the infiltrated parenchyma or presence of neovascularity. Invasion of both gray and white matter is seen, but mitosis, necrosis, and angiogenesis are distinctly uncommon (77–79). 20

䡠

Radiology

䡠

April 2002

On MR images, gliomatosis manifests as a focal or diffuse infiltrative signal intensity abnormality involving the white matter and expansion of the cortex, usually without contrast enhancement. There is minimal, if any, mass effect or

extensive vasogenic edema (79 – 83). In our series of seven patients with gliomatosis cerebri, the perfusion MR imaging appearance reflected the lack of angiogenesis of this tumor found at histopatholgic examination, and rCBV valCha et al

Figure 11. Photomicrographs show angiocentric growth pattern of PCL. Top left: Low-power view shows tumor cells organized around a cerebral vessel (long arrows), as well as in the neuropil (short arrows). (Hematoxylin-eosin stain; original magnification, ⫻50.) Bottom left: Mediumpower magnification further emphasizes the perivascular growth of tumor cells that results in a narrowing of the vascular lumina (arrows). (Hematoxylin-eosin stain; original magnification, ⫻100.) Top right: High-power view depicts cross sections of vascular channels (arrows) cuffed by lymphoma cells. (Hematoxylin-eosin stain; original magnification, ⫻200.) Bottom right: Reticulin staining highlights angiocentric growth of the lymphoma and demonstrates reduplication of the vascular matrix (arrows). (Original magnification, ⫻400.)

ues tended to be lower than those of lowgrade gliomas and even lower than those of normal-appearing unaffected white matter.

Metastatic Neoplasms Intracranial metastases compose almost 50% of all supratentorial brain tumors (35). Metastatic tumors spread into the central nervous system via hematogenous routes and induce neovascularization as they grow and expand. The newly formed capillaries resemble those of the primary systemic tumor, with gap junctions, fenestrated membranes, and open endothelial junctions, all of which differ radically from normal brain capillaries containing a well-developed blood-brain barrier with tight junctions, a continuous basement membrane, and astrocytic foot processes (84). Intracranial metastatic tumors, especially those that cause neurologic symptoms, are associated with a variable degree of vasogenic edema. The capillaries in the area of vasogenic edema are consistently normal in appearance, and no tumor cells are found beyond the macroscopic boundaries of the tumor (85). The imaging diagnosis of intracraVolume 223

䡠

Number 1

nial metastases is usually straightforward and uncomplicated. Metastatic tumors tend to be multiple, well-circumscribed, and favor the gray matter–white matter junction and the patients tend to have a known history of systemic malignancy. On the other hand, solitary metastases, which can occur in 30%–50% of cases, can pose a diagnostic challenge with regard to differentiation from a primary glioma, particularly when there is no history of systemic cancer or the medical work-up fails to identify a primary systemic malignancy. Conventional MR imaging characteristics of the solitary metastasis and primary glioma are nonspecific and cannot be used with confidence to differentiate between the two. Both tumors tend to enhance and may be associated with a variable degree of peritumoral edema, which is defined as the area of hyperintensity on T2-weighted images in immediate contact with the enhancing tumor margin. In our study of 14 solitary metastases and 25 primary gliomas, perfusion MR imaging could not be used to differentiate the two reliably since both are highly vas-

cular tumors demonstrating increased rCBV (P ⬎ .05) (unpublished data, 2002). There was, however, a statistically significant difference (P ⬍ .001) in rCBV within the peritumoral region, where the primary glioma values were higher than the solitary metastasis values. Our experience suggests a promising role for perfusion MR imaging in differentiating solitary metastasis from primary glioma on the basis of differences in peritumoral rCBV measurements (86). This difference in peritumoral rCBV can, in part, be explained by differences in pathophysiology: In metastatic tumors, peritumoral edema represents pure vasogenic edema caused by increased interstitial water due to leaky capillaries (85). In other words, in metastatic tumors, there is no histologic evidence of tumor beyond the outer enhancing margin of the tumor, and the peritumoral region represents reaction of the surrounding intrinsically normal but edematous brain parenchyma. In high-grade gliomas, on the other hand, the peritumoral region represents a variable combination of vasogenic edema and tumor cells infiltrating along the perivascular spaces (87). It has been shown that neoplastic cells can be found in some high-grade gliomas not only outside the enhancing margin but also well beyond the outer edge of the peritumoral zone, as seen on T2-weighted MR images (61). Differentiation between low-grade glioma and metastatic neoplasms can be made on the basis of the usual lack of marked contrast enhancement of the former. Also, peritumoral edema, usually quite marked in solitary metastases, tends to be minimal or absent in low-grade gliomas.

Primary Cerebral Lymphoma Over the past two decades, the incidence of primary cerebral lymphoma (PCL) has substantially increased in both immunocompromised and immunecompetent individuals—a phenomenon that cannot be entirely explained by changes in tumor classification, the increased prevalence of acquired immunodeficiency syndrome, or the growing number of organ transplantations. The incidence ratios of PCL and GBM have risen from 1:250 in 1974 to 1:6 for the years 1981–1990, with PCL now accounting for 6.6%–15.4% of all primary brain tumors (88,89). Unlike other high-grade intracranial neoplasms, PCL is treated with combined high-dose chemotherapy and radiation therapy without surgery. Surgical intervention is usually limited to

Intracranial Mass Lesions: Echo-planar Perfusion MR Imaging

䡠

21

biopsy to obtain tissue for pathologic diagnosis (90 –92). It has been shown that surgical resection of PCL does not necessarily alter prognosis and can lead to profound functional deficits and increased postoperative morbidity. More important, when patients with PCL undergo steroid therapy to decrease intracranial pressure before biopsy, the pathologic findings can be difficult to interpret and a definitive diagnosis cannot be established. In addition, owing to the predilection of PCL to occur near eloquent areas of the brain (eg, deep gray matter, corpus callosum, and subependymal locations), gross total resection may in fact be contraindicated and the performance of biopsy may be challenging. A minimally invasive and accurate diagnosis of PCL is, therefore, an important goal that could clearly alter patient treatment and reduce the risk to the patient. Radiologic diagnosis of PCL remains a challenge, however. Conventional MR imaging findings of PCL can be similar to those of other intracranial tumors or even to those of demyelinating lesions. PCL lesions usually enhance; can be multiple; can favor, as mentioned previously, a deep gray matter or subependymal location; frequently involve the corpus callosum; and can appear similar to GBM. The ability of perfusion MR imaging to help detect and quantify tumor angiogenesis can potentially be useful in differentiating high-grade glioma from PCL. One of the most striking histopathologic features of PCL, shown in Figure 11, is the angiocentric growth pattern—tumor cells forming multiple, thick layers around the host vessels—and widening of the perivascular space. Neovascularization is not a prominent feature, although vascular abnormalities such as tumor invasion of endothelial cells and even into the vessel lumen can often be seen. There are a few published reports on the perfusion MR imaging characteristics of PCL. In a preliminary study, Sugahara et al (33) demonstrated that rCBV values of cerebral lymphomas (five PCLs and three secondary lymphomas) were low. On the other hand, Ernst et al (93) showed that the rCBV of PCL was significantly higher than that of toxoplasmosis in immunocompromised patients, thus allowing differentiation between those two entities. In our series of 19 consecutive patients with PCL, the maximum rCBV ranged from 0.42 to 3.41 (mean, 1.44 ⫾ 0.67) (unpublished data, 2001). When compared with the mean rCBV of 5.5 ⫾ 4.5 in 51 patients with GBM, the difference in mean rCBV was statistically 22

䡠

Radiology

䡠

April 2002

Figure 12. PCL in a 72-year-old man. (a) Transverse contrast-enhanced T1-weighted (600/14) MR image shows homogeneous enhancement and expansion of the splenium of the corpus callosum (arrows). (b) rCBV map demonstrates no evidence of a marked increase in vascularity (rCBV range, 0.89 –1.10).

significant (P ⬍ .001) (94). An example of PCL with marked enhancement following contrast agent administration and low rCBV is shown in Figure 12. At times, it is more difficult to differentiate PCL from dominant masslike demyelinating plaques, also called TDLs, but in our experience PCLs tend to have higher rCBV than TDLs (more discussion on this topic is presented later).

Extraaxial Neoplasms Intracranial extraaxial neoplasms are tumors arising from tissue other than brain parenchyma, such as the meninges, dura, calvarium, ventricle, choroid plexus, pineal gland, or pituitary gland, with meningioma being the most common type of neoplasm. Extraaxial neoplasms tend to be highly vascular and, unlike intraaxial tumors, completely lack a blood-brain barrier and, therefore, demonstrate intense enhancement after administration of a contrast agent. Differentiation between intra- and extraaxial neoplasms is usually straightforward and does not pose a serious diagnostic challenge. There are characteristic conventional MR imaging features that can be used to identify these tumors, such as buckling of adjacent cortex, widening of cerebrospinal fluid spaces, and displacement of subarachnoid veins. There are instances, however, when this differentiation can be difficult, especially when there is brain invasion by extraaxial le-

sions or dural involvement by intraaxial tumors. To the best of our knowledge, there are no published results on the differentiation between extra- and intraaxial tumors by using perfusion MR imaging. In our preliminary results comparing 15 extraaxial tumors (12 meningiomas, three choroid plexus tumors) with 15 similarly appearing high-grade intraaxial tumors (10 GBMs, five metastases), the rCBVs of extraaxial tumors were significantly higher than those of intraaxial tumors (unpublished data, 2000). An example of an rCBV map of an atypical meningioma is shown in Figure 13. It is important to emphasize, however, that in extraaxial tumors there is immediate and persistent leakage of contrast agent during the bolus phase of dynamic imaging, due to complete absence of the blood-brain barrier. Owing to severe leakage of contrast agent, the rCBV calculation can give erroneously low or high values, depending on the imaging and data processing methods used. Therefore, the rCBV is not a reliable parameter for differentiating intra- and extraaxial neoplasms. In our study, we found it helpful to evaluate the signal intensity–time curve during the first pass to differentiate between the two tumor types (unpublished data, 2000). In extraaxial tumors, the curve does not return to baseline, whereas in high-grade intraaxial tumors there is an initial and partial reCha et al

Figure 13. Atypical meningioma in an 81-year-old woman. (a) Transverse contrast-enhanced T1-weighted (600/14) MR image shows intense enhancement of a right frontal lobe tumor (arrow). (b) rCBV map shows increased vascularity (red) of the tumor.

Figure 14. Signal intensity–time curve of atypical meningioma (same patient as in Fig 13) displays immediate leakage of contrast agent (arrow) due to absence of a blood-brain barrier, which is characteristic of an extraaxial tumor. An example of a partially disrupted but not completely absent blood-brain barrier, as in the case of GBM, is shown in Figure 5.

turn of the curve that levels off. Figure 14 demonstrates a characteristic signal intensity–time curve of an extraaxial neoplasm (meningioma shown in Fig 13), which can be compared with that of the intraaxial tumor (GBM) shown in Figure 5.

Nonneoplastic Lesions There are a few nonneoplastic lesions of the brain— cerebral infections, TDLs, and, less commonly, infarcts—that may be confused with and misdiagnosed as brain tumor. The conventional MR imaging appearance of these lesions can be nonspecific and poses a serious challenge in differentiation from brain tumors. Volume 223

䡠

Number 1

Even with the use of a contrast agent, the distinction can still be difficult because any process that disrupts the integrity of the blood-brain barrier can result in enhancement. Perfusion MR imaging potentially offers a different mechanism for differentiation by allowing assessment of variations in the vascularity of these lesions. Central nervous system infections.—With the advent of antibiotics, the incidence of intracranial bacterial abscesses has decreased substantially. With the growing numbers of patients with acquired immunodeficiency syndrome and with organ transplants, however, the prevalence of opportunistic infection is on the rise,

especially infections caused by viral or protozoan organisms. Radiologic diagnosis of intracranial infections can be challenging because of the variable appearance of lesions secondary to different offending microbes and different stages of manifestation. On MR images, an intracranial infection can manifest as a focal mass or a diffuse process, depending on the offending organism and the stage of infection. Bacterial infections can manifest as a nonspecific diffuse signal intensity abnormality with or without enhancement in the cerebritis stage, which precedes the formation of a well-defined abscess. A cerebral abscess, usually caused by a pyogenic bacterial organism, often manifests as a ring-enhancing mass with marked surrounding edema. Encephalitides, usually caused by a virus or a protozoan organism, can appear similar to diffuse cerebritis. The perfusion MR imaging appearance of various intracranial infections is, not surprisingly, variable, depending again on the type of infection. Our limited experience with perfusion MR imaging of various pathologically proved intracranial infections reveals variable rCBV measurements. We studied two cases each of herpes encephalitis and toxoplasma encephalitis and six cases of bacterial abscess (unpublished data, 2001). Herpes and toxoplasmosis demonstrated homogeneously low rCBVs, whereas abscesses showed increased rCBVs. In fact, all bacterial abscesses demonstrated areas of increased rCBV, with maximum rCBV consistently found in the region just superficial and adjacent to the enhancing capsular rim, as shown in Figure 15. It should be pointed out that even without rCBV measurements, the characteristic hypointense rim of the abscess capsule seen on T2-weighted MR images, also shown in Figure 15, can be used prospectively to differentiate between abscess and tumor. In herpes encephalitis and toxoplasmosis (Fig 16), there is no increase in rCBV, with values lower than those of normal white matter. Tumefactive demyelinating lesions.—TDLs can mimic intracranial neoplasms and pose a diagnostic dilemma both at clinical presentation and at conventional MR imaging. At imaging, TDLs and high-grade intracranial neoplasms can both demonstrate contrast enhancement, perilesional edema, varying degrees of mass effect, and central necrosis (95–97). Furthermore, TDLs can be confused with high-grade glial neoplasms at histopathologic evaluation owing to the

Intracranial Mass Lesions: Echo-planar Perfusion MR Imaging

䡠

23

Figure 15. Pyogenic abscess in a 45-year-old man. (a) Transverse T2-weighted (3,400/119) MR image shows a heterogeneous mass (arrowheads) with moderate edema (arrows) in the right parietal lobe. (b) Transverse contrast-enhanced T1-weighted (600/14) MR image shows irregular rim enhancement (arrows). (c) rCBV map demonstrates zone of increased vascularity (arrows) is just outside the enhancing rim of the abscess.

Figure 16. Toxoplasma encephalitis in a 47-year-old man. (a) Transverse T2-weighted (3,400/119) MR image shows heterogeneous hyperintense signal intensity abnormality in both frontal lobes that extends into the genu of the corpus callosum. (b) Transverse contrast-enhanced T1-weighted (600/14) MR image shows irregularly enhancing lesions (arrows) in the frontal lobes. (c) rCBV map demonstrates no evidence of increased vascularity, which is suggestive of a nonneoplastic lesion. The patient underwent stereotactic biopsy, the results of which showed toxoplasma encephalitis.

presence of hypercellularity and atypical reactive astrocytes with mitotic figures (95). A single dominant TDL, in particular, is often misdiagnosed as a brain tumor, leading to unnecessary and possibly harmful biopsy or even resection. There are a few reported cases where patients 24

䡠

Radiology

䡠

April 2002

with TDL underwent not only surgery but also radiation therapy because of a mistaken diagnosis of lymphoma or glioma (95,96). In light of this diagnostic dilemma, there are several reports in the pathology literature (95,98 –100) that have stressed the need for special stains

for myelin and axons to help establish the correct diagnosis of TDL. Numerous reports in the radiology literature (96, 97,100,101) also have addressed TDL masquerading as brain tumor and leading to unnecessary surgery. One of the key histopathologic differCha et al

Figure 17. TDL in a 46-year-old woman. (a) Left: Transverse T2-weighted (3,400/119) MR image shows multiple periventricular high-signalintensity abnormalities, with a dominant masslike lesion (large arrow) in the right posterior frontal lobe region. Middle: Transverse fluid-attenuated inversion recovery (9,000/110/2,500) MR image shows the smaller lesions (arrows) more distinctly. Right: Transverse contrast-enhanced T1weighted (600/14) MR image shows thick rim enhancement (arrowheads) of the dominant lesion. (b) Series of transverse gradient-echo echo-planar (1,000/54) MR images obtained before (left), at peak arrival of (middle), and after (right) bolus injection of contrast agent display a prominent vessel-like structure (arrow) within the lesion.

ences between TDL and high-grade brain tumors is the absence of frank angiogenesis in the former. The blood vessels in areas of demyelination are intrinsically normal without evidence of neovascularization, in contrast to the angiogenesis seen in tumors. In our study (102) of 12 TDLs and 11 brain tumors that had similar conventional MR imaging features, we found higher rCBV values in the tumors than in the TDLs. The rCBV values of TDLs ranged from 0.22 to 1.79 (mean, 0.88 ⫾ 0.46). The rCBV values of intracranial neoplasms ranged from 1.55 to Volume 223

䡠

Number 1

19.20 (mean, 6.47 ⫾ 6.52). This difference was statistically significant (P ⫽ .009). The difference in rCBVs between TDLs and four cases of primary cerebral lymphoma was less pronounced but still statistically significant (P ⫽ .005). In that study (102), we also observed prominent vessel-like structures, which were not seen on conventional MR images (Fig 17a), running through the center of the TDLs at the moment of peak contrast enhancement (Fig 17b). One of the hallmarks of demyelinating plaques is their propensity to occur close to

periventricular tributary or radial veins. We therefore postulate that these prominent structures are periventricular veins or venules. A dynamic series of perfusion MR images depicts these venous structures during passage of gadopentetate dimeglumine through the venous system, thus providing another potential mechanism for differentiation between TDLs and high-grade brain tumors. Cerebrovascular disease.—Diagnosis of acute or hyperacute stroke has been revolutionized by the widespread application of diffusion-weighted MR imaging,

Intracranial Mass Lesions: Echo-planar Perfusion MR Imaging

䡠

25

Figure 18. Presumed infarct in a 35-year-old woman. (a) Transverse contrast-enhanced T1-weighted (600/14) MR image shows a heterogeneously hypointense lesion (solid arrows) with irregular linear enhancement in the right frontoparietal region. Note fiducial marker (open arrow) for surgical planning. (b) Transverse fluid-attenuated inversion recovery (9,000/110/2,500) MR image shows heterogeneous high-signal-intensity abnormality. (c) rCBV map shows an area of susceptibility (arrow) and no increase in vascularity, thus suggesting a nonneoplastic lesion. Surgery was canceled, and at 6-week follow-up MR examination, the lesion had almost completely disappeared without any intervening therapy.

which is sensitive to alterations in molecular motion during the onset and progression of early cerebral ischemia. Diffusion-weighted imaging clearly demonstrates the presence and distribution of acute infarct and, more important, allows exclusion of infarct in patients presenting with symptoms of cerebral ischemia due to another disease entity. Measurements of blood flow obtained with perfusion MR imaging have been used mainly as an adjunctive tool to diffusion-weighted imaging to provide information on altered regional tissue perfusion. Blood flow measurements require acquisition of the arterial input function (ie, concentration of contrast agent in arterial blood) and a different analysis of the dynamic data (103,104). Perfusion MR imaging in acute stroke can potentially be used to identify the “ischemic penumbra,” ischemic yet viable and not fully infarcted areas of the brain surrounding an infarct core. It is beyond the scope of this article to discuss the current literature on perfusion MR imaging of acute ischemic stroke in detail, and only some key points can be outlined here. Several excellent recent articles on the role of perfusion imaging in acute stroke are available (105–107). Diffusion-weighted MR imaging demonstrates areas of restricted water move26

䡠

Radiology

䡠

April 2002

ment that occur during the early onset of ischemia. In the majority of cases, such an area resembles the minimal size of the extent of the final infarct volume (108). If the area of diffusion abnormality is compared with the area of impaired or altered perfusion, different patterns of “mismatch” can be recognized. It is assumed that in cases where the area of the diffusion abnormality is larger than that of reduced perfusion, reperfusion has taken place but permanent damage, as represented by the area of abnormality, has already occurred. In cases where the area of the diffusion abnormality is smaller than the area of reduced perfusion, it is believed that the area with normal diffusion but decreased perfusion—the ischemic penumbra—may represent potentially salvageable tissue or tissue at risk. This distinction between still-recoverable and dead brain tissue, based on diffusionweighted and perfusion MR imaging findings, might be valuable in the selection of patients with acute stroke who may benefit from thrombolytic therapy. Perfusion MR imaging in conjunction with diffusion-weighted imaging may contribute to improved patient care and prognosis by playing an important role in the initial screening of patients with hyperacute or acute stroke and in the selection and evaluation of patients for

immediate intervention with thrombolytic therapy. As pointed out by Powers and Zivin (109), however, a randomized trial is needed to demonstrate that patients in whom the choice of therapy is based on the results of perfusion and diffusion-weighted MR imaging have a better outcome than does a similar group whose therapy is determined without such input. The aforementioned application of perfusion MR imaging in acute infarct mainly involves analysis of information on blood flow and mean transit time. The blood volume information (ie, rCBV) obtained from perfusion MR imaging can also be a valuable diagnostic tool to help differentiate infarct from brain tumor in some cases. Although not common, acute or subacute infarcts— especially those that are atypical in location or in clinical manifestation or venous infarcts that do not follow an arterial vascular territory— can be confused and misdiagnosed as neoplasm, and the patient may undergo unnecessary and even harmful surgical intervention. In these cases, MR imaging CBV maps can be helpful because any infarct, arterial or venous, demonstrates low rCBV due to lack of or decreased perfusion. An example of such a case is demonstrated in Figure 18. Cha et al

CONCLUSION Perfusion MR imaging is a useful complement to traditional anatomic imaging and is becoming increasingly used in clinical practice to help diagnose, manage, and understand intracranial mass lesions. MR imaging CBV maps provide quantitative estimates of regional blood volume that can be used to help (a) differentiate tumor from nonneoplastic lesions, (b) evaluate different tumor types, (c) predict tumor grade, (d) serially and noninvasively monitor tumor progression, (e) monitor the efficacy of antiangiogenic cancer therapy, and (f ) delineate areas of altered perfusion in acute infarct. There are, however, several important limitations of the technique, such as susceptibility artifacts, relative (not absolute) quantification of CBV, and inaccurate estimation of CBV in situations of severe disruption or absence of the blood-brain barrier. With recognition of its strengths and potential pitfalls, perfusion MR imaging can be used as part of the routine evaluation of intracranial mass lesions, to help improve the diagnostic accuracy, understand the pathophysiology, detect and quantify angiogenesis, and serve as an arbiter to assess novel therapies that target blood vessels. References 1. Peters AM. Fundamentals of tracer kinetics for radiologists. Br J Radiol 1998; 71: 1116 –1129. 2. Guyton AC. Cerebral blood flow, the cerebrospinal fluid, and brain metabolism. In: Basic neuroscience: anatomy and physiology. 2nd ed. Philadelphia, Pa: Saunders, 1991; 285–287. 3. Weisskoff R, Belliveau J, Kwong K, Rosen B. Functional MR imaging of capillary hemodynamics. In: Potchen E, ed. Magnetic resonance angiography: concepts and applications. St Louis, Mo: Mosby, 1993; 473– 484. 4. Axel L. Cerebral blood flow determination by rapid-sequence computed tomography: theoretical analysis. Radiology 1980; 137:679 – 686. 5. Zierler KL. Circulation times and the theory of indicator-dilution methods for determining blood flow and volume. In: Handbook of physiology. Baltimore, Md: Williams & Wilkins, 1962; 585– 615. 6. Zhang W, Williams DS, Koretsky AP. Measurement of rat brain perfusion by NMR using spin labeling of arterial water: in vivo determination of the degree of spin labeling. Magn Reson Med 1993; 29:416 – 421. 7. 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; 192:513–520. Volume 223

䡠

Number 1

8.

9. 10.

11.

12.

13.

14.

15.

16.

17.

18. 19.

20.

21.

22.

Kim SG. Quantification of relative cerebral blood flow change by flow-sensitive alternating inversion recovery (FAIR) technique: application to functional mapping. Magn Reson Med 1995; 34: 293–301. Detre JA, Leigh JS, Williams DS, Koretsky AP. Perfusion imaging. Magn Reson Med 1992; 23:37– 45. Tanabe JL, Yongbi M, Branch C, Hrabe J, Johnson G, Helpern JA. MR perfusion imaging in human brain using the UNFAIR technique: un-inverted flow-sensitive alternating inversion recovery. J Magn Reson Imaging 1999; 9:761–767. Kim SG, Ackerman JJ. Multicompartment analysis of blood flow and tissue perfusion employing D2O as a freely diffusible tracer: a novel deuterium NMR technique demonstrated via application with murine RIF-1 tumors. Magn Reson Med 1988; 8:410 – 426. Pekary AE, Li HJ, Chan SI, Hsu CJ, Wagner TE. Nuclear magnetic resonance studies of histone IV solution conformation. Biochemistry 1975; 14:1177–1184. McLaughlin AC, Ye FQ, Pekar JJ, Santha AK, Frank JA. Effect of magnetization transfer on the measurement of cerebral blood flow using steady-state arterial spin tagging approaches: a theoretical investigation. Magn Reson Med 1997; 37:501–510. Ewing JR, Branch CA, Fagan SC, et al. Fluorocarbon-23 measure of cat cerebral blood flow by nuclear magnetic resonance. Stroke 1990; 21:100 –106. Villringer A, Rosen BR, Belliveau JW, et al. Dynamic imaging with lanthanide chelates in normal brain: contrast due to magnetic susceptibility effects. Magn Reson Med 1988; 6:164 –174. Rosen BR, Belliveau JW, Vevea JM, Brady TJ. Perfusion imaging with NMR contrast agents. Magn Reson Med 1990; 14: 249 –265. Tofts PS, Kermode AG. Measurement of the blood-brain barrier permeability and leakage space using dynamic MR imaging. I. Fundamental concepts. Magn Reson Med 1991; 17:357–367. Tofts PS. Modeling tracer kinetics in dynamic Gd-DTPA MR imaging. J Magn Reson Imaging 1997; 7:91–101. Roberts HC, Roberts TP, Brasch RC, Dillon WP. Quantitative measurement of microvascular permeability in human brain tumors achieved using dynamic contrast-enhanced MR imaging: correlation with histologic grade. AJNR Am J Neuroradiol 2000; 21:891– 899. Zheng J, Ehrhardt JC, Cizadlo T, Yuh WT. Comparison of inversion recovery asymmetrical spin-echo EPI and gradient-echo EPI for brain motor activation study. J Magn Reson Imaging 1997; 7:843– 847. Stables LA, Kennan RP, Gore JC. Asymmetric spin-echo imaging of magnetically inhomogeneous systems: theory, experiment, and numerical studies. Magn Reson Med 1998; 40:432– 442. Young IR, Cox IJ, Coutts GA, Bydder GM. Some considerations concerning susceptibility, longitudinal relaxation time constants and motion artifacts in vivo human spectroscopy. NMR Biomed 1989; 2:329 –339.

23.

24.

25. 26.

27. 28. 29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

Weisskoff RM, Zuo CS, Boxerman JL, Rosen BR. Microscopic susceptibility variation and transverse relaxation: theory and experiment. Magn Reson Med 1994; 31:601– 610. Cha S, Lu S, Johnson G, Knopp EA. Dynamic susceptibility contrast MR imaging: correlation of signal intensity changes with cerebral blood volume measurements. J Magn Reson Imaging 2000; 11:114 –119. Mansfield P. Multi-planar image formation using NMR spin echoes. J Phys Chen 1977; 10:L55–L58. Johnson G, Hutchison JMS. The limitations of NMR recalled-echo imaging techniques. J Magn Reson 1985; 63:14 – 30. DeLaPaz RL. Echo-planar imaging. RadioGraphics 1994; 14:1045–1058. Edelman RR, Wielopolski P, Schmitt F. Echo-planar MR imaging. Radiology 1994; 192:600 – 612. Bruening R, Wu RH, Yousry TA, et al. Regional relative blood volume MR maps of meningiomas before and after partial embolization. J Comput Assist Tomogr 1998; 22:104 –110. 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. Knopp EA, Cha S, Johnson G, et al. Glial neoplasms: dynamic contrast-enhanced T2*-weighted MR imaging. Radiology 1999; 211:791–798. Cha S, Knopp EA, Johnson G, et al. Dynamic contrast-enhanced T2-weighted MR imaging of recurrent malignant gliomas treated with thalidomide and carboplatin. AJNR Am J Neuroradiol 2000; 21:881– 890. Sugahara T, Korogi Y, Shigematsu Y, et al. Perfusion-sensitive MRI of cerebral lymphomas: a preliminary report. J Comput Assist Tomogr 1999; 23:232– 237. Sugahara T, Korogi Y, Shigematsu Y, et al. Value of dynamic susceptibility contrast magnetic resonance imaging in the evaluation of intracranial tumors. Top Magn Reson Imaging 1999; 10:114 –124. Burger PC, Vogel FS. The brain: tumors. In: Burger PC, Vogel FS, eds. Surgical pathology of the central nervous system and its coverings. New York, NY: Wiley, 1982; 223–266. Sugahara T, Korogi Y, Kochi M, et al. Correlation of MR imaging-determined cerebral blood volume maps with histologic and angiographic determination of vascularity of gliomas. AJR Am J Roentgenol 1998; 171:1479 –1486. Folkman J. Seminars in medicine of the Beth Israel Hospital, Boston: clinical applications of research on angiogenesis. N Engl J Med 1995; 333:1757–1763. O’Reilly MS, Holmgren L, Shing Y, et al. Angiostatin: a circulating endothelial cell inhibitor that suppresses angiogenesis and tumor growth. Cold Spring Harb Symp Quant Biol 1994; 59:471– 482. O’Reilly MS, Boehm T, Shing Y, et al. Endostatin: an endogenous inhibitor of angiogenesis and tumor growth. Cell 1997; 88:277–285.

Intracranial Mass Lesions: Echo-planar Perfusion MR Imaging

䡠

27

40.

41.

42.

43.

44.

45.

46.

47.

48. 49.

50.

51.

52.

53.

54.

55.

56.

28

䡠

Kirsch M, Strasser J, Allende R, Bello L, Zhang J, Black PM. Angiostatin suppresses malignant glioma growth in vivo. Cancer Res 1998; 58:4654 – 4659. Wu Z, O’Reilly MS, Folkman J, Shing Y. Suppression of tumor growth with recombinant murine angiostatin. Biochem Biophys Res Commun 1997; 236: 651– 654. Miller BL. A review of chemical issues in 1H NMR spectroscopy: N-acetyl-L-aspartate, creatine and choline. NMR Biomed 1991; 4:47–52. Poptani H, Gupta RK, Roy R, Pandey R, Jain VK, Chhabra DK. Characterization of intracranial mass lesions with in vivo proton MR spectroscopy. AJNR Am J Neuroradiol 1995; 16:1593–1603. Kugel H, Heindel W, Ernestus RI, Bunke J, du Mesnil R, Friedmann G. Human brain tumors: spectral patterns detected with localized H-1 MR spectroscopy. Radiology 1992; 183:701–709. Bruhn H, Frahm J, Gyngell ML, et al. Noninvasive differentiation of tumors with use of localized H-1 MR spectroscopy in vivo: initial experience in patients with cerebral tumors. Radiology 1989; 172:541–548. Preul MC, Caramanos Z, Collins DL, et al. Accurate, noninvasive diagnosis of human brain tumors by using proton magnetic resonance spectroscopy. Nat Med 1996; 2:323–325. Shimizu H, Kumabe T, Shirane R, Yoshimoto T. Correlation between choline level measured by proton MR spectroscopy and Ki-67 labeling index in gliomas. AJNR Am J Neuroradiol 2000; 21: 659 – 665. Castillo M, Kwock L. Proton MR spectroscopy of common brain tumors. Neuroimaging Clin N Am 1998; 8:733–752. Castillo M, Kwock L, Mukherji SK. Clinical applications of proton MR spectroscopy. AJNR Am J Neuroradiol 1996; 17: 1–15. Ott D, Hennig J, Ernst T. Human brain tumors: assessment with in vivo proton MR spectroscopy. Radiology 1993; 186: 745–752. Falini A, Giovanna C, Origgi D, et al. Proton magnetic spectroscopy and intracranial tumors: clinical perspectives. J Neurology 1996; 243:706 –714. Dowling C, Bollen AW, Noworolski SM, et al. Preoperative proton MR spectroscopic imaging of brain tumors: correlation with histopathologic analysis of resection specimens. AJNR Am J Neuroradiol 2001; 22:604 – 612. Burger PC, Vogel FS, Green SB, Strike TA. Glioblastoma multiforme and anaplastic astrocytoma: pathologic criteria and prognostic implications. Cancer 1985; 56:1106 –1111. Burger P. Malignant astrocytic neoplasms: classification, pathology, anatomy, and response to therapy. Semin Oncol 1986; 13:16 –20. Wesseling P, Ruiter DJ, Burger PC. Angiogenesis in brain tumors: pathobiological and clinical aspects. J Neurooncol 1997; 32:253–265. Plate KH, Mennel HD. Vascular morphology and angiogenesis in glial tumors. Exp Toxicol Pathol 1995; 47:89 – 94. Radiology

䡠

April 2002

57.

58.

59.

60.

61.

62. 63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

Runge VM, Clanton JA, Price AC, et al. The use of Gd DTPA as a perfusion agent and marker of blood-brain barrier disruption. Magn Reson Imaging 1985; 3:43–355. Roman-Goldstein SM, Barnett PA, McCormick CI, et al. Effects of Gd-DTPA after osmotic BBB disruption in a rodent model: toxicity and MR findings. J Comput Assist Tomogr 1994; 18:731–736. Willems JG, Alva-Willems JM. Accuracy of cytologic diagnosis of central nervous system neoplasms in stereotactic biopsies. Acta Cytol 1984; 28:243–249. Fratkin JD, Ward MM, Roberts DW, Sullivan MM. CT-guided stereotactic biopsy of intracranial lesions: correlation between core biopsy and aspiration smear. Diagn Cytopathol 1986; 2:126 –132. Kelly PJ, Daumas-Duport C, Scheithauer BE, Kall BA, Kispert DB. Stereotactic histologic correlations of computed tomography and magnetic resonance imaging defined abnormalities in patients with glial neoplasms. Mayo Clin Proc 1987; 62:450 – 459. Valk PE, Dillon WP. Radiation injury of the brain. AJNR Am J Neuroradiol 1991; 12:45– 62. Dooms GC, Hecht S, Brant-Zawadzki M, Berthiaume Y, Norman D, Newton TH. Brain radiation lesions: MR imaging. Radiology 1986; 158:149 –155. Babu R, Huang PP, Epstein F, Budzilovich GN. Late radiation necrosis of the brain: case report. J Neurooncol 1993; 17:37– 42. Chan YL, Leung SF, King AD, Choi PH, Metreweli C. Late radiation injury to the temporal lobes: morphologic evaluation at MR imaging. Radiology 1999; 213:800 – 807. Chong VF, Fan YF, Chan LL. Temporal lobe necrosis in nasopharyngeal carcinoma: pictorial essay. Australas Radiol 1997; 41:392–397. Burger PC, Mahley MS Jr, Dudka L, Vogel FS. The morphologic effects of radiation administered therapeutically for intracranial gliomas: a postmortem study of 25 cases. Cancer 1979; 44: 1256 –1272. Morris JG, Grattan-Smith P, Panegyres PK, O’Neill P, Soo YS, Langlands AO. Delayed cerebral radiation necrosis. Q J Med 1994; 87:119 –129. Di Chiro G, Oldfield E, Wright DC, et al. Cerebral necrosis after radiotherapy and/or intraarterial chemotherapy for brain tumors: PET and neuropathologic studies. AJR Am J Roentgenol 1988; 150: 189 –197. Kim EE, Chung SK, Haynie TP, et al. Differentiation of residual or recurrent tumors from post-treatment changes with F-18 FDG PET. RadioGraphics 1992; 12:269 –279. Olivero WC, Dulebohn SC, Lister JR. The use of PET in evaluating patients with primary brain tumours: is it useful? J Neurol Neurosurg Psychiatry 1995; 58: 250 –252. Ricci PE, Karis JP, Heiserman JE, Fram EK, Bice AN, Drayer BP. Differentiating recurrent tumor from radiation necrosis: time for re-evaluation of positron emission tomography? AJNR Am J Neuroradiol 1998; 19:407– 413.

73.

74.

75.

76.

77.

78. 79.

80.

81.

82.

83.

84. 85.

86.

87.

de Vries B, Taphoorn MJ, van Isselt JW, Terhaard CH, Jansen GH, Elsenburg PH. Bilateral temporal lobe necrosis after radiotherapy: confounding SPECT results. Neurology 1998; 51:1183–1184. Thompson TP, Lunsford LD, Kondziolka D. Distinguishing recurrent tumor and radiation necrosis with positron emission tomography versus stereotactic biopsy. Stereotact Funct Neurosurg 1999; 73:9 –14. Yoshii Y, Moritake T, Suzuki K, Fujita K, Nose T, Satou M. Cerebral radiation necrosis with accumulation of thallium 201 on single-photon emission CT. AJNR Am J Neuroradiol 1996; 17:1773– 1776. Cha S, Johnson G, Yuz M, et al. The role of contrast-enhanced perfusion MR imaging in differentiating between recurrent tumor and radiation necrosis (abstr). Radiology 1999; 213(P):188. Artigas J, Cervos-Navarro J, Iglesias JR, Ebhardt G. Gliomatosis cerebri: clinical and histological findings. Clin Neuropathol 1985; 4:135–148. del Carpio–O’Donovan R, Korah I, Salazar A, Melancon D. Gliomatosis cerebri. Radiology 1996; 198:831– 835. Schoenen J, De Leval L, Reznik M. Gliomatosis cerebri: clinical, radiological and pathological report of a case with a stroke-like onset. Acta Neurol Belg 1996; 96:294 –300. Porta-Etessam J, Berbel A, Martinez-Salio A, et al. Gliomatosis cerebri: MRI, SPECT and the study of pathology (letter). Rev Neurol 1999; 29:287–288. [Spanish] Rippe DJ, Boyko OB, Fuller GN, Friedman HS, Oakes WJ, Schold SC. Gadopentetate-dimeglumine-enhanced MR imaging of gliomatosis cerebri: appearance mimicking leptomeningeal tumor dissemination. AJNR Am J Neuroradiol 1990; 11:800 – 801. Spagnoli MV, Grossman RI, Packer RJ, et al. Magnetic resonance imaging determination of gliomatosis cerebri. Neuroradiology 1987; 29:15–18. Essig M, Schlemmer HP, Tronnier V, Hawighorst H, Wirtz R, van Kaick G. Fluid-attenuated inversion-recovery MR imaging of gliomatosis cerebri. Eur Radiol 2001; 11:303–308. Long DM. Capillary ultrastructure in human metastatic brain tumors. J Neurosurg 1979; 51:53–58. Strugar J, Rothbart D, Harrington W, Criscuolo GR. Vascular permeability factor in brain metastases: correlation with vasogenic brain edema and tumor angiogenesis. J Neurosurg 1994; 81:560 – 566. Cha S, Law M, Johnson G, et al. Peritumoral region: differentiation between primary high-grade neoplasms and solitary metastasis using dynamic contrastenhanced T2*-weighted echo-planar perfusion MR imaging (abstr). In: Proceedings of the 38th Annual Meeting of the American Society of Neuroradiology. Atlanta, Ga: American Society of Neuroradiology, 2000; 22. Strugar JG, Criscuolo GR, Rothbart D, Harrington WN. Vascular endothelial growth/permeability factor expression in human glioma specimens: correlation with vasogenic brain edema and tumorCha et al

88.

89. 90.

91.

92.

93.

94.

associated cysts. J Neurosurg 1995; 83: 682– 689. Eby NL, Grufferman S, Flannelly CM, Schold SC, Vogel FS, Burger PC. Increasing incidence of primary brain lymphoma in the US. Cancer 1988; 62: 2461–2465. Schabet M. Epidemiology of primary CNS lymphoma. J Neurooncol 1999; 43: 199 –201. Herrlinger U, Schabet M, Clemens M, et al. Clinical presentation and therapeutic outcome in 26 patients with primary CNS lymphoma. Acta Neurol Scand 1998; 97:257–264. Ling SM, Roach M, Larson DA, Wara WM. Radiotherapy of primary central nervous system lymphoma in patients with and without human immunodeficiency virus: ten years of treatment experience at the University of California San Francisco. Cancer 1994; 73:2570 – 2582. Reni M, Ferreri AJ, Garancini MP, Villa E. Therapeutic management of primary central nervous system lymphoma in immunocompetent patients: results of a critical review of the literature. Ann Oncol 1997; 8:227–234. Ernst TM, Chang L, Witt MD, et al. Cerebral toxoplasmosis and lymphoma in AIDS: perfusion MR imaging experience in 13 patients. Radiology 1998; 208: 663– 669. Frazzini VI, Cha S, Lu S, Johnson G, Knopp EA, Kricheff II. Dynamic contrast enhanced T2*-weighted echo-planar perfusion MR imaging of primary CNS lymphoma and glioblastoma multi-

Volume 223

䡠

Number 1

95.

96.

97.

98.

99.

100.

101.

forme (abstr). In: Proceedings of the 37th Annual Meeting of the American Society of Neuroradiology. San Diego, Calif: American Society of Neuroradiology, 1999; 185. Zagzag D, Miller DC, Kleinman GM, Abati A, Donnenfeld H, Budzilovich GN. Demyelinating disease versus tumor in surgical neuropathology: clues to a correct pathological diagnosis. Am J Surg Pathol 1993; 17:537–545. Hunter SB, Ballinger WE Jr, Rubin JJ. Multiple sclerosis mimicking primary brain tumor. Arch Pathol Lab Med 1987; 111:464 – 468. Giang DW, Poduri KR, Eskin TA, et al. Multiple sclerosis masquerading as a mass lesion. Neuroradiology 1992; 34: 150 –154. Prineas JW, McDonald WI. Demyelinating diseases. In: Greenfield’s neuropathology. 6th ed. Vol 1. New York, NY: Wiley, 1997; 814 – 846. Kepes JJ. Large focal tumor-like demyelinating lesions of the brain: intermediate entity between multiple sclerosis and acute disseminated encephalomyelitis? a study of 31 patients. Ann Neurol 1993; 33:18 –27. Nesbit GM, Forbes GS, Scheithauer BW, Okazaki H, Rodriguez M. Multiple sclerosis: histopathologic and MR and/or CT correlation in 37 cases at biopsy and three cases at autopsy. Radiology 1991; 180:467– 474. Kurihara N, Takahashi S, Furuta A, et al. MR imaging of multiple sclerosis simulating brain tumor. Clin Imaging 1996; 20:171–177.

102.

103.

104.

105.

106.

107.

108.

109.

Cha S, Pierce S, Knopp EA, et al. Dynamic contrast-enhanced T2*-weighted MR imaging of tumefactive demyelinating lesions. AJNR Am J Neuroradiol 2001; 22:1109 –1116. Rempp KA, Brix G, Wenz F, Becker CR, Guckel F, Lorenz WJ. Quantification of regional cerebral blood flow and volume with dynamic susceptibility contrast-enhanced MR imaging. Radiology 1994; 193:637– 641. Ostergaard L, Sorensen AG, Kwong KK, Weisskoff RM, Gyldensted C, Rosen BR. High resolution measurement of cerebral blood flow using intravascular tracer bolus passages. II. Experimental comparison and preliminary results. Magn Reson Med 1996; 36:726 –736. Keir SL, Wardlaw JM. Systematic review of diffusion and perfusion imaging in acute ischemic stroke. Stroke 2000; 31: 2723–2731. Fisher M, Albers GW. Applications of diffusion-perfusion magnetic resonance imaging in acute ischemic stroke. Neurology 1999; 52:1750 –1756. Ueda T, Yuh WT, Taoka T. Clinical application of perfusion and diffusion MR imaging in acute ischemic stroke. J Magn Reson Imaging 1999; 10:305–309. Lev MH, Rosen BR. Clinical applications of intracranial perfusion MR imaging. Neuroimaging Clin N Am 1999; 9:309 – 331. Powers WJ, Zivin J. Magnetic resonance imaging in acute stroke: not ready for prime time. Neurology 1998; 50:842– 843.

Intracranial Mass Lesions: Echo-planar Perfusion MR Imaging

䡠

29