| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Neuroradiology |
1 From the Department of Radiology (S.G.W., S.C., G.J., P.L., M.L., D.L.K., S.D.P.) and Department of Environmental Medicine, Division of Biostatistics (X.X.), New York University School of Medicine, 550 First Ave, New York, NY 10016. Received June 7, 2001; revision requested August 1; revision received September 27; accepted March 1, 2002. S.G.W. supported by the Swiss National Science Foundation/Karger Stiftung and by Novartis Stiftung. S.C. is a recipient of an RSNA Seed Grant 3. Address correspondence to G.J. (e-mail: johnson@mcmri19.med.nyu.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Three independent observers (neuroradiology fellows) who were blinded to the histopathologic diagnosis performed rCBV measurements in 50 patients with various intracranial mass lesions. Three different methods were compared. With method 1, placement of a single region of interest was guided by a color overlay map. With methods 2 and 3, the highest rCBV value and the mean of repeated rCBV measurements, respectively, were recorded. Calculations of the intraclass correlation coefficient, coefficient of variation (CV), and descriptive statistics were used to determine the levels of reproducibility. A multiple linear regression model was used to evaluate for possible explanatory factors for interobserver variance.
RESULTS: Method 2 had, overall, the best reproducibility of all techniques, with an intraclass interobserver correlation coefficient of 0.71 (indicating good agreement), interobserver CV of 30%, and intraobserver CV in the range of 32%–41%. Measurement variations between observers correlated significantly (P < .001) with increasing rCBV values.
CONCLUSION: In this study, interobserver and intraobserver reproducibility of rCBV measurements were clinically acceptable.
© RSNA, 2002
Index terms: Brain, blood flow, 13.36 • Brain neoplasms, 13.36 • Brain neoplasms, MR, 13.12141, 13.121416, 13.12143
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Dynamic contrast material–enhanced T2- or T2*-weighted perfusion MR imaging has been shown to have great potential in the evaluation of a variety of intracranial mass lesions, especially tumors. For example, rCBV measurements may correlate with the degree of vascularity in gliomas and have been shown to be higher in high-grade than in low-grade tumors (3,4). Moreover, in a recent study, rCBV measurements have been used to assess tumor activity during antiangiogenic therapy (5). In all of these studies, perfusion MR imaging provided valuable pathophysiologic information not available with conventional MR imaging. To date, however, these results have been obtained by experienced radiologists working in a research environment on selected predefined disease entities. The reproducibility of results of perfusion MR imaging must be established in a more clinically realistic setting before the technique can find more widespread use.
The purpose of our study was to assess inter- and intraobserver reproducibility for different techniques of rCBV measurement in patients with intracranial mass lesions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Deliberate selection was necessary to include patients with a variety of intracranial mass lesions with different characteristics and CBV values, because random selection would have resulted in an overwhelming preponderance of patients with high-grade gliomas—by far the most common lesion type seen in our clinical practice due to the referral pattern. A higher number of glioblastomas than other lesion types was included because they are the most common malignant brain tumor in the general population (6). Ideally, representative proportions of each lesion type would have been included. However, reliable epidemiologic data on the distribution of the diseases we investigated are not available.
For each type of lesion, cases for evaluation were selected at random from our department’s perfusion MR imaging database by a neuroradiologist (S.G.W.) who was unaware of the imaging findings. Patients were excluded when the perfusion MR imaging study was technically inadequate due to gross patient movement or the lesion could not be identified due to severe susceptibility artifacts. Further criteria for case selection were the following: the patient had to have undergone no surgery (including biopsy) before perfusion MR imaging, and a histopathologically confirmed diagnosis was available (except in demyelinating lesions, where a diagnosis was rendered on the basis of either biopsy results or clinical and laboratory findings). Approval for this study was not required by our institutional review board, nor was informed consent.
Observers
CBV measurements were calculated independently by three board-certified radiologists who were neuroradiology fellows (M.L., D.L.K., S.D.P.). All observers were familiar with the principles of perfusion MR imaging but had little experience with quantitative evaluation of perfusion MR imaging studies. The observers were blinded to the patient’s history and to the diagnosis of the intracranial mass lesion.
MR Imaging and Postprocessing
In all patients, conventional transverse T2-weighted (repetition time msec/echo time msec, 3,400/119) MR images of the brain were obtained; transverse T1-weighted (600/14) MR images were also obtained before and after the administration of contrast material. Dynamic contrast-enhanced perfusion MR imaging was performed with a fat-suppressed T2*-weighted echo-planar imaging sequence and the following parameters: 1,000/54; field of view, 230 x 230 mm; section thickness, 5 or 7 mm; data matrix, 128 x 128; and in-plane voxel size, 1.8 x 1.8 mm. Between five and seven sections were obtained with a gap of 0%–30% of the section thickness to cover the entire lesion volume identified on T2-weighted images. A series of 60 multisection acquisitions was acquired at 1-second intervals. The first 10 acquisitions were obtained before the injection of contrast material to establish a precontrast baseline. At the 10th acquisition, 0.1 mmol/kg of body weight of gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Wayne, NJ) was injected with a power injector (Spectris; Medrad, Pittsburgh, Pa) at a rate of 5 mL/sec through an 18- or 20-gauge intravenous catheter; this was immediately followed by a bolus injection of saline (at 5 mL/sec to a total of 20 mL).
The echo-planar MR images were transferred to a Sun Ultra 10 workstation (Sun Microsystems, Santa Clara, Calif) for postprocessing. Image analysis and CBV calculations were performed with software developed in-house in the C and IDL programming languages (Research Systems, Boulder, Colo). Details of data processing are fully described by Knopp et al (4), and only a summary will be given here. During the first pass of the bolus of paramagnetic contrast material, there is a decrease in signal intensity on T2*-weighted images. The concentration of gadolinium can be calculated from signal intensity changes to obtain a plot of tissue gadolinium concentration over time. The area under this curve is proportional to the regional CBV (1). Correction for contrast material recirculation and leakage (which invalidate the CBV calculation) was performed by subtracting baseline signal intensity measurements from the measurements obtained as the bolus of contrast material moved through the patient. The beginning and end of the bolus were defined by the images at which the signal intensity came within 1 SD of the mean signal intensity observed before and after bolus administration, respectively. The baseline values calculated between these two points were then subtracted from the signal intensity measurements obtained during the movement of the bolus of contrast material through the tissue.
CBV values were expressed relative to those of an internal reference, normal-appearing contralateral white matter. CBV values can be calculated either in regions of interest (ROIs) or on a pixel-by-pixel basis to form color overlay maps. For this study, one neuroradiologist (S.G.W.) created color maps in a standardized fashion. CBV values between a threshold rCBV value of 1.25 and a maximum rCBV value of 5.00 (in relation to normal-appearing contralateral white matter) were displayed on a green-to-red color scale. CBV values that fell below the threshold were not displayed as color values; instead, the signal intensity values in the unprocessed images were displayed in a conventional gray scale to depict the underlying anatomy. Values that exceeded 5 were displayed as red, the color at the maximum end of the scale.
Image Analysis
The three observers independently evaluated 50 cases within 5 weeks. Cases were evaluated in groups of 10 that were chosen randomly by each reader. Groups of 10 were chosen so as not to tire the readers but still allow completion of the study in a reasonable time. After this first evaluation, a second evaluation was performed with the same techniques for obtaining and recording rCBV values. This second evaluation was performed 6 weeks later, and the presentation order of the cases was changed to avoid recall bias in the observers. Each case consisted of a hard copy of the conventional MR images and the perfusion MR images that were stored at the workstation. In addition to the unprocessed perfusion MR images, color rCBV maps were available to the observers. The observers obtained rCBV values in ROIs in the lesion and inspected the raw perfusion MR images without the overlay color map to ensure that ROIs were not placed over large vessels. Three different protocols for obtaining lesion rCBV values were used.
Method 1: single.—Each observer evaluated the conventional MR images and the color overlay maps and chose a single ROI in a location that he or she believed contained the highest rCBV.
Method 2: maximum.—Each observer picked a minimum of four ROIs where he or she believed high rCBV values would be found. If the tumors were heterogeneous and these four rCBV values were inconsistent, they were advised to pick one or two more ROIs, for a total of five or six ROIs.
The highest of the rCBV values found in these four to six ROIs and in the ROI selected in method 1 were recorded.
Method 3: mean.—The means of the three highest rCBV values in the four to six ROIs selected in method 2 and the ROI selected in method 1 were calculated.
The average of multiple readings (used in method 3) should best characterize a uniform lesion, and this method has been used previously (5). The maximum reading (used in method 2), on the other hand, may better characterize malignancies because a tumor is as malignant as its most malignant (and hypervascular) region. This method has also been used previously (3,4). We also wished to investigate method 1 to determine whether a quick protocol of assessing a single ROI could produce results similar to those produced with more elaborate analyses.
In five patients in whom multiple lesions were found, the largest lesion was chosen for evaluation. To minimize confounding factors in the rCBV analysis, the size of the ROIs assessed was kept constant (radius = 1 image pixel, 1.8 mm). The observers were advised to complete the evaluation of a given case within 15 minutes. In our clinical experience, 15 minutes is more than enough to perform the analysis, so this limit is unlikely to affect outcome.
Conventional MR images were analyzed by two neuroradiologists (S.G.W., S.C.) for the presence of contrast enhancement, heterogeneity, cyst or necrosis, and susceptibility artifacts; these criteria were adopted, in part, from a study of the classification of gliomas by Dean et al (7). An estimate of lesion volume was obtained by multiplying the product of the three largest perpendicular lesion diameters by 0.5.
To assess errors due to variations in the normal-appearing contralateral white matter used as the internal standard, 10 ROIs (radius = 1 pixel) were evaluated in normal-appearing white matter around an initial ROI of the same size. The resulting ratios were recorded. This analysis was performed by a neuroradiologist (S.G.W.).
Data Analysis
Four analyses were performed. Absolute rCBV values obtained with methods 1 and 2 were compared. Inter- and intraobserver reproducibility were both quantified with the intraclass correlation coefficient, the coefficient of variation (CV), and by calculating differences between pairs of observers. Linear regression was used to determine whether interobserver differences were correlated with lesion characteristics. Finally, the precision of measurements of normal-appearing white matter was evaluated.
CBV measurement with different methods.—A paired two-tailed Student t test was used to determine whether acquiring measurements from multiple ROIs significantly increased measured values. Only methods 1 and 2 were compared in this analysis.
Inter- and intraobserver reproducibility.—Inter- and intraobserver reproducibility was assessed with the intraclass correlation coefficient Ri (8). This is a true index of agreement between observers, in contrast to the conventional Pearson product moment correlation coefficient, which is a measure of linear association rather than of agreement. Ri is derived from a two-way mixed analysis of variance with subjects treated as a random effect and observers treated as a fixed effect. In accordance with Oppo et al (9), the following criteria for clinically relevant agreement were used: an Ri value less than 0.40 was considered poor, an Ri value of 0.40–0.59 was considered fair, an Ri value of 0.60–0.74 was considered good, and an Ri value greater than 0.74 was considered excellent.
To further quantify the inter- and intraobserver reproducibility, the CV was calculated for every lesion according to the following formula: 100 x SD/mean. The averaged value across the patients was recorded.
Interobserver variability was also quantified by calculating the relative paired difference in rCBV measurements between each pair of observers divided by the mean rCBV measurement for each lesion. The paired difference between observers i and j (Dij) is thus
Correlation of interobserver differences with lesion characteristics.—A multiple linear regression model was used to evaluate possible associations between interobserver variation and measured rCBV value, lesion volume, contrast enhancement, the presence of necrosis or cyst, lesion heterogeneity, and the presence of susceptibility artifacts. For this evaluation, interobserver variation was defined as the SD of rCBV values obtained with method 2, because initial results had indicated that this method is most reliable.
Estimation of precision of measurements of white matter.—The precision of measurements of normal-appearing white matter was estimated by measuring rCBV in 10 ROIs close to an initial reference ROI (also in white matter) in 10 patients. Variation within patients was measured by calculating the CV in each patient and taking the mean across patients. Variation across the patient group was measured by calculating the mean in each patient and then the CV of these means.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
In practice, it proved relatively simple to avoid placing ROIs over large vessels. The susceptibility loss of signal intensity is so great around these vessels that they become obvious, especially on MR images acquired during the arrival of the bolus of contrast material (Fig 1).
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
CBV measurements obtained with dynamic contrast-enhanced susceptibility-weighted perfusion MR imaging are subject to systematic and random errors. Systematic errors, such as those caused by violations of the assumptions governing intravascular indicator dilution theory, can cause under- or overestimation of rCBV (10,11). In principle, however, systematic errors can be minimized with the use of appropriate techniques and models. Random errors due to noise in the data and tissue variability pose a more difficult problem.
In this study, three approaches—method 1, where a single rCBV value was recorded; method 2, where the highest rCBV value obtained in four to six ROIs was recorded; and method 3, where the mean of the three highest rCBV values was recorded—to determining rCBV values from perfusion MR imaging data were compared to discover the technique that yielded the most reliable rCBV value. The intraclass correlation of all techniques was classified as "good." Likewise, averaged CV and absolute paired difference measurements were similar. Although no technique was clearly better than the others, a definite trend was observed, with method 2 generally resulting in the best agreement, followed by successively lower agreements with method 3 and method 1. In particular, the number of cases with marked discrepancies between observers (differences of >100%) was much lower with method 2 (16 of 150 comparisons) than with method 3 (24 of 150) and method 1 (40 of 150). Moreover, method 1 generally yielded significantly lower rCBV values than did method 2. Method 3 did not result in improved interobserver or intraobserver reliability, and, not surprisingly, it yielded lower rCBV values. Thus method 2 appears to be the best approach for measuring rCBV values. Of note, intraobserver reliability was more dependent on the skill of the observer than on the technique used.
The color-encoded overlay map used in this study encoded a limited range of rCBV values: Pixels with rCBV values greater than five were all encoded with the same color. Thus, use of the map as a guide for placing ROIs in lesions with large areas of high rCBV (>5) was rather unreliable. While it is possible to color encode a wider range of rCBV values, this would impair detection of pixels with lower but still abnormal rCBVs. A variable range for the color map can confuse interpretations, because rCBV values corresponding to a particular color can vary. Moreover, in an automatically scaled range, the highest value will correspond to veins, and, therefore the total rCBV range will be very large. However, the relatively poor reliability of the method 1 can only in part be explained by the limited color range of the maps because considerable fractional errors in lesions with rCBV values less than 5 were observed; these were presumably due to the inherent limitations of the technique. Notwithstanding these limitations, the color overlay map demonstrates the overall perfusion characteristics of the lesion and serves as a useful, if limited, "road map" for placement of ROIs.
Even when method 2 was used, the CV between the observers in our study was 30%, and the averaged absolute paired differences between pairs of observers were in the range of 38%–51%. Although use of this method will still allow reliable characterization of a lesion as hyperperfused or "hot" in the majority of cases, it would certainly be desirable to have a more reliable technique. We believe that the relatively large differences between observers can be explained in part by the inexperience of the observers. However, part of the variation can certainly be attributed to noise in the measurements in both the lesions and the reference ROIs. As an estimate for this variation, we found a CV on the order of 20% for repeated measurements of normal-appearing white matter. This may also explain, at least in part, why variations in rCBV measurements were larger with larger rCBV values. The rCBV is calculated by dividing the measurement of the lesion by the measurement of normal-appearing white matter. Thus, the same amount of noise in the white matter denominator will produce larger errors when the numerator is large. Furthermore, it can be assumed that large rCBV values will tend to make ROI placement more critical. CBV values will vary from the peak value somewhere in the lesion to normal values in surrounding unaffected white matter. Thus, the CBV "gradient" will be greater with large CBV values, and a small misplacement of the ROI away from the region of peak CBV will result in a larger error. The correlation of measurement variation with CBV is reassuring. It suggests that large errors will usually only occur when rCBV is substantially elevated, and hence the likelihood of an incorrect clinical conclusion is minimized.
Apart from this positive correlation, we did not find any other lesion characteristics that were significantly related to variations in rCBV measurement. This suggests that the technique can be applied without special caution to the whole range of intracranial mass lesions.
There were well-recognized limitations in our study. First, we used a relatively small ROI for the measurement of rCBV. This will tend to increase the level of noise when compared with the use of larger ROIs. On the other hand, we found use of larger ROIs to be impractical in most lesions due to averaging of rCBV measurements over heterogeneous regions and impractical in normal-appearing white matter due to inclusion of large blood vessels. For similar reasons, and unlike other researchers (3), we did not use voxel smoothing. The influence of ROI size and voxel smoothing on the reliability of rCBV measurements could usefully be evaluated.
A second limitation of our study was the atypical spectrum of lesions studied. This approach ensured, however, that a variety of lesions with a large range of rCBV values were analyzed. Thus, it was possible to analyze the influence of CBV on interobserver variation. A wide spectrum of rCBV values would not have been guaranteed if consecutive cases had been selected.
We have made the implicit assumption that a lesion can be characterized by its highest rCBV value. Although this may not be true in general, we believe that the assumption is justified in the identification and assessment of tumors—the area in which rCBV measurements have been found most useful. CBV is correlated with the degree of neovascularization (3) and hence malignancy. Although areas of low CBV are also found in tumors, these represent areas of necrosis. Because the tumor is as malignant as its most malignant region, we believe that characterizing a tumor by its highest CBV value is valid. Knowledge of the distribution of CBV values may, however, be useful in the assessment of other lesions.
Our results apply only to measurements of rCBV. There are a number of other methods of assessing cerebral microcirculation. A few examples are spin-echo acquisitions, MR methods that incorporate correction for possible T1 effects in leaky tumors (12,13), blood flow measurements that require an arterial input function (14), and methods based on pharmacokinetic modeling (15). It would be inappropriate to extrapolate the results from this study to any other technique.
We found no correlation between rCBV and tumor volume. However, our estimation of tumor volume was based on an approximate formula (the product of the three largest perpendicular lesion diameters multiplied by 0.5). We believe that the use of this simple formula is justified for several reasons. First, the formula is commonly used in radiologic studies (16). Second, researchers have shown that there is no added benefit in reporting bidimensional or tridimensional measurements over reporting the maximum axial dimension of tumors (17). Third, any estimate of tumor volume is, of necessity, approximate because many factors (eg, MR imaging sequence used, window setting) also influence the result. Finally, use of the more exact scale factor of 
/6 (0.524) rather than 0.5 would have no influence on results of linear regression analysis.
Finally, we wished to perform this study in as typical a clinical setting as possible. To this end, our observers were chosen because they had relatively limited experience with CBV measurements. There were, however, several ways in which our setting was atypical. The observers were fellows and were therefore relatively inexperienced overall. There was a greater proportion of unusual lesions (eg, tumefactive demyelinating lesions) than would occur in a typical workload. During the study period, more cases were seen than might be typical. Nonetheless, we do not believe that these shortcomings introduce serious bias into our results.
In conclusion, our results demonstrate that the reproducibility of rCBV measurements obtained by using dynamic contrast-enhanced T2*-weighted imaging is clinically acceptable for the evaluation of intracranial mass lesions. The most reliable method of obtaining the highest rCBV value in a lesion is to select from multiple measurements. There is a tendency for greater variations among observers as rCBV increases.
|
FOOTNOTES |
|---|
Abbreviations: CBV = cerebral blood volume, CV = coefficient of variation, rCBV = relative CBV, ROI = region of interest
Author contributions: Guarantor of integrity of entire study, S.G.W.; study concepts, S.G.W., G.J.; study design, S.G.W., S.C.; literature research, S.G.W., S.C., G.J.; clinical studies, S.G.W., S.C., M.L., D.L.K., S.D.P.; data acquisition, S.G.W., S.C., P.L., M.L., D.L.K., S.D.P.; data analysis/interpretation, S.G.W., G.J., X.X.; statistical analysis, S.G.W., G.J., X.X.; manuscript preparation, S.G.W., G.J.; manuscript definition of intellectual content, S.G.W., S.C., G.J.; manuscript editing, S.G.W., G.J.; manuscript revision/review and final version approval, S.G.W., S.C., G.J., M.L., S.D.P., X.X.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Bisdas, M. Kirkpatrick, P. Giglio, C. Welsh, M.V. Spampinato, and Z. Rumboldt Cerebral Blood Volume Measurements by Perfusion-Weighted MR Imaging in Gliomas: Ready for Prime Time in Predicting Short-Term Outcome and Recurrent Disease? AJNR Am. J. Neuroradiol., April 1, 2009; 30(4): 681 - 688. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K.E. Emblem, D. Scheie, P. Due-Tonnessen, B. Nedregaard, T. Nome, J.K. Hald, K. Beiske, T.R. Meling, and A. Bjornerud Histogram Analysis of MR Imaging-Derived Cerebral Blood Volume Maps: Combined Glioma Grading and Identification of Low-Grade Oligodendroglial Subtypes AJNR Am. J. Neuroradiol., October 1, 2008; 29(9): 1664 - 1670. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Hirai, R. Murakami, H. Nakamura, M. Kitajima, H. Fukuoka, A. Sasao, M. Akter, Y. Hayashida, R. Toya, N. Oya, et al. Prognostic Value of Perfusion MR Imaging of High-Grade Astrocytomas: Long-Term Follow-Up Study AJNR Am. J. Neuroradiol., September 1, 2008; 29(8): 1505 - 1510. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. B. Caseiras, J.S. Thornton, T. Yousry, C. Benton, J. Rees, A.D. Waldman, and H.R. Jager Inclusion or Exclusion of Intratumoral Vessels in Relative Cerebral Blood Volume Characterization in Low-Grade Gliomas: Does It Make a Difference? AJNR Am. J. Neuroradiol., June 1, 2008; 29(6): 1140 - 1141. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. E. Emblem, B. Nedregaard, T. Nome, P. Due-Tonnessen, J. K. Hald, D. Scheie, O. C. Borota, M. Cvancarova, and A. Bjornerud Glioma Grading by Using Histogram Analysis of Blood Volume Heterogeneity from MR-derived Cerebral Blood Volume Maps Radiology, June 1, 2008; 247(3): 808 - 817. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Law, R. J. Young, J. S. Babb, N. Peccerelli, S. Chheang, M. L. Gruber, D. C. Miller, J. G. Golfinos, D. Zagzag, and G. Johnson Gliomas: Predicting Time to Progression or Survival with Cerebral Blood Volume Measurements at Dynamic Susceptibility-weighted Contrast-enhanced Perfusion MR Imaging Radiology, May 1, 2008; 247(2): 490 - 498. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Danchaivijitr, A. D. Waldman, D. J. Tozer, C. E. Benton, G. Brasil Caseiras, P. S. Tofts, J. H. Rees, and H. R. Jager Low-Grade Gliomas: Do Changes in rCBV Measurements at Longitudinal Perfusion-weighted MR Imaging Predict Malignant Transformation? Radiology, April 1, 2008; 247(1): 170 - 178. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Ay, E. M. Arsava, M. Vangel, B. Oner, M. Zhu, O. Wu, A. Singhal, W. J. Koroshetz, and A. G. Sorensen Interexaminer Difference in Infarct Volume Measurements on MRI: A Source of Variance in Stroke Research Stroke, April 1, 2008; 39(4): 1171 - 1176. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.S. Seo, K.-H. Chang, D.G. Na, B.J. Kwon, and D.H. Lee High b-Value Diffusion (b = 3000 s/mm2) MR Imaging in Cerebral Gliomas at 3T: Visual and Quantitative Comparisons with b = 1000 s/mm2 AJNR Am. J. Neuroradiol., March 1, 2008; 29(3): 458 - 463. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Lu, E. Pollack, R. Young, J.S. Babb, G. Johnson, D. Zagzag, R. Carson, J.H. Jensen, J.A. Helpern, and M. Law Predicting Grade of Cerebral Glioma Using Vascular-Space Occupancy MR Imaging AJNR Am. J. Neuroradiol., February 1, 2008; 29(2): 373 - 378. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Lin, K.S. Kazmi, M. Law, J. Babb, N. Peccerelli, and B.K. Pramanik Measuring Elevated Microvascular Permeability and Predicting Hemorrhagic Transformation in Acute Ischemic Stroke Using First-Pass Dynamic Perfusion CT Imaging AJNR Am. J. Neuroradiol., August 1, 2007; 28(7): 1292 - 1298. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Y. Poussaint, P.C. Phillips, S. Vajapeyam, F.H. Fahey, R.L. Robertson, S. Osganian, U. Ramamurthy, R.V. Mulkern, S.T. Treves, J.M. Boyett, et al. The Neuroimaging Center of the Pediatric Brain Tumor Consortium-Collaborative Neuroimaging in Pediatric Brain Tumor Research: A Work in Progress AJNR Am. J. Neuroradiol., April 1, 2007; 28(4): 603 - 607. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Law, R. Young, J. Babb, E. Pollack, and G. Johnson Histogram Analysis versus Region of Interest Analysis of Dynamic Susceptibility Contrast Perfusion MR Imaging Data in the Grading of Cerebral Gliomas AJNR Am. J. Neuroradiol., April 1, 2007; 28(4): 761 - 766. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Inglese, S.-J. Park, G. Johnson, J. S. Babb, L. Miles, H. Jaggi, J. Herbert, and R. I. Grossman Deep Gray Matter Perfusion in Multiple Sclerosis: Dynamic Susceptibility Contrast Perfusion Magnetic Resonance Imaging at 3 T Arch Neurol, February 1, 2007; 64(2): 196 - 202. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. V. Spampinato, J. K. Smith, L. Kwock, M. Ewend, J. D. Grimme, D. L. A. Camacho, and M. Castillo Cerebral Blood Volume Measurements and Proton MR Spectroscopy in Grading of Oligodendroglial Tumors Am. J. Roentgenol., January 1, 2007; 188(1): 204 - 212. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M P Sampat, G J Whitman, T W Stephens, L D Broemeling, N A Heger, A C Bovik, and M K Markey The reliability of measuring physical characteristics of spiculated masses on mammography Br. J. Radiol., December 1, 2006; 79(Special_Issue_2): S134 - S140. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Law, R. Young, J. Babb, M. Rad, T. Sasaki, D. Zagzag, and G. Johnson Comparing Perfusion Metrics Obtained from a Single Compartment Versus Pharmacokinetic Modeling Methods Using Dynamic Susceptibility Contrast-Enhanced Perfusion MR Imaging with Glioma Grade. AJNR Am. J. Neuroradiol., October 1, 2006; 27(9): 1975 - 1982. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Weber, S. Zoubaa, M. Schlieter, E. Juttler, H. B. Huttner, K. Geletneky, C. Ittrich, M. P. Lichy, A. Kroll, J. Debus, et al. Diagnostic performance of spectroscopic and perfusion MRI for distinction of brain tumors Neurology, June 27, 2006; 66(12): 1899 - 1906. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Provenzale, S. Mukundan, and D. P. Barboriak Diffusion-weighted and Perfusion MR Imaging for Brain Tumor Characterization and Assessment of Treatment Response. Radiology, June 1, 2006; 239(3): 632 - 649. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.L. Boxerman, K.M. Schmainda, and R.M. Weisskoff Relative cerebral blood volume maps corrected for contrast agent extravasation significantly correlate with glioma tumor grade, whereas uncorrected maps do not. AJNR Am. J. Neuroradiol., April 1, 2006; 27(4): 859 - 867. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Law, S. Oh, J. S. Babb, E. Wang, M. Inglese, D. Zagzag, E. A. Knopp, and G. Johnson Low-Grade Gliomas: Dynamic Susceptibility-weighted Contrast-enhanced Perfusion MR Imaging--Prediction of Patient Clinical Response Radiology, February 1, 2006; 238(2): 658 - 667. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. A. Taschner, S. G. Wetzel, M. Tolnay, J. Froehlich, A. Merlo, and E. W. Radue Characteristics of Ultrasmall Superparamagnetic Iron Oxides in Patients with Brain Tumors Am. J. Roentgenol., December 1, 2005; 185(6): 1477 - 1486. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Bulakbasi, M. Kocaoglu, A. Farzaliyev, C. Tayfun, T. Ucoz, and I. Somuncu Assessment of Diagnostic Accuracy of Perfusion MR Imaging in Primary and Metastatic Solitary Malignant Brain Tumors AJNR Am. J. Neuroradiol., October 1, 2005; 26(9): 2187 - 2199. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. C. M. Maia Jr, S. M. F. Malheiros, A. J. da Rocha, C. J. da Silva, A. A. Gabbai, F. A. P. Ferraz, and J. N. Stavale MR Cerebral Blood Volume Maps Correlated with Vascular Endothelial Growth Factor Expression and Tumor Grade in Nonenhancing Gliomas AJNR Am. J. Neuroradiol., April 1, 2005; 26(4): 777 - 783. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Miller, H. H. Pien, D. Sahani, A. G. Sorensen, and J. H. Thrall Imaging Angiogenesis: Applications and Potential for Drug Development J Natl Cancer Inst, February 2, 2005; 97(3): 172 - 187. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Law, K. Kazmi, S. Wetzel, E. Wang, C. Iacob, D. Zagzag, J. G. Golfinos, and G. Johnson Dynamic Susceptibility Contrast-Enhanced Perfusion and Conventional MR Imaging Findings for Adult Patients with Cerebral Primitive Neuroectodermal Tumors AJNR Am. J. Neuroradiol., June 1, 2004; 25(6): 997 - 1005. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Law, A. M. Saindane, Y. Ge, J. S. Babb, G. Johnson, L. J. Mannon, J. Herbert, and R. I. Grossman Microvascular Abnormality in Relapsing-Remitting Multiple Sclerosis: Perfusion MR Imaging Findings in Normal-appearing White Matter Radiology, June 1, 2004; 231(3): 645 - 652. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Law, S. Yang, J. S. Babb, E. A. Knopp, J. G. Golfinos, D. Zagzag, and G. Johnson Comparison of Cerebral Blood Volume and Vascular Permeability from Dynamic Susceptibility Contrast-Enhanced Perfusion MR Imaging with Glioma Grade AJNR Am. J. Neuroradiol., May 1, 2004; 25(5): 746 - 755. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Law, S. Yang, H. Wang, J. S. Babb, G. Johnson, S. Cha, E. A. Knopp, and D. Zagzag Glioma Grading: Sensitivity, Specificity, and Predictive Values of Perfusion MR Imaging and Proton MR Spectroscopic Imaging Compared with Conventional MR Imaging AJNR Am. J. Neuroradiol., November 1, 2003; 24(10): 1989 - 1998. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Yang, M. Law, D. Zagzag, H. H. Wu, S. Cha, J. G. Golfinos, E. A. Knopp, and G. Johnson Dynamic Contrast-Enhanced Perfusion MR Imaging Measurements of Endothelial Permeability: Differentiation between Atypical and Typical Meningiomas AJNR Am. J. Neuroradiol., September 1, 2003; 24(8): 1554 - 1559. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
) and method 2 (
). The dashed line represents the mean difference in rCBV measurements between the two observers when method 1 was used; the solid line represents the corresponding value when method 2 was used.