| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Neuroradiology |
1 From the Departments of Radiology (S.L., D.A., G.J., M.L. R.I.G.) and Pathology (D.Z.), New York University Medical Center, NY. Received April 30, 2003; revision requested July 11; final revision received November 11; accepted December 19. Supported by RSNA Research Resident Grant RR0201. Address correspondence to S.L., 564 First Ave, #19D, New York, NY 10016.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: In this study, diffusion-tensor MR imaging was performed preoperatively in 40 patients with intracranial neoplasms, including meningiomas, metastatic lesions, glioblastomas multiforme, and low-grade gliomas. Histograms of mean diffusivity (MD) and fractional anisotropy (FA) were used to analyze both the tumor and the associated T2 signal intensity abnormality. An additional metric, the tumor infiltration index (TII), was evaluated. The TII is a measure of the change in FA presumably caused by tumor cells infiltrating the peritumoral edema. Student t test and least-squares linear regression analyses were performed.
RESULTS: Peritumoral MD and FA values indicated no statistically significant difference between intra- and extraaxial lesions or between high- and low-grade gliomas. Regarding intraaxial tumors, the measured mean peritumoral MD of metastatic lesions, 0.733 x 10–3 mm2/sec ± 0.061 (SD), was significantly higher than that of gliomas, 0.587 ± 0.093 x 10–3 mm2/sec (P < .05). There was also a statistically significant difference between the TIIs of the edema surrounding meningiomas and metastases (mean, 0 ± 35) and the TIIs of the edema surrounding gliomas (mean, 64 ± 59) (P < .05).
CONCLUSION: Peritumoral diffusion-tensor MR imaging metrics enable the differentiation of solitary intraaxial metastatic brain tumors from gliomas. In addition, the TII enables one to distinguish presumed tumor-infiltrated edema from purely vasogenic edema.
© RSNA, 2004
Index terms: Brain, diffusion, 10.91 • Brain neoplasms, diagnosis, 10.30, 10.363, 10.366, 10.38 • Brain neoplasms, MR, 10.121416, 10.12143, 10.12144 • Magnetic resonance (MR), diffusion tensor, 10.12144
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Diffusion is the molecular movement of bulk water. In large volumes, diffusion is equal in all directions. However, in tissues, the presence of cell membranes can impede free diffusion. Apparent diffusion coefficient (ADC) measurements can therefore be used to estimate the size of the extracellular space. The utility of the ADC has been the focus of most of the more recently published literature addressing the characterization of intracranial neoplasms. ADCs, other than enabling differentiation between epidermoids and arachnoid cysts (4), have represented only a limited contribution in the preoperative differentiation of the histologic features of tumors. No consistent differences in ADC measurements among gliomas of different grades or among gliomas, meningiomas, and metastases have been observed (5–8). Indeed, there have been published editorials attesting that "the bloom is off the rose" when it comes to ADC imaging of brain tumors (9).
It has also been recognized that in the presence of well-organized elongated structures such as white matter bundles, diffusion may not be equal in all directions. In this setting, diffusion can be described by using the diffusion tensor, which is a 3 x 3 matrix that quantifies direction-dependent diffusion. Several useful indexes of diffusion can be derived by using an advanced MR technique called diffusion-tensor imaging. The first diffusion-tensor MR imaging metric is the mean diffusivity (MD), which, like the ADC, is a measure of magnitude. The second metric is the fractional anisotropy (FA), which is a measure of the directionality of molecular motion; in the brain, this parameter may reflect white matter integrity.
The purpose of our study was to determine whether the diffusion-tensor MR imaging metrics (hereafter, diffusion-tensor metrics) of brain tumors and of the associated peritumoral edema can be used to differentiate intra- from extraaxial lesions, metastatic lesions from gliomas, and high- from low-grade gliomas (Fig 1). For this study, the term peritumoral edema referred to the region of tissue surrounding the T2 signal intensity abnormality; this region may have consisted solely of vasogenic edema or of edema with infiltrating tumor cells.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
As part of an ongoing study, patients who are scheduled to undergo surgery for intracranial tumors at our institution now are examined with preoperative diffusion-tensor MR imaging. For the current study, the acquired diffusion-tensor MR images were visually inspected for the presence of peritumoral T2 signal intensity abnormalities, and image postprocessing and analysis were performed. Then, the pathologic diagnoses, as reported by one of the authors (D.Z.) in the neuropathology department, were reviewed and the first 10 patients with a peritumoral T2 signal intensity abnormality in each tumor category were included. Poorly defined tumors, such as some low-grade gliomas, were considered to have no identifiable peritumoral region and thus were excluded.
MR Imaging
MR imaging was performed by using a 1.5-T system (Magnetom Vision; Siemens, Iselin, NJ) with a diffusion-weighted echo-planar sequence and the following parameters: 4,000/98 (repetition time msec/echo time msec), 20 contiguous sections, a 5.0-mm section thickness, four signals acquired, a 128 x 128 matrix, a 240 x 240-mm field of view, and a total imaging time of 1 minute 44 seconds. A total of seven image sets were acquired: six with noncolinear diffusion-weighting gradients and a b value of 1,000 sec/mm2 and one without diffusion weighting. The diffusion gradients in the x, y, and z directions were, respectively, (1, 1, and 0); (0, 1, and 1); (1, 0, and 1); (–1, 1, and 0); (0, –1, and 1); and (1, 0, and –1). The dual-echo diffusion-tensor MR imaging sequence was carefully designed to minimize eddy currents and geometric distortion (10). All raw data were transferred to a computer workstation (Sun Ultra 10; Sun Microsystems, Palo Alto, Calif) for analysis by using programs in interactive data language (IDL; Research Systems, Boulder, Colo) that were developed in house.
Image Analysis
The MD and FA were calculated voxel by voxel by using the following standard algorithms:
|
|
1, 2, and 3 are the three eigenvalues describing a diffusion tensor and 
is one-third times the sum of these three eigenvalues. For each image section, these values were used to construct MD and FA overlay maps. After inspection of the conventional MR images, one of the authors (S.L.), who had approximately 4 years of experience in brain tumor imaging, placed two regions of interest (ROIs) on a T2-weighted MR image with guidance from a board-certified neuroradiologist (M.L.) who had more than 5 years of neuroradiology experience. One ROI included the area surrounding the entire peritumoral T2 signal intensity abnormality, which included the tumor, and the other ROI surrounded only the enhancing part of the tumor. Subtracting the second ROI (enhancing tumor only) from the first ROI (entire abnormality) yielded the isolated peritumoral region. It was often difficult to delineate a clear margin between nonenhancing low-grade gliomas and the associated peritumoral T2 signal intensity abnormalities. In these cases, the signal intensity differences seen on conventional nonenhanced T1- and T2-weighted MR images were used to best delineate the peritumoral edema from the tumor.
Histograms of MD and FA values in the tumor and in the peritumoral edema were constructed. From these histograms, mean MD and FA values and SDs were calculated for each patient (11).
Tumor Infiltration Index
During the course of this study it became apparent that in patients with meningiomas and metastatic lesions, peritumoral MD and FA were approximately linearly related, whereas gliomas demonstrated lower FA. We hypothesized that in all intracranial neoplasms, increases in peritumoral MD are determined primarily on the basis of the free extracellular water content (12–14), with corresponding decreases in FA as the extracellular space enlarges. However, in gliomas, mechanisms such as white matter disruption or displacement (15–18) lead to further decreases in FA. Therefore, in the peritumoral edema, we calculated a tumor infiltration index (TII) as follows: TII = FAexp – FAobs, where FAexp is the FA that one would expect for the corresponding MD if the edema were not infiltrated with tumor and FAobs is the measured FA. The TII is multiplied by 103 because of its small scale.
Statistical Analyses
For comparisons between extra- and intraaxial tumors, between metastatic lesions and gliomas, and between high- and low-grade gliomas, statistical analysis consisted of unpaired Student t testing, with significance indicated by P < .05. For each comparison, the diffusion-tensor metrics were analyzed independently—that is, t testing to analyze MD values was performed separately from t testing to analyze FA values.
For the peritumoral measurements, a scatterplot of FA versus MD values was used to determine Pearson correlation coefficients (R). A coefficient greater than 0.80 indicated a strong correlation (19). A similar scatterplot was created for the intratumoral measurements. For the relationships with strong correlations, a least-squares linear regression fit was calculated and 95% CI bands were drawn by using the Working-Hotelling method.
In the equation used to calculate the TII, the FAexp is determined by using the linear regression fit for the cases of tumors with purely vasogenic edema—namely, the meningiomas and metastatic lesions. Ninety-five percent CIs were calculated by using the formula mean ± (1.96 · [SD/
(n)]), where n is the sample size. We also used the unpaired Student t test to compare the TII of vasogenic edema with the TII of presumed tumor-infiltrated edema.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The ADC is a rough measure of the magnitude of diffusion, whereas diffusion-tensor metrics such as MD and FA are measures of magnitude and directionality, respectively. In this study, we obtained histograms of MD and FA in and around four common brain tumor types—specifically, meningiomas, metastatic lesions, glioblastomas multiforme, and low-grade gliomas. The tumor types were grouped, comparisons were made, and the following results were observed:
1. Extraaxial tumors (ie, meningiomas) could not be distinguished from the other tumor types by using peritumoral diffusion-tensor metrics. However, conventional contrast-enhanced MR imaging is usually sufficient for identifying meningiomas. The fact that the peritumoral diffusion around meningiomas is no different from that around other brain tumors tells us that the extraaxial location of these tumors does not alter the diffusion characteristics of purely vasogenic edema. In other words, the layers of arachnoid mater and pia mater between the tumor and the brain parenchyma do not prevent the influx of peritumoral extracellular water. It has been postulated that this influx may be mediated by the secretion of substances such as prostaglandins and vascular endothelial growth factor (23,24) or caused by the direct hydrostatic exudation of water through arachnoidal disruptions and/or adhesions (1,2).
2. Among the intraaxial tumors, metastatic lesions demonstrated significantly higher peritumoral MD than did gliomas; these results support the findings of Krabbe et al (6). This distinction is especially useful when one encounters a common clinical dilemma: Solitary intraaxial metastases and high-grade gliomas often have similar conventional MR imaging appearances. However, these two lesions can now be differentiated by measuring the peritumoral MD. The marked peritumoral influx of water is probably secondary to greatly increased vascular permeability, which is presumably caused by the high expression of vascular endothelial growth factor, a trait that is common among metastatic lesions (25,26).
3. Regarding gliomas, the intratumoral MD in the low-grade gliomas was somewhat higher than that in the glioblastomas multiforme. Although this observation is consistent with the findings of Kono et al (5), who measured significantly higher ADCs in low-grade astrocytomas, we did not observe the same level of statistical significance. According to our study findings, the utility of diffusion-tensor metrics (both peritumoral and intratumoral) in the preoperative grading of gliomas is still limited.
Perhaps the key finding of this study is the ability to differentiate two types of peritumoral edema (ie, purely vasogenic vs tumor infiltrated) that are fundamentally different. Increases in ADC have been attributed to increased free extracellular water content, the primary component of peritumoral edema (22). In several studies (18,27,28), investigators have used diffusion-tensor imaging of high-grade gliomas to demonstrate that the increased extracellular water content leads to increased MD and decreased FA. In another study (29), investigators confirmed that both metastatic lesions and high-grade gliomas induce increased peritumoral MD and decreased peritumoral FA, but they also postulated that a second factor may contribute to the peritumoral FA changes around gliomas. That second factor was believed to be the infiltration of tumor cells into the adjacent brain parenchyma, which results in the white matter deviation and disorganization seen in tractography studies (15,16).
In the current study, we attempted to quantify the distinct contribution of tumor infiltration to FA changes by using a more robust method of data analysis. On the basis of a higher TII, presumed tumor-infiltrated edema, such as that adjacent to gliomas, can now be distinguished from vasogenic edema composed purely of extracellular water, such as the edema adjacent to meningiomas and metastases. In the future, the TII may be useful for clinical applications such as determining tumor resectability and the surgical approach, determining postoperative treatment options and assessing treatment responses (30), and determining likelihoods and/or rates of recurrence.
There were limitations to our study. Whenever ROIs are drawn, some degree of subjectivity is involved. To decrease this user dependence, a tumor was analyzed as a whole by using histograms rather than analyzed as small scattered ROIs; this en bloc analysis was similarly applied to the peritumoral edema. We faced another challenge in determining the margin between the tumor and the edema, which can be indistinct, especially in low-grade gliomas. Such poorly defined neoplasms without an identifiable peritumoral region were excluded. A by-product of the strict selection criteria was the relatively small sample size for each type of tumor.
Finally, histologic examination of the peritumoral edema to correlate the changes in the diffusion-tensor metrics with the physical measurements of water content and tumor cell concentration was not performed. Therefore, metrics such as the TII represent a presumed quantity of tumor infiltration and not a proved one. For true histologic correlation, future investigations will most likely require the use of animal models.
In summary, peritumoral diffusion-tensor metrics cannot be used to distinguish intraaxial from extraaxial lesions or to determine the grade of gliomas preoperatively. However, peritumoral MD values can be helpful in distinguishing solitary intraaxial metastatic lesions from gliomas. In addition, we introduce the TII, which enables one to distinguish presumed tumor-infiltrated edema from vasogenic edema composed purely of extracellular water. These capabilities of diffusion-tensor MR imaging are helpful in current diagnostic scenarios and conceivably will be useful for broader applications in the future.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, S.L.; study concepts, S.L., R.I.G., M.L.; study design, S.L., R.I.G., G.J.; literature research, S.L., D.Z.; clinical and experimental studies, S.L., D.A.; data acquisition, S.L., D.A.; data analysis/interpretation, S.L., G.J., M.L.; statistical analysis, S.L., G.J., M.L.; manuscript preparation, S.L.; manuscript definition of intellectual content, S.L., G.J., M.L., R.I.G.; manuscript editing, revision/review, and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C.-H. Toh, M. Castillo, A.M.-C. Wong, K.-C. Wei, H.-F. Wong, S.-H. Ng, and Y.-L. Wan Primary Cerebral Lymphoma and Glioblastoma Multiforme: Differences in Diffusion Characteristics Evaluated with Diffusion Tensor Imaging AJNR Am. J. Neuroradiol., March 1, 2008; 29(3): 471 - 475. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. G. P. Nucifora, R. Verma, S.-K. Lee, and E. R. Melhem Diffusion-Tensor MR Imaging and Tractography: Exploring Brain Microstructure and Connectivity Radiology, November 1, 2007; 245(2): 367 - 384. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Mikuni, T. Okada, R. Enatsu, Y. Miki, S.-i. Urayama, J. A Takahashi, K. Nozaki, H. Fukuyama, and N. Hashimoto Clinical significance of preoperative fibre-tracking to preserve the affected pyramidal tracts during resection of brain tumours in patients with preoperative motor weakness J. Neurol. Neurosurg. Psychiatry, July 1, 2007; 78(7): 716 - 721. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Okada, Y. Miki, K. Kikuta, N. Mikuni, S. Urayama, Y. Fushimi, A. Yamamoto, N. Mori, H. Fukuyama, N. Hashimoto, et al. Diffusion Tensor Fiber Tractography for Arteriovenous Malformations: Quantitative Analyses to Evaluate the Corticospinal Tract and Optic Radiation AJNR Am. J. Neuroradiol., June 1, 2007; 28(6): 1107 - 1113. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Stadlbauer, C. Nimsky, S. Gruber, E. Moser, T. Hammen, T. Engelhorn, M. Buchfelder, and O. Ganslandt Changes in Fiber Integrity, Diffusivity, and Metabolism of the Pyramidal Tract Adjacent to Gliomas: A Quantitative Diffusion Tensor Fiber Tracking and MR Spectroscopic Imaging Study AJNR Am. J. Neuroradiol., March 1, 2007; 28(3): 462 - 469. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N M DEsouza, S A Reinsberg, E D Scurr, J M Brewster, and G S Payne Magnetic resonance imaging in prostate cancer: the value of apparent diffusion coefficients for identifying malignant nodules Br. J. Radiol., February 1, 2007; 80(950): 90 - 95. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Inglese, S. Brown, G. Johnson, M. Law, E. Knopp, and O. Gonen Whole-Brain N-Acetylaspartate Spectroscopy and Diffusion Tensor Imaging in Patients with Newly Diagnosed Gliomas: A Preliminary Study AJNR Am. J. Neuroradiol., November 1, 2006; 27(10): 2137 - 2140. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Stadlbauer, O. Ganslandt, R. Buslei, T. Hammen, S. Gruber, E. Moser, M. Buchfelder, E. Salomonowitz, and C. Nimsky Gliomas: Histopathologic Evaluation of Changes in Directionality and Magnitude of Water Diffusion at Diffusion-Tensor MR Imaging Radiology, September 1, 2006; 240(3): 803 - 810. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Okada, N. Mikuni, Y. Miki, K.-i. Kikuta, S.-i. Urayama, T. Hanakawa, Y. Fushimi, A. Yamamoto, M. Kanagaki, H. Fukuyama, et al. Corticospinal Tract Localization: Integration of Diffusion-Tensor Tractography at 3-T MR Imaging with Intraoperative White Matter Stimulation Mapping--Preliminary Results Radiology, September 1, 2006; 240(3): 849 - 857. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G G Fan, Q L Deng, Z H Wu, and Q Y Guo Usefulness of diffusion/perfusion-weighted MRI in patients with non-enhancing supratentorial brain gliomas: a valuable tool to predict tumour grading? Br. J. Radiol., August 1, 2006; 79(944): 652 - 658. [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]
|
|
|
|
|
|
E. Goebell, S. Paustenbach, O. Vaeterlein, X.-Q. Ding, O. Heese, J. Fiehler, T. Kucinski, C. Hagel, M. Westphal, and H. Zeumer Low-Grade and Anaplastic Gliomas: Differences in Architecture Evaluated with Diffusion-Tensor MR Imaging Radiology, April 1, 2006; 239(1): 217 - 222. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Cha Update on Brain Tumor Imaging: From Anatomy to Physiology AJNR Am. J. Neuroradiol., March 1, 2006; 27(3): 475 - 487. [Full Text] [PDF]
|
|
|
|
|
|
A. S. Field Diffusion Tensor Imaging at the Crossroads: Fiber Tracking Meets Tissue Characterization in Brain Tumors AJNR Am. J. Neuroradiol., October 1, 2005; 26(9): 2168 - 2169. [Full Text] [PDF]
|
|
|
|
|
|
T. P.L. Roberts, F. Liu, A. Kassner, S. Mori, and A. Guha Fiber Density Index Correlates with Reduced Fractional Anisotropy in White Matter of Patients with Glioblastoma AJNR Am. J. Neuroradiol., October 1, 2005; 26(9): 2183 - 2186. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Kim, S.-K. Lee, J. D. Lee, Y. W. Kim, and D. I. Kim Decreased Fractional Anisotropy of Middle Cerebellar Peduncle in Crossed Cerebellar Diaschisis: Diffusion-Tensor Imaging-Positron-Emission Tomography Correlation Study AJNR Am. J. Neuroradiol., October 1, 2005; 26(9): 2224 - 2228. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K Tsuchiya, A Fujikawa, M Nakajima, and K Honya Differentiation between solitary brain metastasis and high-grade glioma by diffusion tensor imaging Br. J. Radiol., June 1, 2005; 78(930): 533 - 537. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Pauleit, F. Floeth, K. Hamacher, M. J. Riemenschneider, G. Reifenberger, H.-W. Muller, K. Zilles, H. H. Coenen, and K.-J. Langen O-(2-[18F]fluoroethyl)-L-tyrosine PET combined with MRI improves the diagnostic assessment of cerebral gliomas Brain, March 1, 2005; 128(3): 678 - 687. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Yousem, S.-K. Lee, D. I. Kim, and J. Kim Invited Commentary * Authors' Response RadioGraphics, January 1, 2005; 25(1): 66 - 68. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
