JOURNAL OF MAGNETIC RESONANCE IMAGING 28:566 –573 (2008)
Original Research
Diffusion Tensor Imaging and Fiber Tractography of C6 Rat Glioma Taketoshi Asanuma, DVM, PhD,1,2 Sabrina Doblas, MS,1 Yasvir A. Tesiram, PhD,1 Debra Saunders,1 Rebecca Cranford, BS,1 Jamie Pearson, BS,1 Andrew Abbott, MS,1 Nataliya Smith, PhD,1 and Rheal A. Towner, PhD1* Purpose: To apply diffusion tensor images using 30 noncollinear directions for diffusion-weighted gradient schemes to characterize diffusion tensor imaging (DTI) features associated with C6 glioma-bearing rat brains, and ideally visualize fiber tractography datasets. Materials and Methods: Fiber tractographies of normal male Fischer 344 rat brains were constructed from DTI datasets acquired with a 30 noncollinear diffusion gradient scheme. Cultured C6 cell were intracranially injected into the cortex of male Fischer 344 rats. The time course of the tumor growth was monitored with DTI and fiber tractography using diffusion-weighting gradients in 30 noncollinear directions. Results: Fiber tractographies through the corpus callosum (CC) were easily visualized with the 30-direction gradient scheme, and the fiber trajectories of the motor cortex and striatum were well represented in normal rats. Fiber tractography indicated that the neuronal fibers of the CC were compressed or disappeared by growing C6 glioma, which affected surrounding brain tissue. Conclusion: We have demonstrated in this study that fiber tractography with the 30 noncollinear diffusion gradient scheme method can be used to help provide a better understanding regarding the influence of a tumor on the surrounding regions of normal brain tissue in vivo. Key Words: MRI; diffusion tensor; eigenvector; fiber tractography; neuronal fibers; glioma; C6 gliomas; rat. J. Magn. Reson. Imaging 2008;28:566 –573. © 2008 Wiley-Liss, Inc.
1 Small Animal MRI Facility, Oklahoma Medical Research Foundation, Oklahoma City, Oklahoma. 2 Laboratory of Radiation Biology, Graduate School of Veterinary Medicine, Hokkaido University, Sapporo, Japan. Contract grant sponsor: Oklahoma Center for the Advancement of Science and Technology; Contract grant number: fMRI-002. *Address reprint requests to: R.A.T., Small Animal Magnetic Resonance Imaging Facility, Free Radical Biology and Aging Research Program, Oklahoma Medical Research Foundation, 825 N.E. 13th St. (MS 60), Oklahoma City, OK 73104. E-mail:
[email protected] Received November 13, 2007; Accepted May 14, 2008. DOI 10.1002/jmri.21473 Published online in Wiley InterScience (www.interscience.wiley.com).
© 2008 Wiley-Liss, Inc.
DIFFUSION-WEIGHTED IMAGING (DWI), provides tissue contrast that depends on the diffusion of water molecules and has fast become an important clinical tool for early detection of brain injuries, allowing visualization of ischemic regions within normal tissue. Since Moseley et al (1) reported that DWI could detect lesions associated with middle cerebral artery occlusion in a hyperacute phase DWI has been used clinically to evaluate human brain injuries. DWI has been previously used in the assessment of potential therapeutic agents against malonate-induced ischemic injury (2). Recently, the diffusion tensor imaging (DTI) technique, which is an advanced form of diffusion imaging, has permitted the detailed visualization of white matter structural integrity and connectivity (3). Diffusion anisotropy reports on abnormalities in neuronal diseases such as stroke (3), multiple sclerosis (4), epilepsy (5), schizophrenia (6), and gliomas (7–13). A number of diffusion-weighted studies of rodent glioma models have been reported to reveal unique DTI features in apparent diffusion coefficients (ADC) or fractional anisotropy (FA) (14,15). The fiber tractography technique is novel, and is a fairly well-accepted method for noninvasive visualization of neuronal fiber bundles in human brain. This is achieved by calculating the principal diffusion direction and magnitude (principal eigenvector) to generate trajectory maps (16 –18). Fiber tractography has been successfully applied to various human neuronal disorders such as ischemic stroke (19), amyotrophic lateral sclerosis (20), and gliomas (21,22). Some researchers have also reported the importance of using fiber tractography in isolated rat spinal cords (23) and the hippocampus (24). From two-dimensional serial multislice DTI datasets or eigenvector maps it is not easy to understand the degree of integrity and connectivity of the neuronal fibers that surround a brain tumor, and therefore fiber tractography can be used to reflect the directions and connectivity of neuronal fiber bundles. Even though fiber tractography may not represent an accurate neuronal fiber image, this method may be used to help understand the influence of a glioma on surrounding tissue neuronal fiber tracts. In this study we used fiber tractography to visualize the trajectories of neuronal fiber tracts in rats in vivo to assess the impact of gliomas on normal tissue. Serial
566
Fiber Tractography of Rat Gliomas
DTI with a high signal-to-noise ratio (SNR) and the use of robust three eigenvector components were used to create fiber tractographs of rat brains in vivo. From our pilot studies using 6- and 12-direction gradient schemes (14,23–26), the fiber trajectories of the corpus callosum (CC) were easily visualized; however, the fiber trajectories of the motor cortex or striatum were not well represented and appeared distorted in normal rat brain regions. Based on these pilot studies, in order to obtain optimized fiber trajectories in glioma-bearing rat brains DTIs were collected with a 30 noncollinear directions diffusion-weighted gradient scheme in this study. In addition to obtaining optimized fiber tractographs in rat brain, this study also reports on observations regarding the DTI properties of C6 rat gliomas in vivo. MATERIALS AND METHODS Glioma Cell Injection The animal study was conducted in accordance with the guidelines and approval of the Institutional Animal Care and Use Committee (IACUC). The C6 glioma-bearing rat model used was a modification of the model described by Kobayashi et al (27). Three-month-old male Fischer 344 rats (250 –350 g) were fed a cholinedeficient diet from 5 days before C6 cell implantation until the time of euthanasia. The numbers of normal and glioma-bearing rats used were 3 and 6, respectively. It has been found that choline deficiency helps glioma cell seeding and growth (27). Each animal was anesthetized (3% isoflurane at 2.5 L/min oxygen) and placed in a stereotaxic device (Stoelting, Wood Dale, IL). The skin of the head was cleaned and an incision was made to expose the skull to identify the bregma and lateral positions. A 1-mm hole was drilled (using a dentist drill) through the skull 2 mm anterior and 2 mm lateral to the bregma, in the right-hand side of the skull. Rat glioma C6 cells (106 in 10 L cell culture media: Dulbecco’s modified Eagle’s medium [DMEM]; GibcoInvitrogen, Carlsbad, CA), maintained in our laboratory and originally obtained from ATCC (American Type Culture Collection, Manassas, VA) in an ultra-low gelling temperature agarose (1%; Sigma-Aldrich, St. Louis, MO), were injected in the cortex at a 3-mm depth at a rate of 2 L/min. A waiting time of 2 minutes was implemented following injection and bone wax was put in the burr-hole to prevent any cell suspension reflux. The incision was sutured (4-0 suture; Oasis, Mettawa, IL) and covered with surgical glue (LiquiVet, Oasis). The surgery was performed under sterile conditions. The syringes (Hamilton, 25 L, 25-G needle; Hamilton, Reno, NV) containing the C6 cells were kept at 37°C until the moment of injection. MRI Experiments MRI experiments were performed on a Bruker Biospec 7 T/30 cm horizontal bore magnet (Bruker BioSpin MRI, Germany). Transmitter and receiver radiofrequency (RF) coils used were a quadrature volume coil and a surface coil, respectively. Rats were anesthetized (2.5% isoflurane at 2.5 L/min oxygen) and wrapped with a hot-water-circulated blanket that kept them at 37.5 ⫾
567
1.0°C throughout the experimental protocol. MRI experiments were done after 1, 2, 3, and 4 weeks after cell implantation. These times were based on the average survival time of C6 glioma-bearing rats, and to follow glioma growth on a weekly basis. Proton densityweighted (PDw) and T2-weighted (T2w) images were acquired using a 2D multislice double-echo RARE imaging pulse sequence (repetition time [TR] ⫽ 3 seconds, RARE factor ⫽ 6, echo times [TEs] ⫽ 32.6 and 96.18 msec, field of view ⫽ 2.56 ⫻ 2.56 cm2, 256 ⫻ 256 matrix, and 4 averages). Respiratory gating was used to trigger acquisition of the phase-encoding steps in the imaging sequence for both the RARE imaging and the DTI sequences. Diffusion tensor data were acquired with a spin echo multishot echo planar imaging (EPI) pulse sequence. The receiver bandwidth was 200 kHz. The diffusion-weighting gradient schemes with 30 noncollinear directions were used based on the Jones_30 directions method (25,26). For all studies the acquisition parameters were TR ⫽ 3 seconds, TE ⫽ 50 msec, 21 interleaved and contiguous slices with a slice thickness of 1 mm, a field of view of 2.56 ⫻ 2.56 cm2, and a b-factor of 1000 sec/mm2. Total acquisition time for all the methods was ⬇70 minutes. All EPI data were acquired with a four-shot EPI sequence, eight excitations, 128 ⫻ 128 matrix, and zero-filled in k-space to construct a 256 ⫻ 256 image matrix. No residual motion artifacts or ghosts were found in the images in either the animal investigations or phantom studies done previously. The DTI data acquisition and reconstruction were performed using ParaVision software (v. 4.0; Bruker Biospin MRI). Diffusion tensor elements at each voxel were calculated with ParaVision software to obtain three eigenvalues, three eigenvectors, an FA map, a Trace map, and an S0 image. The principal eigenvector maps were created by using DTI Visualization software (Bruker BioSpin MRI). The components of the principal eigenvectors were displayed as RGB colors: red for left– right, green for cranial– caudal (in– out, perpendicular to the slice plane), blue for dorsal–ventral (top– down) directions. By using a built-in macro application from the ParaVision software, time course changes in FA and ADC values taken from region of interests (ROIs) in the following tissue areas were obtained: glioma, ipsilateral CC, contralateral CC, ipsilateral cerebral cortex (cx), and contralateral cx. ROIs from each time course were manually drawn on FA maps in reference to the principal eigenvector (1) map, and then the same position and size of each ROI was automatically overlapped onto Trace maps. Finally, the macro application calculated the values of FA and ADC for each ROI. Data are presented as the mean ⫾ SE (standard error). The variance ratio was estimated by using an F-test, and differences in the means from each group were determined by a Student’s t-test or Welch’s t-test. The minimum level of statistical significance was set at P ⬍ 0.05. Fiber Tractography All 2D multislice DTI imaging datasets were converted to DICOM format for these analyses. 3D fiber tract maps were created by VOLUME-ONE and dTV software (Image Computing and Analysis Laboratory, Depart-
568
ment of Radiology, University of Tokyo Hospital, Tokyo, Japan), and are available at http://www.volume-one.org and http://www.ut-radiology.umin.jp/people/ masutani/dTV.htm, respectively (18 –21). Fiber trajectories such as neuronal fiber images of the CC and the peripheral neuronal fibers of the CC were visualized at four ROIs, which were axial slice positions 3.05, 2.00, 0.95, and – 0.1 mm anterior to the bregma (28). In determining these tracts we set the threshold of FA to 0.2 (29), with the resulting FA values of the CC and cortex being 0.30 ⫾ 0.03 and 0.18 ⫾ 0.02, respectively. Tractographies were displayed by using a directionalcoloring code: red for left–right, blue for cranial– caudal, and green for dorsal–ventral directions. Gliomas were visualized on the 3D fiber tract maps. By using a Freehand ROI editor in the dTV.II application, glioma regions were obtained from signal intensities of the glioma in each T2w image slices, and then each glioma region was compiled into a 3D representation and overlaid onto respective fiber trajectories. Furthermore, 3Drendered images were separately reconstructed with the T2w-EPI data from contiguous slices in the normal rat by using VOLUME-ONE software, and illustrated 3D in the same direction as that observed from each fiber tractography dataset. Histological Examination After all MRI experiments were completed, rats were euthanized by carbon dioxide asphyxiation, followed by transcardial perfusion with phosphate-buffered saline, and then fixed in 10% buffered formalin. Whole brains were removed from the skull and immersed in fixative (formalin). Seven-m thick paraffin-fixed coronal sections were stained with hematoxylin-eosin (H&E) or Luxol Fast Blue eosin (LFBE) for the detection of the myelin sheath. RESULTS Thirty Noncollinear Diffusion-Weighted Gradient Schemes in Normal Rat Brain Figure 1 shows the representative results of the colored principal eigenvector maps obtained from 30 noncollinear directions in one normal rat. This was particularly illustrated within the neuronal fibers in the glioma regions, which were accurately visualized (1–3 mm anterior to the bregma; B2.5, B1.5, and B0.5 in Fig. 1). In addition, the structure of the olfactory bulb, hippocampus, and cerebellum can be visualized in excellent detail (5.5 mm anterior and 10.5 mm posterior to the bregma; B5.5 and B-10.5 in Fig. 1, respectively). To evaluate the directions of the principal eigenvector, the principal eigenvector maps (Fig. 2a) were compared with histological sections stained specifically for the myelin sheath with LFBE, and these are presented in Fig. 2. By using Adobe Photoshop (San Jose, CA), a representative colored principal eigenvector map was divided into three images based on red, green, and blue (RGB) for a better understanding of the long axis of the diffusion ellipsoid (Fig. 2b– d, respectively). These separate images corresponded to the directions of the long axis of the diffusion ellipsoid; left–right, cranial– caudal,
Asanuma et al.
Figure 1. Colored principal eigenvector maps applied in 30 noncollinear directions in normal rat brain. B5.5, B2.5, B1.5, B0.5, B-0.5, B-3.5, B-5.5, and B-10.5 are the slice positions, with plus and minus being either anterior or posterior to the bregma, respectively; the number (2.5, 5.5) refers to the distance (mm) from the bregma. Eigenvector components were mapped to the following color schemes: red is left–right, blue is dorsal–ventral (top– down), and green is cranial– caudal (perpendicular to the slice plane). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley. com.]
and dorsal–ventral directions, respectively. White squares (e–l) in Fig. 2a correspond to histological regions in Fig. 2e–2l. In Fig. 2e the direction of the CC was from left–right, and the eigenvector map is well visualized in this direction (Fig. 2a,b). The neuronal fibers in the cx near the CC were in various directions; skewed (right–left and dorsal–ventral) in the cingulate cortex (cg), dorsal–ventral in the motor cortex (mc), and right– left in the somatosensory cortex (sc) (Fig. 2f–i, respectively). The principal eigenvector map was found to well reflect these directions (Fig. 2a– d). The directions of the anterior commissure anterior part (aca) and the lateral olfactory tract (lot), which are well-defined bundles of nerve fibers, were mainly perpendicular (cranial– caudal direction) to the slice plane, and were slightly in right–left or dorsal–ventral directions. In the principal eigenvector map the directions were similarly dorsal– ventral rather than in cranial– caudal directions (Fig. 2a,c,d). Furthermore, in the striatum (Str) the directions of the neuronal bundles were dorsal–ventral in Fig. 2l, with green colors mainly observed in Fig. 2a (in addition, Fig. 2c,d, respectively). These results indicate that the directions of the CC were accurately obtained with the 30 diffusion gradient datasets. Figure 3 shows representative fiber trajectories of the CC and the cg which are clearly distinguishable by the 30 noncollinear directions method. Having established that the 30 noncollinear diffusion-weighting gradient scheme method was the most suitable for the visualization of rat brain neuronal fibers, all the DTI datasets from the C6 glioma cell injected rats were acquired using the 30-direction gradient scheme method. Time Course Change of C6 Glioma Growth in Living Rat Brain Figure 4 shows typical PDw and T2w images, Trace, FA, and principal eigenvector maps at 1, 2, 3, and 4 weeks after C6 glioma cell implantation. C6 glioma growth was
Fiber Tractography of Rat Gliomas
569
Figure 2. a: Representative colored principal eigenvector map at B0.5 (0.5 mm anterior to the bregma); cg: cingulate cortex, mc: motor cortex, sc: somatosensory cortex, CC: corpus callosum, Str: striatum, aca: anterior commissure anterior part, lot: lateral olfactory tract. Eigenvector components are the same as in Fig. 1. b– d: The representative colored principal eigenvector map was divided into three images based on red, green, and blue, corresponding to the directions of the long axis of diffusion ellipsoid; left–right, cranial– caudal, and dorsal– ventral directions, respectively. White squares (e–l) in Fig. 2a correspond to histological images in Fig. 2e–l (LFBE-stained sections). Scale bars ⫽ 250, 100, 50 m in e, f–i,l, and j– k, respectively. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]
detected in all intracerebrally injected rats. At 1 and 2 weeks after cell implantation, tumor regions are easily recognized and are characterized by a demarcation between the glioma and the surrounding regions in both PDw and T2w images. At 3 and 4 weeks postcell-im-
Figure 3. Representative fiber tractographies obtained from normal rat brains with 30 directions methods datasets. At the bottom of each column 3D-rendered images are illustrated in the same directions as those observed for each fiber tractography dataset. The directions of the fiber trajectories of the neuronal fibers were colored as follows: red is left–right, green is dorsal–ventral, and blue is cranial– caudal (perpendicular to the slice plane). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]
plantation, however, the border between the glioma and the surrounding regions were not well delineated in either PDw or T2w images. Instead, a diffuse region of high signal intensity results as the glioma grows. Obvi-
Figure 4. Representative PDw, and T2w images, Trace, FA, and principal eigenvector (1) maps at 1, 2, 3, and 4 weeks after C6 cell implantation. Eigenvector components were mapped to the following color schemes: red is left–right, blue is dorsal–ventral (top– down), and green is cranial– caudal (perpendicular to the slice plane). Arrows indicate the location of the C6 glioma. Closed arrowheads and open arrowheads indicate the regions of the ipsilateral and the contralateral corpus callosum, respectively, in FA and principle eigenvector maps. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]
570
ously, establishing the precise size and shape of the tumor is somewhat difficult from these images but the calculation of other parameters from the diffusion tensor provides additional metrics not easily derived from PDw, T2w images and Trace maps alone (arrows in Fig. 4 PDw, T2w, and Trace maps). Fractional anisotropy and principal eigenvector maps are additional metrics that provide some unique results on the nature of C6 glioma growth and its effect on surrounding tissue. From the FA maps, the regions of glioma were found to have low values (arrows in Fig. 4 FA maps). FA and principal eigenvector maps of Fig. 4 show morphological changes in the CC (arrowheads and open arrow indicate the regions of the ipsilateral and the contralateral CC, respectively). For the ipsilateral CC the fiber directional positions seemed to shift toward the ventral from the dorsal of the brain at 2 weeks after cell implantation, and then at 4 weeks the FA values of the growing glioma decreased and disappeared (arrowheads in the FA and principal eigenvector maps). For the contralateral CC the fiber directional positions in the central regions shifted toward the temporal region as a result of the growing glioma (open arrowheads in FA and principal eigenvector maps). In other words, the shape of the CC was distorted, caused by compression of the surrounding tissue during glioma growth. In the principal eigenvector maps, the directions of the CC changed with time over the course of the experimental period, ie, these directions were initially perpendicular (in the cranial– caudal direction) to the slice plane at 1 and 2 weeks, and then at 3 weeks, yellowcolored regions were observed in a part of the ipsilateral CC. In the RGB color-encoding images, yellow is expressed as the mixing of red and green. This means that the cranial– caudal direction (green) of the neuronal fibers were forced into a right–left direction (red) from the glioma growth. Figure 5 show the time course changes in FA and ADC values within the glioma, ipsilateral CC, contralateral CC, ipsilateral cx, and contralateral cx regions (mean ⫾ SE, n ⫽ 6). Figure 5a shows representative ROIs for FA and Trace maps in reference to the colored principal eigenvector map at 2 weeks. FA values of the C6 glioma were found to have low values of 0.11– 0.24 during the experimental time periods (closed circle in Fig. 5b, as well as Fig. 4). Our results show that a growing glioma has no specific direction and that the glioma itself has isotropic diffusion. For the ipsilateral CC, FA values increased at 3 weeks (0.38 ⫾ 0.02), and decreased at 4 weeks (0.29 ⫾ 0.04) (closed square in Fig. 5b). For the contralateral CC, FA values gradually increased at 4 weeks (from 0.26 ⫾ 0.01 to 0.34 ⫾ 0.04, open square in Fig. 5b). These results also indicate that the shape of the CC was distorted due to compression of the surrounding tissue from the glioma growth. In fact, at 4 weeks ADC values of the ipsilateral CC were higher than they were initially at 1 week (0.70 ⫾ 0.02 and 0.57 ⫾ 0.03 ⫻ 10⫺3 mm2/s, respectively, in Fig. 5c). These high ADC values might reflect degradation or disappearance of CC as a result of the glioma. For both ipsilateral and contralateral cerebral cortexes, FA values gradually increased at 4 weeks (closed and open
Asanuma et al.
Figure 5. a: Time course changes in (b) FA and (c) ADC values taken from ROIs in the following regions: glioma (closed circle), the ipsilateral CC (closed square), the contralateral CC (open square), the ipsilateral cx (closed triangle), and the contralateral cx (open triangle). Statistical significance in the ipsilateral region (closed square) was obtained (*P ⬍ 0.05) when compared to the contralateral CC (open square).
triangles in Fig. 5b, respectively), and ADC values gradually decreased at 3 weeks; however, both ADC values quickly increased at 4 weeks (closed and open triangles in Fig. 5c, respectively). These results imply that compression of the surrounding tissue due to glioma growth lasted for 3 weeks, and then degradation occurred at 4 weeks. Four weeks is the average survival time of C6 glioma-bearing rats, and therefore the incensement of the FA and ADC values might reflect a collapse in function of the central nervous system. Fiber Tractography of C6 Glioma-Bearing Rat Brain Figure 6 shows representative fiber tractography datasets for a C6 glioma-bearing rat brain at various times. Glioma regions were visualized as brown-colored voxels. It was easy to visualize the regions where the fiber trajectories had been compressed and transformed by glioma growth. At 1 week after cell implantation, although the cell injection-trace was observed in PDw and T2w images (Fig. 4), the fiber trajectories of the ipsilateral CC and fibers of the motor and cingulate
Fiber Tractography of Rat Gliomas
571
open arrowheads). The contralateral fiber trajectories in the motor and cingulate cortex (blue-colored fibers) were pushed aside by the glioma, but the connectivity of these fiber tracts remained intact (Fig. 6, white arrows). Histological Examination of C6 Glioma Brain at 4 Weeks After Cell Implantation
Figure 6. Representative fiber tractography dataset from one rat (n ⫽ 6) with a C6 glioma. At the bottom of each column 3D-rendered images illustrate the same direction observed for each fiber tractography dataset. Glioma regions are visualized as brown-colored voxel overlays. The directions of the fibers were colored as follows: red is left–right, green is dorsal–ventral (top– down), and blue is cranial– caudal (perpendicular to the slice plane). Small white closed arrowheads depict a shifting of the contralateral CC toward the left side. Open arrowheads indicate that the contralateral CC is not visible. White arrows indicate that part of the fiber trajectory of the ipsilateral CC is shifted toward the ipsilateral (left) side.
cortex were almost intact and similar to that of the contralateral side. At 2 weeks, although these fiber trajectories had been altered by the glioma, the connectivity of these fibers seemed to be intact (Fig. 6, closed arrowheads). The fiber trajectories had not been pushed aside on the cerebral cortex side. At 3 weeks after cell implantation the fiber tracts had been pushed aside on the ventral side or contralateral side by the glioma growth (Fig. 6, closed arrowheads). In the principal eigenvector maps at 3 weeks (Fig. 4), yellow-colored regions were observed within the ipsilateral CC. We propose that the direction of the fiber trajectories were forced to be altered as a result of the glioma growth. However, in the contralateral side (left), the shape of the fiber tracts did not change, and this connectivity seemed to be intact. At 4 weeks the glioma occupied the ipsilateral side, and the ipsilateral fiber trajectories had disappeared around the glioma (Fig. 6,
Figure 7a shows an H&E-stained section of the brain at 4 weeks after cell implantation. In the border between the glioma and normal tissues the shapes of the neuronal fibers and cells were found to be elongated ellipses, whereas in the normal tissues the shapes were oval. An ischemic region (is) is depicted between the normal and glioma tissue regions. Interestingly, the directions of the long axis of these elongated neuronal fibers or cells were parallel to the boundary side of the glioma. These results indicate that the glioma had compressed the normal tissues. In addition, vacuolation surrounding the neuronal cells was observed in the border region. This finding appears to reflect the presence of an ischemic region induced by the pressure of the glioma growth. Figure 7b,c shows the LFBE-stained sections of the ipsilateral and contralateral sides of the brain at 4 weeks after cell implantation, respectively. In the ipsilateral striatum, no neuronal fibers were observed (Fig. 7b). In the contralateral side, although the shape of the CC was deformed, its presence was still easily confirmed by histology. In addition, the neuronal fibers were clearly observed in the contralateral striatum. Within the necrotic center of the glioma (Fig. 7d) there was widespread vacuolation and colliquative necrosis (Fig. 7d, arrowheads and arrows, respectively). Within the glioma there was also demylelination depicted (Fig. 7f, arrows) in comparison to the nuclei of normal neural cells (Fig. 7e, open arrowheads). DISCUSSION High SNR and high-resolution MRI signals from small laboratory animals such as rats are easily achieved with high magnetic field strengths (30). Since DWI typically has low SNR, it is important to try and minimize the noise during data acquisition in order to acquire accurate FA, Trace, and fiber-imaging datasets in reasonable timeframes. It is therefore necessary to discuss the most suitable sets of diffusion-weighting gradient schemes that are required to visualize the neuronal fiber tracts of glioma-bearing rat brains. With the 30 direction method (25), the fiber trajectories of the CC and the neuronal fibers in the motor and the cingulate cortex (Fig. 3) could be clearly visualized (Figs. 1, 4). Skare et al (26) reported that it was necessary to increase the number of diffusion-weighting gradient schemes, as this greatly reduces the standard deviation of the FA and its dependence on the orientations of the diffusion ellipsoid. Taken in conjunction with the LFBE-stained sections, it can be said that the 30 directions method provides an accurate representation of the directions of the fiber trajectories (Figs. 2, 3). By using many diffusion-weighting gradient schemes such as 42, 162, or 642 directions, this would provide higher
572
Asanuma et al.
Figure 7. Hematoxylin-eosin (H&E) and Luxol First Blue-eosin (LFBE)-stained sections at 4 weeks after C6 cell implantation. a: An ischemic region (is) is observed between the glioma (g) and normal cerebral (n) tissues. nc, necrotic center in the glioma. b,c: LFBE-stained sections of the ipsilateral and contralateral sides of the brain at 4 weeks after C6 cell implantation, respectively. g, glioma; str, striatum; aca, anterior commissure anterior region; cc, corpus callosum. (d) Necrotic center (nc) of a glioma (g). Closed arrowheads and arrows depict widespread vacuolation and colliquative necrosis, respectively. e: Neuronal fiber bundles (open arrowheads) in a normal brain region. f: Demyelination of neuronal fibers (arrows) in an ischemic region of a glioma (g).
SNR and more accurate diffusion ellipsoids (31); however, this would also result in long sequence times, which may be unsuitable. Figure 5 shows a time course change of Trace, where the ADC values of both the ipsilateral and contralateral CCs decreased gradually from 1 to 3 weeks, and then the ADC value of ipsilateral CC was found to increase at 4 weeks. This result and histological examination (Fig. 7c) indicated that the ipsilateral CC disappeared at 4 weeks. However, it is difficult to determine what occurred in the ipsilateral CC from 1 to 3 weeks with ADC data only, since there was no difference in ADC values between the ipsilateral and contralateral CCs that was observed during this period. In the time course graph for FA (Fig. 5b), FA values of the ipsilateral CC were found to increase at 2 weeks, whereas the values at 4 weeks decreased. Also noted, the FA values of the contralateral CC gradually increased at 2 weeks. Taken together with the results shown in Fig. 6 (fiber tractography data), these results indicate that the FA values correlate with a compression due to the glioma growth, and additionally that the FA values indicated that ischemia may be associated with the disappearance of neuronal fibers (Fig. 7a,f). Similar histological changes resembling what was observed in our study, between the glioma and normal tissues, were previously reported (32) in a study on ischemic necrosis-induced edema present in surrounding normal tissues, with a related loss of the myelin sheath. LFB-eosin-stained histology revealed that the neuronal fibers disappeared in these regions (Fig. 7c), and that fiber tractography also revealed the disappearance of neuronal fibers due to the fact that no fiber trajectory was observed surrounding the glioma at 4 weeks (open arrowheads in Fig. 6 at 4 weeks). We hypothesize that the dramatic change in FA values (high FA values at 3 weeks and then a decrease at 4 weeks) is due to the disappearance of the neuronal fibers in the glioma. In conclusion, the method of using 30 noncollinear directions was found to be suitable to visualize neuro-
nal fibers in rats. The fiber tractography method can also be useful to help better understand the influence of peripheral tissues surrounding a tumor. These results indicate that normal brain tissue can be subjected to a sudden compression resulting from C6 glioma growth. Moreover, DTI was able to provide a more detailed characterization of diffusion within a living rat brain with a developing glioma. Although fiber tractography is not an accurate representation of a neuronal fiber image, this method does reflect the directions and connectivity of neuronal fibers, and can provide additional information regarding the influence of glioma growth on surrounding tissues. ACKNOWLEDGMENT The authors thank Ms. Virginia Oblander (Comparative Medicine (LARC), OMRF) for assisting in the histological examinations. REFERENCES 1. Moseley ME, Butts K, Yenari MA, Marks M, de Crespigny A. Clinical aspects of DWI. NMR Biomed 1995;8:387–396. 2. Asanuma T, Ishibashi H, Konno A, Kon Y, Inanami O, Kuwabara M. Assessment of neuroprotective ability of a spin trap, ␣-phenyl-Ntert-butylnitrone, against malonate-induced ischemic injury of rat brain by apparent water diffusion coefficient mapping. Neurosci Lett 2002;329:281–284. 3. Mukherjee P, Bahn MM, McKinstry RC, et al. Differences between gray matter and white matter water diffusion in stroke: diffusiontensor MR imaging in 12 patients. Radiology 2000;215:211–220. 4. Guo AC, MacFall JR, Provenzale JM. Multiple sclerosis: diffusion tensor MR imaging for evaluation of normal-appearing white matter. Radiology 2002;222:729 –736. 5. Rugg-Gunn FJ, Eriksson SH, Symms MR, Barker GJ, Duncan JS. Diffusion tensor imaging of cryptogenic and acquired partial epilepsies. Brain 2001;124:627– 636. 6. Kubicki M, Westin CF, Maier SE, et al. Uncinate fasciculus findings in schizophrenia: a magnetic resonance diffusion tensor imaging study. Am J Psychiatry 2002;159:813– 820. 7. Beauchesne PD, Pedeux RM, Bonmartin A, et al. Intracerebral C6 glioma model in female hairless rats: assessment by using MRI and follow-up of irradiation. Anticancer Res 2003;23:3755–3760.
Fiber Tractography of Rat Gliomas 8. Lu S, Ahn D, Johnson G, Cha S. Peritumoral diffusion tensor imaging of high-grade gliomas and metastatic brain tumors. AJNR Am J Neuroradiol 2003;24:937–941. 9. Beppu T, Inoue T, Shibata Y, et al. Measurement of fractional anisotropy using diffusion tensor MRI in supratentorial astrocytic tumors. J Neurooncol 2003;63:109 –116. 10. Price SJ, Pena A, Burnet NG, et al. Tissue signature characterisation of diffusion tensor abnormalities in cerebral gliomas. Eur Radiol 2004;14:1909 –1917. 11. Provenzale JM, McGraw P, Mhatre P, Guo AC, Delong D. Peritumoral brain regions in gliomas and meningiomas: investigation with isotropic diffusion-weighted MR imaging and diffusion-tensor MR imaging. Radiology 2004;232:451– 460. 12. Stieltjes B, Schluter M, Didinger B, et al. Diffusion tensor imaging in primary brain tumors: reproducible quantitative analysis of corpus callosum infiltration and contralateral involvement using a probabilistic mixture model. Neuroimage 2006;31:531–542. 13. Tsuchiya K, Fujikawa A, Nakajima M, Honya K. Differentiation between solitary brain metastasis and high-grade glioma by diffusion tensor imaging. Br J Radiol 2005;78:533–537. 14. Zhang J, van Zijl PC, Laterra J, Salhotra A, Lal B, Mori S, Zhou J, Unique patterns of diffusion directionality in rat brain tumors revealed by high-resolution diffusion tensor MRI. Magn Reson Med 2007;58:454 – 462. 15. Bennett KM, Hyde JS, Rand SD, et al. Intravoxel distribution of DWI decay rates reveals C6 glioma invasion in rat brain. Magn Reson Med 2004;52:994 –1004. 16. Melhem ER, Mori S, Mukundan G, Kraut MA, Pomper MG, van Zijl PC. Diffusion tensor MR imaging of the brain and white matter tractography. AJR Am J Roentgenol 2002;178:3–16. 17. Basser PJ, Pajevic S, Pierpaoli C, Duda J, Aldroubi A. In vivo fiber tractography using DT-MRI data. Magn Reson Med 2000;44:625– 632. 18. Masutani Y, Aoki S, Abe O, Hayashi N, Otomo K. MR diffusion tensor imaging: recent advance and new techniques for diffusion tensor visualization. Eur J Radiol 2003;46:53– 66. 19. Kunimatsu A, Aoki S, Masutani Y, Abe O, Mori H, Ohtomo K. Three-dimensional white matter tractography by diffusion tensor imaging in ischaemic stroke involving the corticospinal tract. Neuroradiology 2003;45:532–535. 20. Aoki S, Iwata NK, Masutani Y, et al. Quantitative evaluation of the pyramidal tract segmented by diffusion tensor tractography: feasi-
573
21.
22.
23.
24.
25.
26.
27. 28. 29.
30.
31.
32.
bility study in patients with amyotrophic lateral sclerosis. Radiat Med 2005;23:195–199. Akai H, Mori H, Aoki S, et al. Diffusion tensor tractography of gliomatosis cerebri: fiber tracking through the tumor. J Comput Assist Tomogr 2005;29:127–129. Mori S, Frederiksen K, van Zijl PC, et al. Brain white matter anatomy of tumor patients evaluated with diffusion tensor imaging. Ann Neurol 2002;51:377–380. Schwartz ED, Duda J, Shumsky JS, Cooper ET, Gee J. Spinal cord diffusion tensor imaging and fiber tracking can identify white matter tract disruption and glial scar orientation following lateral funiculotomy. J Neurotrauma 2005;22:1388 –1398. Shepherd TM, Ozarslan E, King MA, Mareci TH, Blackband SJ. Structural insights from high-resolution diffusion tensor imaging and tractography of the isolated rat hippocampus. Neuroimage 2006;32:1499 –1509. Jones DK, Horsfield MA, Simmons A. Optimal strategies for measuring diffusion in anisotropic systems by magnetic resonance imaging. Magn Reson Med 1999;42:515–525. Skare S, Hedehus M, Moseley ME, Li TQ. Condition number as a measure of noise performance of diffusion tensor data acquisition schemes with MRI. J Magn Reson 2000;147:340 –352. Kobayashi N, Allen N, Clendenon NR, Ko LW. An improved rat brain-tumor model. J Neurosurg 1980;53:808 – 815. Paxinos G, Watson C. The rat brain in stereotaxic coordinates. San Diego: Academic Press; 1998. Kunimatsu A, Aoki S, Masutani Y, et al. The optimal trackability threshold of fractional anisotropy for diffusion tensor tractography of the corticospinal tract. Magn Reson Med Sci 2004;3:11–17. Asanuma T, Hirano Y, Yamamoto K, Kon Y, Shimokawa S, Kuwabara M. MR imaging of hepatic injury in the LEC rat under a high magnetic field (7.05 T). J Vet Med Sci 1999;61:239 –244. Le Bihan D, Mangin JF, Poupon C, et al. Diffusion tensor imaging: concepts and applications. J Magn Reson Imaging 2001;13:534 – 546. Asanuma T, Inanami O, Tabu K, Waki K, Kon Y, Kuwabara M. Protection against malonate-induced ischemic brain injury in rat by a cell-permeable peptidic c-Jun N-terminal kinase inhibitor, (L)-HIV-TAT48 –57-PP-JBD20, observed by the apparent diffusion coefficient mapping magnetic resonance imaging method. Neurosci Lett 2004;359:57– 60.
