**3. Diffusion tensor tractography of cerebral white matter in the dog**

Several studies have demonstrated the validity of quantitative diffusion imaging of the large white matter tracts in the brain in vivo [25]. DTI has been extensively used to investigate brains of the dogs ex vivo [11, 34].

There is only one report of the use of DTT in dogs to determinate the feasibility for in vivo examination of the normal appearance of the cerebral white matter.

MRI allows investigators and clinicians to observe the anatomy and injuries of the cerebral white matter (CWM) in dogs. However, dynamic images based on the diffusion tensor (DT) technique are required to assess fiber tract integrity of the CWM. The goal of this study was to determine the feasibility of DTT for in vivo examination of the normal appearance of CWM in dogs through visual and quantitative analysis of the most representative CWM tracts [1].

## **3.1. Materials and methods: experimental animals**

Nine healthy canine patients of varying breeds and genders were prospectively recruited for the study. The dogs received a general physical and neurological exam, and blood samples were taken for a preanesthetic profile. During the imaging procedure, the dogs were anesthetized with diazepam (2 mg/kg IV, valium, Roche, Nutley, New Jersey) and propofol (4 mg/kg IV, recofol, Bayer, Turku, Finland) [1].

## **3.2. Image acquisition technique**

In dogs, it has been used for the evaluation of the spinal cord and its pathologies, providing statistically and visually different images when evaluating fractional anisotropy and ADC in normal dogs compared against dogs with lesions localized in the spinal cord using different

DTI technique has also been applied to image animal brains in vivo and ex vivo [1, 10, 11, 20, 34]. Previous work shows consistent results between diffusion anisotropy of in vivo and ex vivo formalin-fixed mouse brains. This offers a new opportunity to study the brain microstructure with ex vivo DTI, as it avoids motion artifacts and allows for longer imaging time [20].

The use of DTI in white matter tracts found in dog and human brains has the potential for

The corpus callosum has been also studied in dogs with DTI because its aging and pathologies are similar of those found in the human corpus callosum. In Pierce's work, the corpus callosum was segmentated into six major White matter tracts, which will provide grounds for new

Tracing tracts by DTI is relative new, and it is expected that further research will develop new technologies soon. However, recent studies have demonstrated that even simple methodologies are able to visualize cerebral white matter tracts connections in situ for both human and

To date, DT remains the only noninvasive method for visualizing human brain and spinal cord connections. DT suffers from both fundamental and practical limitations that limit its use for modelling brain connections. Unlike many invasive modalities, DT is incapable of determining the direction of information flow, nor can it distinguish single- and multineuron connections. DT may also have difficulty in resolving complex intravoxel fiber crossings or nondominant fiber populations due to limitations in scan time, hardware, or processing methods. Despite its many limitations, DT has been successfully used to model human neuronal connections for over two decades, including several pathways that are putatively deep brain stimulation targets. DT generation can be divided into three separate steps: data acquisition, data processing, and tracking. Each of these steps has several variables that must be considered in order

Different methods for the acquisition and analysis of DTI have been developed and have improved the precision of diffusion tensor measurements in recent years, so, new innovations can be expected. New pulse sequences and diffusion tensor encoding schemes are being developed to improve the spatial resolution, accuracy and to decrease artifacts in diffusion

studying several pathologies by correlating DTI findings with clinical symptoms [11].

types of scanners and software [19, 22, 24, 38–40].

188 Canine Medicine - Recent Topics and Advanced Research

research in both species [34].

animals [15].

**2.4. DTI limitations**

to ensure accurate DT [41].

**2.5. New research on DTI**

tensor measurements [3].

The MRI protocol was carried out in the same 3 Tesla scanner for all dogs. Diffusion tensor imaging was performed on each patient. Moreover, T1- and T2-enhanced images were acquired to obtain a high-resolution anatomical reference. T1- and T2-weighted images and DTI were obtained in different planes (transverse, dorsal, and sagittal). Three-dimensional reconstructions, FA, and ADC values were obtained for the left and right corticospinal tracts, the corpus callosum, the cingulum, and the right and left fronto-occipital fasciculus to visually evaluate and quantify these fiber tracts [1].

#### **3.3. Diffusion tensor imaging tractographies**

Diffusion tensor tractography was performed by importing DTI into image analysis software. Cerebral white matter tracts were identified using regions of interest (ROIs). The software identified tracts based on finding the most favorable path between two manually placed ROIs. Regions of interest were positioned where trajectories of the cerebral white matter fiber tracts were estimated to be, based on veterinary anatomy guides and a human DTI atlas. Highresolution T1-images were placed on top of the colored map to identify connections between anatomical structures. The different tracts were identified, delineated, and reconstructed at different points along their trajectory using the color map in the sagittal, dorsal, and transverse planes, which were reconstructed using a fiber-tracking algorithm. Data were coded in red to indicate a right-left direction, green to indicate a dorsoventral direction, and blue to indicate a rostrocaudal direction. The cerebral white matter tracts were assigned to three groups of fibers: projection, commissural, and association fibers [1].

#### **3.4. Data analysis**

Statistical analyses were selected and performed using a commercially available statistical software package (SSPS, version 19, Microsoft, Chicago, IL). Mean tract FA and ADC values, their standard errors, and standard deviation were calculated. A confidence interval of 95% or a significance value of P < 0.05 was used for the mean. A quantitative assessment of ventricular volume (VV) in relation to the brain volume (BV) was also performed using manual segmentation in regions of interest (ROIs) on the image analysis freeware (OsiriX v.3.9.4) in the nine healthy dogs. The means, standard errors, standard deviations, and 95% confidence intervals for the means of the VV in relation to the BV of the right and the left side were obtained [1].

#### **3.5. Results**

Three-dimensional reconstructions of the corticospinal tract, corpus callosum, cingulum, and fronto-occipital fasciculus were generated for each of the nine dogs. Fibers in the corticospinal tract component of the projection fiber group were displayed in blue and green (**Figure 1A**–**C**) [1].

Blue fibers connected cortical areas in the cerebral cortex, the brain stem, and spinal cord. Green fibers connected the corona radiate, internal capsule, and cerebral peduncle. Fibers in the corpus callosum component of the commissural fiber group were displayed in red and connected the two cerebral hemispheres (**Figure 2A**–**C**).

The cingulum component of the association fiber group appeared as long fibers, and these were displayed in blue (**Figure 3A**–**C**) [1].

These fibers had a rostrocaudal orientation and connected cortical areas in each hemisphere. Fibers in the superior and inferior fronto-occipital fasciculus component of the association fiber group were long and displayed in blue (**Figure 4A**–**C**) [1].

**3.3. Diffusion tensor imaging tractographies**

190 Canine Medicine - Recent Topics and Advanced Research

fibers: projection, commissural, and association fibers [1].

connected the two cerebral hemispheres (**Figure 2A**–**C**).

fiber group were long and displayed in blue (**Figure 4A**–**C**) [1].

were displayed in blue (**Figure 3A**–**C**) [1].

**3.4. Data analysis**

**3.5. Results**

(**Figure 1A**–**C**) [1].

Diffusion tensor tractography was performed by importing DTI into image analysis software. Cerebral white matter tracts were identified using regions of interest (ROIs). The software identified tracts based on finding the most favorable path between two manually placed ROIs. Regions of interest were positioned where trajectories of the cerebral white matter fiber tracts were estimated to be, based on veterinary anatomy guides and a human DTI atlas. Highresolution T1-images were placed on top of the colored map to identify connections between anatomical structures. The different tracts were identified, delineated, and reconstructed at different points along their trajectory using the color map in the sagittal, dorsal, and transverse planes, which were reconstructed using a fiber-tracking algorithm. Data were coded in red to indicate a right-left direction, green to indicate a dorsoventral direction, and blue to indicate a rostrocaudal direction. The cerebral white matter tracts were assigned to three groups of

Statistical analyses were selected and performed using a commercially available statistical software package (SSPS, version 19, Microsoft, Chicago, IL). Mean tract FA and ADC values, their standard errors, and standard deviation were calculated. A confidence interval of 95% or a significance value of P < 0.05 was used for the mean. A quantitative assessment of ventricular volume (VV) in relation to the brain volume (BV) was also performed using manual segmentation in regions of interest (ROIs) on the image analysis freeware (OsiriX v.3.9.4) in the nine healthy dogs. The means, standard errors, standard deviations, and 95% confidence intervals for the means of the VV in relation to the BV of the right and the left side were obtained [1].

Three-dimensional reconstructions of the corticospinal tract, corpus callosum, cingulum, and fronto-occipital fasciculus were generated for each of the nine dogs. Fibers in the corticospinal tract component of the projection fiber group were displayed in blue and green

Blue fibers connected cortical areas in the cerebral cortex, the brain stem, and spinal cord. Green fibers connected the corona radiate, internal capsule, and cerebral peduncle. Fibers in the corpus callosum component of the commissural fiber group were displayed in red and

The cingulum component of the association fiber group appeared as long fibers, and these

These fibers had a rostrocaudal orientation and connected cortical areas in each hemisphere. Fibers in the superior and inferior fronto-occipital fasciculus component of the association

**Figure 1.** Images illustrating the corticospinal tract in a healthy dog. (A) DTT image, side view. (B) T1-weighted image, sagittal view. (C) Colored map, transverse sectional view [1].

**Figure 2.** Images illustrating the corpus callosum in a healthy dog. (A) Diffusion tensor tractography (DTT) image, side view. (B) T1-weighted image and diffusion tensor tractography (DTT) image, dorsal view. (C) Diffusion tensor tractography (DTT) image on the colored map, dorsal view [1].

Diffusion Tensor Tractography in Cerebral White Matter http://dx.doi.org/10.5772/66249 193

**Figure 3.** Images illustrating the cingulum in a healthy dog. (A) Diffusion tensor tractography (DTT) image, dorsal view. (B) T1-weighted image and diffusion tensor tractography (DTT) image, dorsal view. (C) Diffusion tensor tractography (DTT) image on the colored map, dorsal view [1].

**Figure 2.** Images illustrating the corpus callosum in a healthy dog. (A) Diffusion tensor tractography (DTT) image, side view. (B) T1-weighted image and diffusion tensor tractography (DTT) image, dorsal view. (C) Diffusion tensor tractog-

raphy (DTT) image on the colored map, dorsal view [1].

192 Canine Medicine - Recent Topics and Advanced Research

**Figure 4.** Images illustrating the fronto-temporo-occipital tract in a healthy dog. (A) Diffusion tensor tractography (DTT), side view. (B) T1-weighted image and diffusion tensor tractography (DTT) image, sagittal view. (C) Diffusion tensor tractography (DTT) image on the colored map, dorsal view [1].

**Figure 5.** Comparison between a healthy dog with large lateral ventricles and a healthy dog with normal ventricles. (A) Diffusion tensor tractography (DTT) image, corticospinal tract with altered topography and corpus callosum due to large lateral ventricles, dorsal view. (B) Diffusion tensor tractography (DTT) image, corticospinal tract with normal downward topography of the fibers, dorsal view. (C) T1-weighted image and diffusion tensor tractography (DTT) image, corticospinal tract with anterior and lateral displacement of the fibers and of the corpus callosum due to large lateral ventricles, dorsal view. (D) T1-weighted image and diffusion tensor tractography (DTT) image, corticospinal tract with normal downward topography of the fibers, dorsal view. (E) Diffusion tensor tractography (DTT) image of the corticospinal tract on the colored map with anterior and lateral displacement of the fibers and of the corpus callosum due to large lateral ventricles, dorsal view. (F) Diffusion tensor tractography (DTT) image, corticospinal tract on the colored map with normal topography of the fibers, dorsal view [1].

These fibers had a rostrocaudal orientation and connected cortical areas in each hemisphere and in the frontal and occipital lobes. Three-dimensional reconstructions of the tracts were homogeneous and uniform in geometry and spatial orientation in eight of the nine healthy dogs. In one dog, tract reconstructions were not homogeneous or uniform because the fibers were displaced, most evident in the corticospinal tract and corpus callosum (**Figure 5**) [1].

There was a significant difference in the VV in relation to the BV in this dog (**Table 1**) [1].

**Figure 4.** Images illustrating the fronto-temporo-occipital tract in a healthy dog. (A) Diffusion tensor tractography (DTT), side view. (B) T1-weighted image and diffusion tensor tractography (DTT) image, sagittal view. (C) Diffusion

tensor tractography (DTT) image on the colored map, dorsal view [1].

194 Canine Medicine - Recent Topics and Advanced Research

R, relation; VV, ventricle volume; BV, brain volume; X, mean; SE, standard error; SD, standard deviation; C.I., 95% confidence interval for the mean.

**Table 1.** Ventricular volume (VV) in relation to the brain volume (BV) in healthy dogs <inlinefx>.

The means, standard errors, standard deviations, and 95% confidence intervals for the means of the FA and ADC values for the six tracts were obtained from all nine dogs (**Table 2**) [1].


FA, fractional anisotropy; ADC, apparent diffusion coefficient; X, mean; SE, standard error; SD, standard deviation; C.I., 95% confidence interval for the mean.

**Table 2.** Fractional anisotropy and apparent diffusion coefficient values of six cerebral white matter tracts in healthy dogs.

Similarities in the FA and ADC values were identified in the nine healthy dogs [1].

#### **3.6. Discussion**

**Healthy dogs (***n* **= 9)**

 **s−1)**

RVV/BV (left) 0.051 0.019 ±0.057 (0.007, 0.095) RVV/BV (right) 0.044 0.016 ±0.047 (0.007, 0.080)

> R (VV/BV) Left R (VV/BV) Right

R, relation; VV, ventricle volume; BV, brain volume; X, mean; SE, standard error; SD, standard deviation; C.I., 95%

The means, standard errors, standard deviations, and 95% confidence intervals for the means of the FA and ADC values for the six tracts were obtained from all nine dogs (**Table 2**) [1].

**FA ADC (10−3 mm2**

Corticospinal (right) 0.391 0.009 ±0.026 (0.371, 0.412) 1.154 0.063 ±0.189 (1.009, 1.299) Corticospinal (left) 0.400 0.006 ±0.018 (0.387, 0.414) 1.104 0.055 ±0.164 (0.977, 1.230) Corpus callosum 0.365 0.007 ±0.022 (0.348, 0.382) 0.975 0.033 ±0.099 (0.899, 1.051) Cingulum 0.336 0.011 ±0.032 (0.311, 0.360) 0.847 0.053 ±0.159 (0.725, 0.969) Fronto-occipital (right) 0.331 0.009 ±0.026 (0.311, 0.350) 0.879 0.009 ±0.028 (0.858, 0.901) Fronto-occipital (left) 0.331 0.010 ±0.029 (0.308, 0.353) 0.913 0.010 ±0.031 (0.888, 0.937) FA, fractional anisotropy; ADC, apparent diffusion coefficient; X, mean; SE, standard error; SD, standard deviation;

**Table 2.** Fractional anisotropy and apparent diffusion coefficient values of six cerebral white matter tracts in healthy

Similarities in the FA and ADC values were identified in the nine healthy dogs [1].

X SE SD C.I. 95% X SE SD C.I. 95%

**Table 1.** Ventricular volume (VV) in relation to the brain volume (BV) in healthy dogs <inlinefx>.

X SE SD C.I. 95%

**Relation ventricle volume/brain volume**

0.000 0.050 0.100 0.150 0.200 0.250

dogs.

12345678 9

196 Canine Medicine - Recent Topics and Advanced Research

**Tracts Healthy dogs (***n* **= 9)**

C.I., 95% confidence interval for the mean.

confidence interval for the mean.

**Relation of Ventricle Volume/Brain Volume** 

The current study was the first to visually and quantitatively describe the trajectory of cerebral white matter fiber tracts in a group of live dogs using DTT for diagnostic purpose. Diffusion tensor tractography-imaging resolution allowed rapid display and identification of the most representative cerebral white matter fiber tracts in this sample population of nine healthy dogs of varying breeds and genders. This technique described anatomical, geometric, and spatial properties of the fibers. In addition, the conduction properties of the fibers could be estimated through FA and ADC quantification. There was homogeneity and uniformity in the threedimensional reconstructions in nearly all dogs. Quantification of the FA and ADC values of the most representative tracts was similar in all nine dogs, thus demonstrating the feasibility of the technique and the analysis of the normal appearance of the cerebral white matter (CWM) in dogs in vivo. The analysis of the images in different planes and the three-dimensional reconstructions of the fiber tracts revealed a visual difference in the normal cerebral white matter appearance in one of the nine healthy dogs. This dog showed an altered topography of the corticospinal tract and corpus callosum due to displacement of the fibers. Quantitative assessment of ventricular volume in relation to the brain volume showed that this dog had larger lateral ventricles in relation to the other dogs. Authors believe this was most likely a normal variant because the dog exhibited no clinical neurological signs at the time of imaging. Other studies have shown that some canine species may exhibit a broad range of normal cerebral ventricular sizes, as assessed by neuroimaging, and that cerebral ventricular size is quite variable in a normal dog [1].

The only values quantified in the current study were those specific to the fiber tract and not to the ROIs because our research focused on demonstrating the feasibility of DTT for displaying and the normal appearance of the most representative cerebral white matter fiber tracts of healthy dogs in vivo. The quantification of FA and ADC values, or the exact description of the ROIs, were not used to visualize fiber tracts, as in our experience these values and descriptions can be obtained at different points of the fiber tract trajectories using the colored map. Future studies are needed to develop a method for preparing DTT templates and to reconstruct cerebral white matter pathways in the healthy dog in vivo using an ROI approach. This method is similar to the method used in humans and provides virtual representations of cerebral white matter tracts that are faithful to the classical ex vivo descriptions that include a detailed anatomical study of canines. The preparation of cerebral white matter templates for the healthy dog in vivo will allow investigators to follow the trajectory of the fibers delineating ROIs in DTI for DTT reconstruction. Improved knowledge of anatomy and the development of a template as a guide for the placement of ROIs could be used in future applications such as teaching and guiding virtual dissections of cerebral white matter tracts in healthy dogs in vivo, and comparison with pathological cases where the anatomy is observed distorted by the underlying disease process. Our findings are preliminary and future research will be needed to evaluate the use of DTT in clinical cases. Indeed, the use of this technique as a diagnostic tool is currently limited because templates and standardized FA and ADC values of the cerebral white matter of healthy dogs in vivo (which are prerequisites for the use of this technique in pathological cases) are lacking. Three previous canine DTI studies were conducted ex vivo, and one was conducted in vivo. In these studies, DTI was used to examine the structure and microstructure of the cerebral white matter but not for diagnostic purposes in canines. The findings of one study supported our study in that they also revealed the presence of association, commissural, and projection fibers in the dog. These data and information are also available in humans, and this allowed us to identify, compare, and reconstruct different fiber tracts in our healthy dogs in vivo. The results of study showed that the DTT reconstructions allowed identification and differentiation of CWM tracts in dogs in vivo, and that these reconstructions were comparable with those obtained in ex vivo canine studies. As new MRI options become increasingly available for daily clinical veterinary practice, novel diffusion techniques such as DTT warrant further exploration. In conclusion, findings from the current study indicated that DTT is a feasible noninvasive technique for in vivo study of CWM fiber tracts in the dog. We believe that the implementation of DTT as a noninvasive diagnostic method will complement conventional MRI, thus allowing investigators to examine the microstructural characteristics of the brain in vivo, and to obtain information on the anatomy, connectivity, and morphology of possible damage to fiber tracts in dogs suffering intracranial pathology or injury. Future research is needed to develop standardized ROI templates for in vivo canine studies and to compare DTT findings with confirmed pathologic findings [1].
