Intervertebral disc disease is a degenerative condition of unknown cause that results in herniation of the disc or disc material into the vertebral canal, compressing the spinal cord or spinal nerve roots [7]. Diagnostic imaging modalities for intervertebral disc disease include survey radiography, myelography, computed tomography (CT), and magnetic resonance imaging (MRI). One study has shown that a correct diagnosis of acute intervertebral disc prolapse (Hansen type I) may be made from survey radiographs in 69 to 72% of dogs [8]. However, false positive and false negative results are common [9]. Though the main roles of survey radiography are to help rule out differential diagnoses and to confirm anatomical landmarks, either myelography or advanced imaging should be performed for definitive diagnosis [15]. Since the development of non-ionic, water soluble, iodinated contrast media, myelography has been the standard diagnostic method for intervertebral disc disease [11]. Although the accuracy of myelography in one canine study was 86 to 97%, this diagnostic modality is invasive with the potential to cause side effects such as seizures and exacerbation of neurological signs [11].

Computed Tomography provides more accurate information and takes less time than myelography, particularly in chondrodystrophoid breeds, and is usually much easier to decide which side the disc material is located on from CT images than from myelograms [15]. In one study, the accuracy of CT and myelographies in diagnosing acute intervertebral disc disease was compared in 20 dogs. CT and myelography was found to be 90 and 88% accurate, respectively, at identifying the major site of disc herniation. Lateralization of disc material was correctly predicted in 96% of dogs with CT and in 92% of dogs with myelography [3].

MRI also provides transverse imaging and is superior to CT when the disc material is not mineralized. If the extruded material is not mineralized, contrast medium may be necessary [15]. Obviously the use of CT with subarachnoid contrast medium invalidates the potential advantage of diagnosing disc herniation without the side effects associated with myelography [11]. However, intervertebral disc herniation, especially in chondrodytrophoid breeds, usually occurs after the nucleus pulposus has become mineralized [6]. Therefore, mineralized intervertebral disc material in the vertebral canal can be detected clearly using CT without the subarachnoid injection of contrast medium [11].

The aim of this study was to describe the computed tomographic characteristics of the acute thoracolumbar intervertebral disc diseases and to pursue any exploitable relationships between clinical parameters and CT imaging characteristics.

The herniation type of intervertebral discs are divided into Hansen types I and II. Type I is more common in chondrodystropic breeds such as the Dachshund and the disc extrusion often results in more severe symptoms than type II. The nucleus pulposus of a chondroid disc is often extruded in an acute or subacute fashion [10]. In this study, all of the dogs which were affected acute thoracolumbar intervertebral disc disease were chondrodystrophic breeds except for two mixed-breeds. Distribution of the 645 cases of intervertebral disc disease by breed, sex, and anatomic site of involvement was studied. More than 53% of the cases were Dachshunds and the other breeds most often represented were Cocker Spaniel, Beagle, and Pekingese in order of decreasing frequency [5]. In this study, Pekingese was represented most frequently, and less for the Maltese, Cocker Spaniel, Shihtzu, Dachshund, in that order. Dachshunds were represented in only 5% of cases. It is thought to be due to possible regional breed population characteristics, although the contributing cause is not clear.

The normal spinal cord had intermediate attenuation on transverse images (mean attenuation: 31.3 ± 8.6 HU; range: 4~55) [11]. In this study, the normal spinal cord had a slightly high mean attenuation, but similar to the range of attenuation (mean attenuation: 23.4 ± 11.8 HU; range: -0.6 ~59.3 HU). There have been reports of identification of the cord and the level of the conus with CT by measuring the spinal canal HU. This value will be helpful to identify the spinal cord when the spinal cord exists with the herniated disc materials. But there were some limitations. In individual patients, HU values often varied randomly and it was impossible to obtain a reading reliable enough to predict the level of the conus. More problematic were the artifacts from the intermittent bony ring of the spinal canal that contains the cord and an unpredictable amount of other bone and gas from the surrounding body. The patient's respiration and other motions add to the artifacts [13].

In thoracolumbar disc disease, over 50% of lesions occur at the T12-T13 and T13-L1 discs, and more than 85% occur between T11-T12 and L2-L3, with the disc extrusion occurs occasionally as far cranial as T9-T10 [15]. The intercapital ligaments are short, transverse fibrous bands that lie ventral to the dorsal longitudinal ligament, joining the rib heads between T2 and T11, and these ligaments buttress the dorsal part of the annulus cranial to T11 and help resist dorsal disc protrusion [18]. In this study, the T13-L1 disc was the most common site of thoracolumbar disc disease, with 53% of lesions occurring at the T12-T13 and T13-L1 discs, and 87% occuring between T11-T12 and L2-3. In dogs, the mobility of the vertebral column is especially large in the thoracolumbar junction, and intervertebral disc disease is recorded most frequently at the most mobile region of the vertebral column in all animals [2]. It is considered that this is why the most common site of disc herniation is the thoracolumbar junction as with the results of the present study.

In the distribution pattern for each breed, the Maltese showed a tendency to have more single pattern disc herniation than other patterns. In the distribution patterns of all affected disc spaces, continuous pattern was more frequent at the thoracic than lumbar spine. There were two continuous patterns of disc herniation at T9-T10 in all affected disc spaces. In both of the cases, there was herniated disc material extended up to 4 intervertebral disc spaces. One of 2 cases had decompressive surgery, epidural hemorrhage was observed at surgery. It is considered that the disc material at T9-T10 extended from the caudal disc space with epidural hemorrhage. In all affected disc spaces, the T12-T13 and T13-L1 disc spaces were also common sites of disc herniation.

Spinal CT scan effectively shows a medial and lateral disc herniation. In localizing the exact level of the intervertebral disc, CT has advantages over myelography [14]. In our study, 75 (88.2%) of 85 disc spaces were ventrally located, the highest rate. And 2 (2.4%) of 85 disc spaces were had dorsally located disc herniation. Both of these cases had surgery and identified the concurrent epidural hemorrhage. In the lateral location of disc herniation, the left side was more frequent than right in all affected disc spaces. But left sides occurred at 3 cases and right sides at 2 cases. So there was no significant difference in the incidence of left and right sides.

Clinical signs of intervertebral disc disease vary from only spinal hyperesthesia to paraplegia with or without pain perception [3]. Spinal cord compression by herniated discs causes a closed type of injury that alters cord function and structure [18]. The severity of signs depends on the velocity at which the compressive force is applied, the degree of compression, the duration of compression, and the ratio of spinal cord diameter to vertebral canal diameter [12]. In humans, the size of the disc herniation in relation to the size of the spinal canal provided the best positive correlation to the clinical findings [17]. There was also a significant positive correlation between the improvement from sciatic pain and the reduction in the size of the individual hernia [4]. In this study, two different methods to measure the size of herniated disc were used. A-index was significant correlated with the clinical grade. Particularly, average of top 5 was slightly more correlated. It could be the reason why some cases of disc herniation were the multiple or continuous patterns.

In humans, measuring the absolute area of the herniation and the spinal to canal is technically demanding, time consuming, and therefore still not suitable for routine use in clinical practice. A simplified way to describe the area of the spinal canal that is occupied by the hernia would be to measure the linear relationship between the size of the disc hernia and that of the spinal canal. The sagittal diameter of the canal and the herniation gave a less precise description of the relative size and morphologic characteristics of a lumbar disc hernia, but the ratio of sagittal diameter correlated more strongly to the degree of sciatic pain than the other two different ratios, which were calculated by area or two perpendicular lines [17]. But in this study, the ratio of sagittal diameter, which was the H-index, was not correlated with clinical grade. The possible contributing causes for this difference may include the variability and severity of clinical signs as well as the location.

Type I disc extrusion is often associated with rupture of the vertebral sinuses and hemorrhage into the epidural space, which can exacerbate spinal cord compression [1]. CT can identify acute hemorrhage in the vertebral canal. The appearance of CT images suggests that epidural hemorrhage can extend over a number of vertebrae to each side of the disc herniation. The herniated material was distinguishable from the spinal cord, but was only slightly more attenuating than the spinal cord (mean attenuation: 59 ± 17 HU; range: 38~98 HU), and extended over distances of up to five vertebrae [11]. In this study, the herniated material with concurrent epidural hemorrhage was extended over distances of up to three vertebrae (4 cases of continuous pattern and 1 case of multiple pattern of disc herniation), and was slightly more attenuating than the spinal cord, which was measured at a non-mineralized region (mean attenuation: 86 ± 26 HU; range: 32~149 HU). Two of three cases with dorsal location of disc material had confirmed epidural hemorrhage at surgery. The epidural hemorrhage could be predicted by the distribution pattern, location and attenuation of herniated disc materials. Although there were statistically significant differences between two groups when measured by two other methods, the measurement of herniated disc material at the non-mineralized region is thought to be more useful than at the total area due to the narrow range of mean HU value.

In chondrodytrophoid breeds, herniated disc materials were usually mineralized and were able to be described by CT. In conclusion, the area ratio of the herniated disc material to spinal canal could be calculated, which was correlated with clinical grade. And the HU value of the herniated disc material was also able to be measured. The HU value of the group with concurrent epidural hemorrhage was lower than the group without concurrent epidural hemorrhage. Therefore, it was considered that CT characteristics such as area ratio, HU value, distribution pattern, and location could be helpful for the accurate evaluation and prediction of the status of the patient with intervertebral disc disease.