Ronaldo C. da Costa
Generalities of Vertebral Column MRI1–10,
12–51, 53–66, 68–70
Magnetic resonance imaging (MRI) of the vertebral column has revolutionized the diagnosis of many spinal conditions such as fibrocartilaginous embolic myelopathy or type III interverte-bral disc disease. MRI is the only modality that allows direct visualization of soft-tissue structures such as the spinal cord and nerve roots, thus allowing characterization of spinal disor-ders well beyond those achieved with myelography and com-puted tomography (CT). The applications of spinal MRI are many and are expanding progressively. In addition to routine MRI sequences, such as T2- and T1-weighted (W) images, spe-cial sequences, such as diffusion tensor imaging and angiogra-phy, are beginning to be used, and may become routine in the future.
Advanced imaging modalities offer significant advantages over survey radiographs and myelography. There is a wealth of data in people supporting the significant benefits of MRI versus myelography, and the few comparative studies available support the findings seen in people. The overall diagnostic sensitivity of MRI is superior to CT, and as such MRI can be used to image the vast majority of spinal disorders, with few exceptions (e.g. spinal trauma caused by gunshot). CT and MRI allow visualization of different structures, so ideally should be seen as complementary imaging modalities. Routine survey radiographs are always rec-ommended prior to proceeding with advanced imaging because the area of interest may be more specifically localized, reduc-ing scannreduc-ing time, and severe spinal/osseous lesions, such as hemivertebrae or discospondylitis, may be identified without the need of advanced imaging studies. However, depending on the clinical status and treatment plan, advanced imaging may still be needed.
Compared to CT, MRI provides superior contrast resolution and is better suited for imaging soft tissues, such as the spinal cord, nerve roots, and intervertebral discs. MR images can be acquired in multiple planes, whereas CT images can only be acquired in one plane (typically transverse). Whereas myelogra-phy is still often used in conjunction with CT of the spine, this is not necessary with MRI because of the ability to alter tissue con-trast by applying different acquisition sequences. Thus, the asso-ciated morbidity often accompanying myelography is avoided.
Disadvantages of MRI compared to CT are the prolonged acqui-sition time and increased expenses.
MRI imaging techniques2, 3, 15, 16, 16, 37–39, 42, 49, 62
Positioning
Straight spinal alignment is extremely important for MRI to allow direct comparison of multiple sites on sagittal images. A trough may be necessary to achieve this in deep-chested dogs.
The spinal segment(s) to be imaged must be in close contact to the RF coil for a satisfactory signal-to-noise ratio. As table sur-face coils designed for spinal imaging are used in most instances, patients are typically positioned in dorsal recumbency.
Plane of acquisition
r Sagittal: often used as the survey series to localize a specific site of pathology as a long segment of the spine can be imaged. The T2-W sequence is preferred for initial assessment.
r Dorsal: useful to evaluate lateralized compressions, mainly when multiple lesions are present.
r Transverse: typically limited to specific sites of pathology as imaging extended lengths of spine in this plane is time intensive.
Image sequences and slice
r T2-W: fluid, such as cerebrospinal fluid (CSF) or edema, will be hyperintense.
r T1-W: fluid, such as CSF or edema, will be hypointense.
r FLAIR (fluid attenuated inversion recovery): pure fluid, such as normal CSF, will be hypointense, whereas edema or abnor-mal fluids will be increased in intensity.
r STIR (short tau inversion recovery) or fat saturation: the typi-cally hyperintense fat (on both T1- and T2-W sequences) will be of low signal intensity. If a sequence is T2-W, CSF can be more readily differentiated from epidural fat. If a sequence is T1-W, a contrast-enhancing lesion can be more readily differ-entiated from fat.
r Gradient echo (GRE): a fast sequence that is very sensitive to inhomogeneities in the magnetic field, though less sensitive to motion artifact because of its speed of acquisition. It is often used to detect areas of hemorrhage, as excessive iron concen-trations in hemorrhagic areas of the spinal cord will cause field inhomogeneity.
r MR myelogram sequence (single-shot turbo spin echo), HASTE (half-Fourier acquisition single-shot turbo spin echo), or heavily T2-W: this sequence is primarily useful to evaluate the dorsal and ventral subarachnoid spaces, thus it is useful to characterize arachnoid diverticula and intradural neoplasms.
Recommended standard planes, slice thicknesses and sequences
r 2–3 mm thick T2-W sagittal images followed by 2–3 mm thick T1-W sagittal images. A 0.5 mm gap between slices, to avoid image cross-talk and resultant decreased signal to noise, or interleaved (no interval/gap) acquisition is typical.
r Sagittal MR myelogram sequence (single-shot turbo spin echo, heavy T2).
r 2–3 mm transverse images through the lesion identified on the sagittal images.
r 2–3 mm dorsal images for lateralized lesions identified on the prior sequences.
r T1-W images after intravenous gadolinium injection may be obtained in all three image planes according to the lesion char-acteristics.
r STIR, FLAIR, fat saturation, and GRE sequences may be nec-essary depending on lesion characteristics.
Field of view (FOV)
Field of view is defined as the vertical or horizontal distance across an image. It should be set specific to the spinal column, which is dependent on patient size. By decreasing FOV to a spe-cific area, the signal-to-noise ratio decreases and thus increases in sampling (NEX; see below) or increases in slice width are nec-essary to maintain adequate spatial and contrast resolution. By increasing NEX, scan time is also increased.
Number of excitements (NEX)
This refers to the number of repeat signal samplings acquired in any given sequence. In general, the more times an area is sam-pled, the greater the signal-to-noise ratio (improved tissue con-trast). The downside is that the greater the NEX, the longer the overall imaging time. NEX values between 2 and 4 are typically required for field strengths over 1 tesla.
Saturation bands
As MRI is extremely sensitive to motion artifact secondary to respiration and blood flow, saturation bands may be placed over the abdomen and thorax to suppress the effects of motion from these areas during scanning.
Contrast imaging
Gadolinium diethylenetriaminepentaacetic acid (MRI) con-trast: indications include identification of vascular lesions like
those secondary to neoplasia, infectious/noninfectious inflam-matory lesions other than related to intervertebral disc hernia-tion, and vascular malformations.
Normal anatomy
The anatomic features of the spinal region being studied must be taken into consideration when interpreting advanced imaging studies. Each region of the vertebral column (cervical, thoracic, lumbar, sacral) has specific anatomic features. A larger degree of anatomic variation is seen in the cervical vertebrae compared to other vertebral regions. Not only are the cervical vertebrae different but also the shape and relationship of the spinal cord to the vertebral canal varies according to the location within the cervical column (Fig. 6.51). The anatomic features of the cervical spine and lumbosacral spine have been reported in detail using CT, MRI, or both.
Due to the exquisite sensitivity of MRI, it is important to assess the imaging findings in the light of the patient’s clinical signs and examination findings. The presence of asymptomatic spinal abnormalities is widely recognized in human medicine.
The presence of multiple spinal abnormalities is commonly seen in aged dogs and cats. Clinically silent spinal cord compres-sion, intervertebral disc degeneration, intervertebral disc pro-trusion, nerve root compression, intervertebral foraminal steno-sis, and vertebral canal stenosis have all been reported without concurrent clinical signs in dogs. Similarly, no correlation exists between the severity of spinal cord or nerve root compression with the severity of clinical signs.
Vascular diseases
Fibrocartilaginous embolic myelopathy (FCEM)1, 18, 20
MRI is the imaging modality of choice for the diagnosis of fibrocartilaginous embolic myelopathy. MRI depicts the spinal
(a) (b) (c)
Figure 6.51 Transverse T2-W images (WI) at the level of the C2–C3 (A), C5–C6 (B), and C7–T1 (C) intervertebral discs of a clinically normal Doberman Pinscher. Notice the different shapes of the spinal cord at each cervical level. The normal spinal cord at C7–T1 has a trapezoid shape. (da Costaet al., 2006.
Reproduced with permission from American Veterinary Medical Association.)15
(a)
(b) (c)
Figure 6.52 Fibrocartilaginous embolic myelopathy (FCEM). (A) Sagittal T2-WI of the cervical spine of a dog with the presumptive diagnosis of FCEM. Hyperintensity is seen at the level of the seventh cervical vertebra (C7) (arrow). (B) Transverse T2-WI at the level of C7 showing the asymmetric hyperintensity. (C) Transverse T1-WI postcontrast administration.
Note mild contrast enhancement at the same level of the hyperintensity seen on 2B.
cord parenchyma allowing visualization of the ischemic spinal cord region. The appearance of FCEM on MRI is typically of a focal (although sometimes extensive) hyperintense T2-W lesion affecting primarily the gray matter and usually clearly asymmetric (Fig. 6.52). The lesion is commonly isointense on T1-W images but can also be hypointense. Contrast enhance-ment is not commonly seen, it was observed in only 12% of cases in a study. It is more likely to see contrast enhancement when MR imaging studies are performed five to seven days after the ischemic event. Spinal cord swelling is a common fea-ture of FCEM. The severity of intramedullary lesions seen on MRI has been correlated with the severity of neurologic signs at presentation and with outcome. Interestingly, the presence of MRI changes was not associated with the timing of imaging. In some cases, primarily those imaged within 24 to 72 hrs after the ischemic injury, no signal changes are seen on MRI. These cases could be diagnosed with diffusion weighted MR imaging, although application of this technique in dogs is challenging.
Inflammatory spinal diseases Discospondylitis8, 10, 28, 29
Discospondylitis has been, by convention, a radiographic diag-nosis. Clinical signs often precede radiographic evidence of vertebral endplate lysis and medullary sclerosis, typically seen with more advanced disease. MRI is much more sensitive for detecting soft-tissue inflammation of the disc that usually
precedes bony changes. MRI may be preferred to CT in order to screen for early cases of discospondylitis where bony changes are not observed radiographically. Two recent studies reported the MRI findings in dogs with discospondylitis. The vertebral body was typically hypointense or isointense in both T2- and T1-W images. The endplates were hypointense or had mixed intensity in T1-W images, with hyperintensity in STIR-W images in most dogs (Fig. 6.53). Contrast enhancement of endplates and par-avertebral tissues was also frequently seen. The intervertebral discs were frequently hyperintense in T2-W and STIR images, isointense in T1-W images and had contrast enhancement in most cases.
Myelitis24
Myelitis is best evaluated with pre- and postintravenous contrast MR studies because contrast enhancement is a frequent finding (Fig. 6.54). Imaging features are very similar to, and are difficult to distinguish from, neoplasia. CSF analysis, surgical biopsy, and the patient’s overall clinical picture are necessary to substantiate the diagnosis of myelitis. A recent study reported STIR muscle hyperintensity in the deep cervical muscles (longus colli) in dogs with meningomyelitis. The finding had a strong association with CSF changes.
Epidural empyema21
Epidural empyema is a neurologic emergency characterized by the accumulation of purulent material within the vertebral canal
(a) (b)
(c) (d)
Figure 6.53 Discospondylitis. (A) Sagittal T2-WI showing subluxation between L1 and L2 with lysis of the endplates and intervertebral disc with heterogeneous hyperintensity of the endplates (arrow). Mild ventral spinal cord compression and spinal cord hyperintensity are also seen. (B) Sagittal T1-WI showing markedly irregular endplates with isointense signal of the intervertebral disc space. (C) Sagittal T1-WI postcontrast injection. Observe marked contrast enhancement of the intervertebral disc space, endplates, and peridural region. (D) STIR WI showing hyperintensity of the disc space and endplates.
(a)
(b) (c)
Figure 6.54 Myelitis—sagittal (A) and transverse (B, C) T1-W MR images after IV gadolinium injection showing the irregular areas of contrast enhancement in the spinal cord (arrows).
(b) (a)
Figure 6.55 Traumatic intervertebral disc herniation. The patient suffered a traumatic injury and had an acute onset of tetraparesis. (A) Sagittal T2-WI shows a focal area of spinal cord hyperintensity between C2–C3. (B) Transverse T2-WI shows the hyperintensity within the spinal cord parenchyma with a focal area of hypointensity (arrow) thought to be a fragment of extruded disc material within the spinal cord (type III disc).
causing fever, spinal pain, and progressive neurologic dysfunc-tion. Neurologic signs are secondary to the combined effects of regional tissue inflammation and spinal cord compression by an epidural mass effect. The epidural masses were hyperintense or had mixed intensity on T2-W, and were hypointense on T1-W images. Mild to moderate contrast enhancement was frequently observed. Spinal cord hyperintensity on T2-W images was a common finding in a study suggesting ischemia, extension of the infection, or focal malacia.
Spinal trauma31, 32, 36, 41, 46
Limited information is available on MRI and spinal trauma. In a study of 11 dogs with spinal fracture or subluxation, MRI iden-tified rupture of soft-tissue structures and/or fracture of two vertebral compartments (based on the three-compartment the-ory; see Chapter 15) in most dogs. The majority of dogs had spinal cord changes (hyperintensity on T2-W images), hemor-rhage, or swelling. Epidural hematoma was also identified in most dogs. Another study compared the diagnostic sensitivity of survey radiographs and CT in dogs with confirmed spinal frac-tures and luxations. Radiographs missed approximately 25% of the lesions detected on CT. CT is routinely used as the screening imaging of choice for spinal trauma in people due to its avail-ability and rapidity of imaging. Studies in humans suggest that MRI and CT should be used as complementary modalities, as CT has exquisite bone resolution whereas MRI has superior soft-tissue resolution. It should be pointed out that large case series in people indicated that MRI may miss a significant proportion of osseous lesions identified on CT.
Traumatic intervertebral disc extrusion is important to be considered in cases of external trauma. In a recent series, the majority (71%) of traumatic disc extrusions did not cause spinal cord compression. MR of the affected area identified spinal cord
hyperintensity on T2-W images and reduced the volume of nucleus pulposus of the intervertebral disc ventral to the area of hyperintensity. (Fig. 6.55).
Congenital spinal diseases Atlantoaxial subluxation9, 50, 52
Atlantoaxial subluxation or instability (AAS) is seen primar-ily in toy-breed dogs. Anatomic abnormalities associated with atlantoaxial subluxation include odontoid process (dens) mal-formation, hypoplasia, or aplasia, and/or weak or poorly devel-oped supporting ligamentous structures of the odontoid pro-cess. The result may be subluxation or luxation of the atlantoax-ial articulation, usually resulting in ventroflexion and a widened angle or space between the dorsal lamina of the atlas and spinous process of the axis (Fig. 6.56). If the odontoid process is intact and normally or partially formed, it may deviate dorsally within the spinal canal resulting in severe compression of the spinal cord, exacerbated by increased ventroflexion of the cranial cer-vical spine. MRI is sensitive for detecting secondary trauma to the spinal cord (hemorrhage, edema) secondary to instability at the articulations. The supporting ligaments can be visualized using high-field MRI and thin slices. Considering the possibility of concurrent neurologic diseases in these toy/small-breed dogs, it may be necessary to perform lumbar CSF collection to rule out inflammatory diseases.
Hemivertebra4, 35
Hemivertebra is commonly recognized in brachycephalic, screw-tail breeds. Besides the tail, the mid- to caudal tho-racic spine is often affected to varying degrees. It results from abnormal, uneven growth between the two halves of one or more vertebrae during development, with often, incomplete fusion between the halves. Because of the abnormal vertebral
(b) (a)
Figure 6.56 Atlantoaxial subluxation. (A) Sagittal T2-WI showing the odontoid process as a hypointense structure ventral to the spinal cord (long white arrow). A curvilinear hypointensity caudal to the cerebellum (*) represents caudal occipital overlapping with medullary kinking (short white arrows). (B) Transverse T1-WI through the odontoid process (*). The normally oval spinal cord has a kidney-bean shape secondary to ventral compression by the dorsally deviated odontoid process.
conformation, secondary spinal canal stenosis and spinal cord compression may occur. One of the greatest challenges in imag-ing these patients is that most patients have associated vary-ing degrees of kyphosis, lordosis, and scoliosis. Achievvary-ing well-positioned planar images is difficult. CT is preferred for a better definition of bone. If neurologic deficits are present, MRI is pre-ferred for spinal cord imaging(Fig. 6.57).
Spinal arachnoid diverticula (“cysts”)26, 40, 61, 62, 69
These diverticula consist of outpouchings of the arachnoid mat-ter or focal dilatations in the subarachnoid space that are filled with CSF. The use of the term “cyst” to describe these lesions is a misnomer, as these outpouchings are not lined by epithelium.
(a)
(b)
Figure 6.57 Hemivertebrae. Sagittal (A, B) T2-WI of the thoracolumbar spine of a French Bulldog with severe lordosis and scoliosis secondary to T7, T8, and T9 hemivertebrae.
These diverticula are commonly located in the cranial cervi-cal (C2–C4) region in large-breed dogs (mainly the Rottweiler), whereas in small-breed dogs the thoracic and thoracolumbar region is more commonly affected. The clinical appearance of these cases is of a chronic progressive cervical or thoracolumbar myelopathy (depending on the location of the diverticulum) fre-quently associated with fecal and/or urinary incontinence. There is still discussion as to whether these diverticula are congeni-tally acquired or secondary to trauma, arachnoiditis, interver-tebral disc disease, previous surgery, hematomas, or meningitis.
It has also been postulated that the arachnoid adhesions, rather than the subarachnoid outpouchings, are the cause of spinal cord compression and the resultant neurologic deficits. As the Rot-tweiler breed is overrepresented, a genetic or hereditary compo-nent is likely in at least some patients. CT myelography and MRI are equally adept at identifying cyst-like lesions of the subarach-noid space. MRI appearance is of a focal accumulation of CSF, usually dorsally, with the cranial aspect ending abruptly with a
“tear-drop” appearance. The MR myelogram sequence may be particularly useful to identify these dilations (Fig. 6.58). Spinal cord hyperintensity on T2-W images may also be seen associated with the subarachnoid dilation.
Spinal neoplasia7, 14, 22, 23, 42, 43, 51, 59, 60
Spinal neoplasia is an important differential diagnosis for dogs presenting with either chronic or acute neurologic signs. Sev-eral classification systems are used to categorize spinal neo-plasms. A common classification categorizes the tumors accord-ing to the location into: extradural, intradural extramedullary, and intramedullary. Extradural tumors such as osteosarcoma and fibrosarcoma are the most common.
Both CT and MRI are sensitive techniques for the diagnosis of spinal tumors. Due to the superior soft-tissue resolution, MRI is usually the preferred imaging method; however, CT is excel-lent for the visualization of osseous lesions, which are commonly observed in spinal tumors.
(a)
(b)
Figure 6.58 Spinal arachnoid diverticula. (A) Sagittal T2-WI showing a teardrop shape accumulation of CSF dorsally between C2–C3 (arrow). (B) HASTE sequence (MR myelogram) showing the conspicuous dilation of the dorsal subarachnoid space.
Lytic and proliferative regions observed on imaging need to be differentiated from infections such as discospondylitis or vertebral osteomyelitis. As a general rule, lytic and prolifera-tive lesions located in the vertebra itself are typical or neopla-sia, while lytic/proliferative lesions centered at the disc space
are caused by infections (discospondylitis). In one study, myel-ography was more useful in differentiating between intradu-ral extramedullary and intramedullary tumors than CT. This same observation was made when comparing myelography and MRI in another study. Careful evaluation of the images in all three planes (transverse, sagittal, dorsal) may assist in defining the location of the tumor. Dorsal images are particularly use-ful. The MRI findings of spinal meningiomas, the most com-mon intradural tumor, have been well described. Most menin-giomas are iso- to hyperintense on T1-W images and hyperin-tense on T2-W images. Homogeneous, strong contrast enhance-ment, and presence of dural tail are also consistently observed (Fig. 6.59).
MRI findings for other spinal tumors have not been well described. Osseous tumors have very variable signal intensity, ranging from iso- to hyper- or hypointense on both T1- and T2-W images. Contrast enhancement is also variable. It is important to use fa- suppression techniques (e.g. STIR) when observing hyperintense lesions on T1- and T2-W images within osseous structures, as these changes may be caused by fat infiltration.
Intramedullary tumors have more consistent imaging patterns.
Due to the common presence of edema, hyperintensity on T2-W images and hypointensity on T1-T2-W images, in relation to the surrounding spinal cord, is commonly observed. The pattern of enhancement is variable.
Primary or secondary nerve sheath tumors (NSTs) are the most common cause of neurogenic lameness. Approximately 45% of the tumors are located in the nerve roots proximal to
(a) (b)
Figure 6.59 Meningioma. T1-W MR images after IV gadolinium injection. (A) Dorsal and (B) transverse images show a large, mostly homogeneous, contrast-enhancing mass at the level of C1 and C2 (arrow). (From da Costa, RC. In Daleck CR, de Nardi AB, Rodaski S.Oncologia em c˜aes e gatos.
Guanabara: Koogan; 2009. Reproduced with permission.)
(a) (b)
Figure 6.60 Primary malignant nerve sheath tumor—distal location. (A) Transverse T2-W MR image reveals a large hyperintense mass at the level of the left axilla just lateral to the first rib (arrows). (B) Transverse T1-W MR image post-IV gadolinium injection shows an inhomogeneously enhancing mass at the same level of (A; arrows). Arrowhead indicates the spinal cord.
the spinal cord, while 55% are located in the plexus area or peripheral nerves. This emphasizes the importance of using a larger field of view when imaging patients suspected of hav-ing NSTs. The imaghav-ing features of NSTs have been described using CT and MRI. The visualization of NST is facilitated using MRI. CT findings commonly observed in dogs with NSTs are
rim enhancement and a hypodense center. On MRI, NSTs are consistently hyperintense on T2-W images, hypointense on T1-W images, and show homogeneous or inhomogeneous con-trast enhancement (Fig. 6.60). The tumor may appear as a dif-fuse brachial plexus nerve thickening or a circumscribed mass (Fig. 6.61).
(a) (b)
Figure 6.61 Primary malignant nerve sheath tumor—proximal location. (A) Dorsal T1-W MR image post-IV gadolinium injection shows an oval contrast-enhancing mass that appears to be displacing the spinal cord medially at the level of C5–C6 (arrow). (B) Transverse T1-W MR image at the level of C5–C6 shows a large contrast-enhancing mass involving the nerve roots and infiltrating into the spinal cord (arrow). (From da Costa, RC. In Daleck CR, de Nardi AB, Rodaski S.Oncologia em c˜aes e gatos.Guanabara: Koogan; 2009. Reproduced with permission.)
Figure 6.62 Multiple sites of thoracolumbar intervertebral disc protrusion in a 9-yr-old German Shepherd. Variable degrees of spinal cord
compression are seen at almost every single disc space starting at T9–T10 (long arrow). Compression is also seen at the lumbosacral space (short arrow). Intervertebral disc degeneration and spondylosis is also observed at multiple sites (arrowhead).
Biopsy is always needed to confirm the diagnosis of spinal or nerve sheath tumors, and to define the specific type of tumor.
Depending on the location of the tumor, biopsy can be guided using CT or ultrasonography.
Degenerative and developmental diseases Intervertebral disc disease
(IVDD)3, 5, 19, 25, 34, 45, 48, 56, 58, 62, 66, 68
Intervertebral disc disease is the most common spinal disease of dogs, and should be a differential diagnosis in any dog older than 1 yr of age with spinal pain and myelopathic signs.
MRI allows clear visualization of IVDD. The normal nucleus pulposus of the intervertebral disc appears a hyperintense ellip-soid area on sagittal T2-W images. Intervertebral disc degener-ation leads to a decrease in signal intensity and the degenerated disc becomes isointense to hypointense relative to the surround-ing annulus fibrosus. The degree of brightness of the T2 signal of the nucleus pulposus correlates with the proteoglycan concen-tration but not with water or collagen concenconcen-tration. MRI allows the detection of disc degeneration at earlier stages of the degen-erative process, before mineralization occurs, which is when it can be visualized using CT and radiographs. It is important to emphasize that disc degeneration per se does not lead to clini-cal signs, except in uncommon cases of discogenic spinal pain.
MRI also allows visualization of the spinal cord, which facilitates comparison when multiple sites are affected (Fig. 6.62). The size of the spinal cord and vertebral canal varies according to differ-ent spinal locations, so the same degree of disc protrusion may lead to different degrees of spinal cord compression according to the spinal location. Sagittal and transverse images should be used concurrently to assess the severity and lateralization of spinal cord compression (Fig. 6.63). It is common to observe lateral-ized disc extrusions causing nerve root compression, and this may not be seen on midsagittal images. Parasagittal images allow visualization of lateralized disc herniations and nerve root com-pression. Hemorrhage associated with disc extrusion can cause
a signal void on MR images. GRE sequences can confirm the presence of hemorrhage.
MRI also allows assessment of the spinal cord parenchyma and detection of spinal cord signal changes. In cases with mul-tiple sites of spinal cord compression, identification of hyperin-tensity on T2-W images indicates the site with the worst com-pression. The spinal cord hyperintensity seen on T2-W images correlate with the severity of clinical signs. The degree of severity of spinal cord compression, however, does not correlate with the severity of clinical signs. Three studies in dogs have indicated that the presence and extension of spinal cord signal changes have prognostic implications. A study suggested that areas of hyperintensity three times longer than the body of L2 were asso-ciated with poor prognosis, with only 20% of dogs with this sig-nal change regaining ambulatory status. The presence of hyper-intensity was a more reliable prognostic indicator than absence of nociception (deep pain perception). Even in noncompressive intervertebral disc extrusions, the extent of spinal cord hyperin-tensity predicted the outcome.
Chronic disc extrusions may be associated with inflamma-tory reaction surrounding the extruded disc leading to a ring enhancement pattern surrounding the disc (Fig. 6.64). The image may be mistaken as a granulomatous or neoplastic lesion.
It is important to be aware of this imaging feature of IVDD to avoid misdiagnosing it as other conditions. Intervertebral disc herniation can cause many different types of imaging patterns and should always be considered in the differential diagnosis for unusual MRI findings. In contrast to nonenhanced CT, normal MRI findings in all three imaging planes rules out the diagnosis of IVDD.
Cervical spondylomyelopathy (CSM)11–13, 17, 44, 47, 53, 63, 64
Cervical spondylomyelopathy is characterized by static and dynamic spinal cord compressions. The disease is commonly caused by osseous compressions in young, giant-breed dogs and by disc protrusion in middle-age to older large-breed dogs. Both MRI and CT have been used in the diagnosis of CSM in dogs.
When planning advanced imaging studies of the cases suspected of having CSM, it is important to plan the FOV to cover the entire cervical spine up to the third thoracic vertebrae. A recent study found compressions at T1–T2 and T2 associated with other cervical compressions in almost 10% of dogs.
MRI has been considered the best imaging technique for humans with cervical spondylotic myelopathy for more than 20 yrs. CT myelography is still used in humans for equivo-cal cases where cerviequivo-cal radiculopathy secondary to foraminal stenosis is the main clinical problem. The main advantage of MRI over CT myelography is the ability to directly visualize the spinal cord. This allows the detection of spinal cord signal changes that are helpful to determine the primary spinal cord lesion(s) in cases with multiple compressions (Fig. 6.65). Two studies indicate that multiple spinal compressions are seen in 63% of dogs with CSM.