Diagnostic Imaging of the Spine--Radiography, Myelography, CT & MRI
World Small Animal Veterinary Association World Congress Proceedings, 2004
Erik R. Wisner, DVM
School of Veterinary Medicine, University of California, Davis
Davis, CA, USA


The approach one takes in imaging the spine depends on the clinical signs of the patient and the suspected etiology and location of the lesion as determined by history and physical examination. For patients with spinal pain or neurological deficits, a step-wise diagnostic approach using survey radiography, myelography, x-ray computed tomography (CT) or magnetic resonance imaging (MRI) might be the preferred approach. For patients with primary spinal skeletal disease or injury, survey radiography with or without computed tomography might be satisfactory for diagnosis. Additionally, there are differing institutional imaging preferences for diagnosis of spinal disease with some hospitals now preferring MR as the initial imaging step in patients with neurological disease localized to the spine. Experience at our institution suggests that although the diagnostic imaging strategy must be tailored to each patient, often different imaging approaches will lead to an accurate diagnosis.


Survey radiography is still the most widely available imaging technique for evaluation of the spine and is an excellent screening test that can be used to dictate additional, more specific and expensive imaging studies such as CT or MRI. Because limited spinal survey examinations can be performed quickly and with minimal patient movement or restraint, it is particularly useful for evaluating patients with suspected spinal trauma and those with gross spinal skeletal anomalies. Cross-table horizontal views of the spine can be employed for trauma patients when spinal instability is a concern. Survey radiography is best used to evaluate skeletal structures and is useful for assessing gross spinal alignment/malalignment, vertebral contours, bone destruction or production, spinal canal stenosis, intervertebral disk space width and disk mineralization. A limited evaluation of soft tissues within the spinal canal and adjacent to the spinal column may also be possible. However, with few exceptions, survey radiology should not be used to establish if spinal cord compression is present. A myelogram or cross-sectional imaging study such as CT or MR is generally necessary to make this determination. Survey radiography, sometimes used in conjunction with myelography or epidurography, is also useful for documenting spinal instability with positional studies such as flexed and extended views of the lumbosacral junction to document positional spinal canal narrowing or with traction views of the cervical spine to verify dynamic spinal cord compressive lesions.


Myelography has been the mainstay of spinal neuro-imaging in veterinary medicine for decades. With the recent advent of CT and MR in veterinary clinical practice, it is tempting to dismiss myelography entirely but this may be misguided. Although CT and MR have, for the most part, supplanted myelography in human medicine, neurological localization of spinal lesions is also much more refined and lesions can be localized very specifically leading to a very accurately targeted cross-sectional imaging study. Because anatomic localization of spinal lesions is often less specific in veterinary patients, myelography can still serve as a valuable localizing tool particularly when followed by CT or MR for more specific diagnosis. In some instances, myelography can be more definitive in characterizing spinal cord compression, cord atrophy and defining the presence and location of syringohydromyelia than can be done with other imaging modalities.

Non-ionic iodinated contrast material is typically injected at the level of the L4-5 or L5-6 intervertebral space. If the lesion is localized to this site or if initial attempts at lumbar contrast injection are unsuccessful, a cisternal contrast injection is performed. Although selective regional myelograms can be performed, we suggest that the entire spine be evaluated to avoid missing clinically silent secondary lesions that might be masked by a more cranial lesion. Standard radiographic views consist of lateral and ventrodorsal projections with the patient neutrally positioned. Dorsoventral views are sometimes acquired in patients with caudal cervical lesions due to the better intrathecal contrast distribution at this level of the spine. Oblique views are often obtained when a focal lesion is identified or suspected from standard views to better define contrast column attenuation or deviation around the lesion. Extended, flexed and traction views are used to document a dynamic, usually compressive, lesion not convincingly seen on neutral views.

Occasionally, additional contrast studies are performed to evaluate lesions of the caudal lumbar spine or lumbosacral junction that may not be demonstrated using conventional myelography. Because the thecal sac of the spinal cord tapers in diameter and does not extend caudally to the to the level of the lumbosacral junction in most patients, epidurography is sometimes used to define the ventral margin of the spinal canal at this level as a basis for diagnosing cauda equina compression. This study is often performed using neutrally positioned and flexed and extended views to detect dynamic compressive lesions associated with instability and intermittent subluxation of the lumbosacral joint. Contrast material is typically injected into the epidural space at the level of the lumbosacral junction or between the most proximal caudal vertebral segments.


As with conventional radiography, computed tomography relies on tissue density differences as the basis for image formation. An x-ray beam passes through and is partially attenuated by the patient as the x-ray tube revolves about the patient while the patient is simultaneously moved through the open ring (gantry) of the CT scanner. A series of small x-ray detectors are positioned opposite the x-ray tube along the circumference of the gantry opening. X-rays that pass through the patient interact with the detector array and each detector measures the deposited x-ray energy. This process is repeated many times as the x-ray tube rotates around the patient and the recorded data are used to construct digital cross-sectional images of the patient. Because each image represents such a thin (0.5-10 mm) slice of the structure in question, the summation effects that are inevitable with conventional radiography are minimized. Images can also be reformatted to depict anatomy in other planes or as three-dimensional renderings. Contrast resolution of CT is excellent compared with conventional radiography as there is significantly better discrimination of tissues having similar densities.

Computed tomography is often employed when survey radiology/myelography has failed to adequately define or characterize a spinal lesion or when a spinal lesion has been sufficiently localized to make a targeted CT study practical. In our practice, most CT studies of the spine performed as part of a neurologic work-up are done after a myelogram and iodinated contrast material is already present within the subarachnoid space delineating the spinal cord and arachnoid/dural margins. When mass lesions are detected, intravenously administered iodinated contrast media is often used to better characterize lesion margins and vascularity. In addition to having greater sensitivity than radiography for detecting subtle bone destruction or production, CT is also excellent for revealing mineralized and non-mineralized disk material, determining the extent and character of spinal cord compression, confirming the presence of cord atrophy or syringohydromyelia and contrast-enhancing masses.


For routine clinical MR imaging, water hydrogen protons within a patient are aligned with the axis of a magnetic field produced by the magnetic resonance scanner. Pulsed radiofrequency (RF) waves directed at the patient cause the hydrogen protons to wobble after which they return to alignment. This phenomenon is referred to as proton relaxation. By manipulating the applied RF pulse, characteristic signals emitted by the tissues can be detected by a receiver within the scanner. Because the relaxation behavior of hydrogen protons depends on their microenvironment, variations in tissue-emitted signals can be used to discriminate between different tissue types. By systematically altering the applied RF pulses, anatomy can be mapped in all three dimensions, voxel by voxel. In addition, by varying the RF pulse sequence, the returning signal from different tissues can be either enhanced or suppressed. As with CT, MR images are typically displayed as thin slices of anatomy on a square or rectangular matrix. Unlike CT, in which image formation depends on tissue density differences, MR relies on differences in chemical properties and resulting magnetic resonance responses of tissues for image formation and tissue discrimination. Contrast resolution of MR is excellent making it particularly useful for imaging soft tissues that are not adequately evaluated using radiography or CT.

The greatest value of MR is its ability to depict and discriminate different normal and abnormal soft tissues. Accurate differentiation of gray and white matter of the spinal cord, detection of central canal dilatation or syrinx formation, detection of edema fluid and hemorrhage, characterization of mineralized and non-mineralized disk material, detection of contrast-enhancing masses or reactive tissues can all be accomplished with MR. As with CT, contrast media is often administered following acquisition of initial images to better delineate vascular or mass lesions. Some institutions have begun to use MR as both an initial screening test as well as a final diagnostic study. Often the initial screening study consists of sagittal images of the entire spine using standard RF pulse sequences followed by axial image acquisitions targeted to specific anatomical areas of interest as dictated by survey findings.


1.  Roberts, R.E. and B.A. Selcer, Myelography and epidurography. Vet Clin North Am Small Anim Pract, 1993. 23(2): p. 307-29.

2.  Gomez, M., et al., Computed tomographic anatomy of the canine cervical vertebral venous system. Vet Radiol Ultrasound, 2004. 45(1): p. 29-37.

3.  Axlund, T.W. and J.A. Hudson, Computed tomography of the normal lumbosacral intervertebral disc in 22 dogs. Vet Radiol Ultrasound, 2003. 44(6): p. 630-4.

4.  Jones, J.C. and K.D. Inzana, Subclinical CT abnormalities in the lumbosacral spine of older large-breed dogs. Vet Radiol Ultrasound, 2000. 41(1): p. 19-26.

5.  Jurina, K. and V. Grevel, Spinal arachnoid pseudocysts in 10 Rottweilers. J Small Anim Pract, 2004. 45(1): p. 9-15.

6.  Mayhew, P.D., et al., Association of cauda equina compression on magnetic resonance images and clinical signs in dogs with degenerative lumbosacral stenosis. J Am Anim Hosp Assoc, 2002. 38(6): p. 555-62.

7.  Lipsitz, D., et al., Magnetic resonance imaging features of cervical stenotic myelopathy in 21 dogs. Vet Radiol Ultrasound, 2001. 42(1): p. 20-7.

8.  Gopal, M.S. and N.D. Jeffery, Magnetic resonance imaging in the diagnosis and treatment of a canine spinal cord injury. J Small Anim Pract, 2001. 42(1): p. 29-31.

9.  Levitski, R.E., D. Lipsitz, and A.E. Chauvet, Magnetic resonance imaging of the cervical spine in 27 dogs. Vet Radiol Ultrasound, 1999. 40(4): p. 332-41.

10. Kippenes, H., et al., Magnetic resonance imaging features of tumors of the spine and spinal cord in dogs. Vet Radiol Ultrasound, 1999. 40(6): p. 627-33.

Speaker Information
(click the speaker's name to view other papers and abstracts submitted by this speaker)

Erik R. Wisner, DVM
School of Veterinary Medicine, University of California-Davis
Davis, CA

MAIN : Diagnostic Imaging : Diagnostic Imaging - Spine
Powered By VIN