Rodney S. Bagley, DVM, DACVIM (Neurology and Internal Medicine)
The use of magnetic resonance (MR) imaging has revolutionized the diagnosis of animals with neurologic disease as well as other body system abnormalities. Access to these modalities for diagnostic imaging in animals has been a historical stumbling-block, however is slowly becoming less of an obstacle as MR imaging facilities dedicated to imaging animals are established. At Washington State University, College of Veterinary Medicine, we have been using MR imaging routinely for diagnosis of animals with neurologic disease since 1996. During this time over 5000 animals have had either intracranial or spinal imaging as part of their diagnostic evaluation. Information gathered from these patients using this imaging modality has lead to significant improvements in understanding of the anatomic and physiologic ramifications of central nervous system disease, and, subsequently, to improvements in treatments of animals with neurologic disease.
Currently, MR must be considered the "gold-standard" for imaging the central nervous system. The anatomic detail and spatial resolution is significantly superior to other modalities. These factors yield an image equal to that visualized at gross dissection thus allowing a rapid familiarity of the diagnostician to these images. Newer pulse sequences have been developed that allow further anatomical clarification of abnormalities of the nervous parenchyma.
Clinical Use of MR Imaging
Unlike computed tomographic (CT) imaging, MR imaging is a noninvasive procedure that does not require ionizing radiation. The physics associated with MR imaging is complex and have reviewed elsewhere. While this information is helpful in determining optimal imaging techniques, it is not necessary for most clinicians to have this knowledge in daily clinical practice.
Numerous excitation and detection sequences are available. The various sequences provide images of the proton density, and the two basic forms of relaxation, named longitudinal (T1) and transverse (T2). The most commonly used imaging sequence to date has been a spin-echo sequence yielding a proton density and a T2-weighted image followed by a T1-weighted sequence before and following the administration of a paramagnetic contrast agent. The most commonly used contrast agent to date utilizes the paramagnetic material gadolinium (Gd) in the form of Gd-DTPA. This contrast agent acts in a matter similar to the iodinated contrast agents utilized in CT. The main difference is that MR does not visualize the gadolinium but produces an image of the protons that are influenced in their relaxation by the presence of Gd. This material does not pass the intact blood-brain-barrier (BBB), and therefore, the only enhanced areas in the normal brain are those that lack a blood-brain barrier including the pituitary and choroid plexus. Most diseases of the CNS disrupt the BBB resulting in a high signal intensity in the presence of the contrast agent.
Magnetic resonance imaging initially was used for diagnosis of animals with intracranial disease. As compared to CT scanning, MR provides superior anatomical detail of the intracranial nervous system and associated structures such as cranial nerves. A variety of structural intracranial diseases are readily observed on MR imaging. These include brain tumors, hemorrhage, hydrocephalus, and encephalitis.
Diagnosis utilizing MR are based on mass effect causing displacement of normal structures and the visualization of contrast-enhanced areas similar. Loss of symmetry, changes in signal intensity, and displacement of normal structures are extremely helpful in arriving at a diagnosis. The improved resolution afforded by MR imaging often allows determining whether lesions are intra- or extra-axial in origin. Extra-axial lesions involve tissue outside the neuronal axis, including meninges, pituitary, and choroid plexus. The location of lesions helps determine potential tissues of origin, and the shape of the mass and visualization of its edges is often useful in describing the degree of malignancy. In general, signal intensity is increased with tumors, especially on the T2-weighted images. Depending on the image sequence, however, tumors may be hypo-intense, iso-intense, or of a heterogeneous intensity as compared to normal surrounding tissue. Extra-axial tumors are often difficult to visualize without contrast enhancement. Meningiomas, especially, are often iso-intense on the proton density, T1-, and T2-weighted images. However, these tumors usually show marked enhancement following the administration of GD-DTPA.
Because non-neoplastic lesions can mimic the appearance of tumors biopsy of the abnormal tissue is often required for determination of appropriate therapy. Hydrocephalus is visualized on all imaging sequences, and any obstructions to flow can be readily detected. In addition, MR is extremely useful to visualize edematous changes within the CNS in the absence of contrast enhancement. Magnetic resonance is extremely sensitive at depiction of hemorrhage, and due to the normal degradation of hemoglobin, the age of hemorrhagic occurrence can be estimated. Magnetic resonance provides better detail and more information than CT. Computed tomography, however, is superior in detecting mineralization and bony changes.
Historically, diagnosis imaging of the spine as centered on spinal radiographs for vertebral imaging and myelography for evaluation of the subarachnoid space. Myelography is an invasive procedure requiring a spinal tap and subarachnoid contrast administration. Complications of this procedure include iatrogenic spinal cord injury during spinal needle placement, seizures, and a mild central nervous system inflammatory reaction associated with the contrast agent administered. Another significant disadvantage of myelography is the inability of this procedure to imaging the spinal cord itself. As numerous spinal cord diseases occur primarily within the spinal cord, these diseases may potentially be missed myelography is used for diagnostic spinal imaging. While computed tomography has advantages over myelography in provided a transverse imaging plane, anatomical detail within the spine is less obvious than with MR imaging. Magnetic resonance imaging, therefore, has become the "gold standard" for diagnostic imaging of the spinal cord in our hospital.
Spinal MR imaging provides images of vertebral bodies, intervertebral disks, spinal cord, exiting spinal nerves and paraspinal anatomy. Sagittal images are initially performed to evaluate the spinal cord. Abnormal areas are re-imaged in a transverse plane, for further clarification. Cortical bone is dark or hypointense, on both T1- and T2-weighted images. The MR appearance of the vertebral column depends mainly on the signal from bone marrow. Medullary bone, rich in fat and blood, is moderately intense (gray) on the T1- and less intense on the T2-weighted images. The ultimate appearance of marrow, however, depends upon the composition of the numerous components that may be present. The spinal cord tends to be isointense to the medullary bone of the vertebral bodies in the T1-weighed study and is typically surrounded by epidural fat which is hyperintense to the spinal cord. This hyperintense outline can be seen on the T1-weighted sagittal and axial images. Epidural fat is easily displaced and hence may not be visible in conditions that result in spinal cord swelling. The intervertebral disks on T1-weighted images are slightly hypointense to isointense to the medullary cavity of the vertebral bodies.
The spinal cord on T2-weighted image is also isointense to the vertebral bodies, however the CSF surrounding the spinal cord emits a bright white (hyperintense) signal. By intensifying the CSF signal on a T2-weighted study, a "myelogram effect" can be achieved. This, in essence, provides for an exclusive image of the subarachnoid space. The normal anatomy of the intervertebral disk is best appreciated on a T2-weighted sagittal MR image. Due to its high water content, the normal nucleus pulposus is bright (hyperintense) while the outer layers of the normal annulus fibrosis are homogeneously dark (hypointense).
Following a neurologic examination, a generalized lesion localization is made and imaging is performed in the area of the suspected lesion. T1- and T2-weighted sagittal images are performed to evaluate for signs of an extradural compression of the spinal cord. Disk degeneration is easily seen on the T2-weighted sagittal and transverse (axial) images of the spinal column. Normal hydrated nucleus pulposus has a hyperintense signal compared to the anulus fibrosus. As the nucleus pulposus loses hydration, (seen in chondroid metaplasia or Hansen type I disk disease), the signal becomes less intense and may appear iso- or hypointense to the anulus fibrosus. Most acute disk extrusions of this type appear as a hypointense lesion (disk) on the T2-weighted sagittal image. The T2-weighted sagittal images may also be used to create the "myelogram effect" which is of use in evaluating compromise of the dural sac by the disk (an extradural compression). This is rarely necessary, however, during routine evaluation. Once a lesion is identified on the sagittal image, a transverse image is performed over the area for further clarification of asymmetry.
Syringomyelia and hydromyelia are diseases where there is abnormal fluid cavities within the spinal cord. These cystic spinal disease may occur congenitally, due to trauma, or from tumor. Magnetic resonance imaging is superior in the detection of these intramedullary processes compared to myelography or even CT.
On T1-weighted images, most intramedullary neoplasms have diminished signal intensity with respect to the cord. On T2-weighted images, they usually have a brighter signal than cord. Most tumors have a non-homogenous signal intensity and indistinct margins between tumor and surrounding normal cord. Tumors are typically vascular and invasive, disrupting the normal integrity of the blood brain barrier. Thus, intravenous injection of a contrast agent generally results in some degree of either diffuse or focal enhancement in the area affected by tumor. Contrast enhancement does not precisely define the tumor borders; neoplastic cells are generally found outside the enhanced portion of the mass.
Although contrast enhancement can not identify the type of tumor, particular enhancing characteristic have been seen with various tumors. Intradural/extramedullary tumors may be more difficult to identify because these tumors may have little contrast with respect to the adjacent spinal cord. Sagittal images may not detect these tumors as they are often lateral to the spinal cord. Intravenous gadolinium-DTPA (Magnevist, Berlex Laboratories, Cedar Knolls, NJ), enhances these tumors by increasing their signal intensity by 200-500% on T1 weighted images.
T2-weighted images are more sensitive than T1- weighted images for the detection of both spinal cord and paraspinal soft tissue injury. Despite its inability to delineate bone injury, MR clearly establishes the presence of associated acute spinal pathology. When examination with MR is abnormal, the area affected may then be accessed by CT if further clarification of bone involvement is necessary.
Imaging artifacts occur with MR as in all modalities. The main causes of artifacts in images of the CNS are generally motion and deformations in the magnetic field from implanted metallic objects. Since the time required for MR studies is 10 to 60 minutes, general anesthesia must be utilized since MR generates considerable noise that would startle animals. If gas anesthesia is used, a high flow rate, non-rebreathing system allows the standard anesthetic equipment to remain safely outside of the strong magnetic field. With general anesthesia, the lack of patient motion yields excellent animal studies. Rarely do animals have any implanted metallic objects that are magnetic enough to perturb the image field, however micro chips and other steel objects (BB pellets) can create artifacts with MR that affect image quality. The most important technical factor that influence the degree of image misinterpretation is poor animal positioning and associated alterations in slice plane direction.