Anatomic, differential diagnosis key to MRI success

Q. What are the pros and cons for magnetic resonance imaging (MRI) in dogs?

A. Dr. Laurent S. Garosi at the 2005 American College of Veterinary Internal Medicine Forum in Baltimore, Md., gave a lecture on use, misuse and abuse of MRI. Some relevant points in this lecture are provided below.

There has seen a dramatic increase in the availability of sophisticated neuro-diagnostic tests (computed tomography and MRI). Despite their relatively high sensitivity, these diagnostic tests often lack specificity in determining the exact nature of the disease process. It still stands true — a precise localization of the problem within the nervous system (anatomic diagnosis) and understanding of the suspected disease process (differential diagnosis) are the keys to successful management.

With unlimited media availability provided by the Internet, the general public has become more informed and sometimes more critical of what can or cannot be done for their beloved companion dog or cat. Consequently, the demand from owners for diagnostic modalities such as MRI has increased. MRI is now more readily available to the veterinary community and considered as an indispensable tool in many referral practices. Unfortunately, advances in therapy seem to lag behind the progress made in the availability of this diagnostic imaging technique. MRI is the preferred imaging method for humans with central nervous system disease. Its main advantages rest on its highly accurate soft-tissue resolution, multiplanar capability and absence of ionizing radiation risk when compared to other imaging techniques.

Numerous MRI radiofrequency pulse sequences have been designed in order to improve soft-tissue contrast resolution. The possibilities are nearly endless. The most common pulse sequences used to image the brain and spinal cord are those based upon the spin-echo. T1-weighted images are useful for visualizing anatomy. Gadolinium-enhanced T1-weighted images allow for identification of regions within the CNS where the blood-brain barrier is not intact. T2-weighted images are used for identifying regions of increased free water (regions of edema, cellular infiltration or inflammation). Other pulse sequences used for small animal MRI include fluid-attenuated inversion recovery (FLAIR) (used to suppress CSF signals in order to examine lesions of the brain parenchyma which are near to the ventricle or subarachnoid space), gradient echo (increased sensitivity of blood products and calcification), diffusion- and perfusion-weighted (use for early detection of infarction), and fat-saturation techniques.

Common MRI artifacts are important to recognize and include motion, susceptibility, signal void, partial volume and signal drop-off artifacts. MRI does require that the animal be anesthetized. Other than the risks associated with this general anesthesia (especially in animals with severely compromised brain function), there is no conclusive evidence for irreversible or hazardous bio-effects related to short-term exposures of humans to static magnetic fields up to 2.0 Tesla. When compared with other imaging techniques, MRI also has the benefit of the absence of ionizing radiation risk.

MRI is likely to add information to that obtained by conventional imaging techniques when:

1. The area of the suspected lesion cannot be evaluated using other means,

2. The information produced by conventional imaging techniques is limited, and

3. A lesion is evident on other imaging techniques but more 3-D information is required for example for treatment planning.

Looking into brain diseases

Because of its excellent soft-tissue resolution, MRI of the brain is indicated in the diagnostic work-up of animals with neurological signs of brain disease. Diseases that affect the brain are divided into a. extra-cranial disease (toxic or metabolic), b. intra-cranial structural brain disease (cerebrovascular, inflammatory or infectious, neoplastic, degenerative, anomalous, or trauma) and c. intra-cranial functional brain disease (mainly diseases caused by abnormal neurotransmission or ion channel disorders, such as primary epilepsy, narcolepsy or diseases classified as movement disorders). Elimination of extra-cranial causes of brain disorder is a prerequisite to MRI evaluation of brain disease. Signal intensity on MRI scans is a reflection of subtle biochemical and biophysical tissue properties. As a result, MRI has excellent soft-tissue contrast resolution and high sensitivity to many disease processes affecting the brain.

Previously considered uncommon, cerebrovascular accidents, and in particular brain infarcts, are increasingly recognized in dogs or cats with the advance of neuro-imaging. The improved resolution of imaging methods has made the identification of small pathologic process such as lacunar infarcts possible. These infarcts are now clearly the most common category of infarcts in dogs. Initial hopes that MRI signal intensity would be diagnostic for specific pathologic processes have not been borne out. Radiologists and neurologists have long tried to correlate histopathologic results and magnetic resonance signal intensity using combination of sequences to distinguish non-neoplastic from neoplastic disease and to further classify different types of neoplastic disease. The combined results of sequences might in some cases be strongly suggestive of a specific pathologic process. However, pathologic processes, such as inflammatory mass and neoplasia, share some MRI characteristics as well as similarities in origin, shape or anatomic site.

In many cases, the interpretation of the imaging findings rely on the clinical understanding of the suspected disease process (differential diagnosis), the use of other diagnostic tests (cerebrospinal fluid analysis, antibody titers, metastatic work-up, coagulation profile, etc.), a good understanding of the disease process mechanism, and/or histological diagnosis after tissue biopsy (ultrasonographic guided, stereotactic or surgical). Magnetic resonance techniques may still hold the key to non-invasive specific histological diagnosis. Magnetic resonance spectroscopy (MRS) provides metabolic information about brain tumors beyond what is obtained from anatomic images. Response to radiation therapy is also reflected by MRS patterns. MRS is also of use in the investigation of intrinsic metabolic disease where, while it is usually of insufficient sensitivity to measure the specific chemical responsible for the disorder, it can detect the secondary chemical pathologic changes. In some situations, these secondary chemical changes may be specific for the disorder.

The absence of abnormalities on MRI evaluation of the brain could indicate a. incorrect anatomical diagnosis (disease affecting parts of the nervous system other than the brain), b. toxicity or metabolic disease affecting the brain function without causing macroscopic parenchymal changes (hepatic encephalopathy, electrolyte imbalance or hypoglycemia), c. diseases causing very subtle parenchymal changes that might not be detected with a low-field magnet (small infarct or hemorrhage and degenerative disease), or d. functional brain disease (primary epilepsy and movement disorders).

Spinal cord diseases: common indication

Diseases affecting the spinal cord are broadly divided into compressive diseases (disk herniation, vertebral fracture/luxation, spinal malformation, neoplasm) and non-compressive diseases (spinal cord malformation, inflammatory/infectious CNS disease, degenerative diseases and vascular accident). Suspected spinal cord disease is a common indication for MRI in small animals. Traditionally, survey radiography and myelography have been the techniques of choice for the investigation of spinal cord disease. However, myelography provides no information about the spinal cord parenchyma other than whether it is compressed, displaced or swollen. MRI offers the double advantage of showing spinal cord parenchymal changes (such as those associated with inflammation, edema, syringomyelia-hydromyelia, infarct or neoplasia) and allowing transverse images (greatly assisting with the localization of the compressive tissue). MRI also has the benefit of being safer than conventional myelography as a subarachnoid injection is not required. Without an accurate anatomic diagnosis (suspected spinal cord segments affected), MRI evaluation of the spinal cord can be extremely time consuming, especially in large dogs, and can lead to over-diagnosis of incidental changes such as mildly herniated disks. Other limitations are mainly technical (magnetic field strength, field of view allowed, accurate positioning of the animal, and setting of different imaging planes).

MRI evaluation of the spinal cord could be normal in the case of a. incorrect anatomical diagnosis (particularly with neuromuscular disease, brain disease, or disease affecting other spinal cord segments than the ones imaged), b. diseases for which the parenchymal changes can only been seen microscopically (especially degenerative disease such as degenerative myelopathy), or c. diseases causing very subtle parenchymal changes that might not be detected with a low strength magnet (vascular disease such as fibrocartilaginous embolism or inflammatory disease of the spinal cord).

Looking into nerve and muscle diseases

The peripheral nervous system consists of 12 pairs of cranial nerves and 36 pairs of spinal nerves that extend from, or to, the spinal cord and brainstem. Diseases affecting the peripheral nervous system can affect one single peripheral nerve (mononeuropathy) or multiple peripheral nerves (polyneuropathy). Common indications of MRI in investigation of peripheral nerve disease include the investigation of a. cauda equina syndrome (very often technically difficult using myelography and epidurography), b. brachial plexopathy, and c. cranial nerve neuropathy. The main limitations of MRI in these cases are mainly technical (choice of adequate imaging plane and sequences), operator-dependent (detailed knowledge of anatomy), and inherent poor specificity in differentiating neoplastic from non-neoplastic disease processes.

Electromyography is better than MRI for muscle disease evaluation, and a muscle biopsy is the only manner in which a definitive diagnosis can be made. However, MRI has proven a useful adjunctive diagnostic procedure for cases of polymyositis where diagnosis is elusive, particularly if a specific anatomic localization cannot be reached by other means. The MRI signal intensity of normal muscle is intermediate between that of fat and cortical bone. MRI studies in dogs have shown a good correlation between uniform hyper-intensity on T1-W and T2-W images showing enhancement after contrast medium administration and muscle inflammation identified histopathologically.

Vascular-related neurological disorder

In addition to stroke, vascular-related neurological disorders include aortic thromboembolism (ischemic neuro-myopathy), congenital portosystemic shunt (hepatic encephalopathy) and other vascular malformations (such as caudal vena cava malformation) as a possible cause of exercise-induced weakness and collapse. In addition to its use for tissue evaluation, MRI (i.e., magnetic resonance angiography or MRA) can non-invasively assess the vascular system. Two techniques can be used: 1. time of flight (TOF) MRA and 2. contrast MRA. The TOF MRA can be implemented on every MR system, is easy to use, and do not need contrast medium injection. TOF may be used in either a 2-D (sensitive to slow-flowing blood) or a 3-D sequence (high spatial resolution and sensitive to fast-flowing blood). Contrast MRA is based on the use of paramagnetic contrast agents in combination with gradient-echo sequences. Paramagnetic contrast media strongly enhance the MR signal due to T1 shortening. Image data can be collected during the whole measurement cycle, demonstrating arterial and venous contrast passage. In all MRA techniques, image contrast is the result of blood motion. Although less invasive, one of the main limitations of MRA is its lower resolution compared with conventional angiography that becomes progressively worse as the vessel luminal size decreases.