Neurobites – Evaluation of MRI


Every week, the BVNS neurologists, residents and interns convene to discuss a human or veterinary neurology/neurosurgery article.

NeuroBites is a digestible synopsis written by Dr. Bush of the article covered in journal club.

Welcome to NeuroBites.

Masciarelli AE, Griffin IV JF, Fosgate GT, et al. Evaluation of magnetic resonance imaging for the differentiation of inflammatory, neoplastic, and vascular intradural spinal cord diseases in the dog. Vet Radiol Ultrasound.

Click HERE for the full article.

Dr. Rivera picked a very important article today which discussed the sensitivity and specificity of MRI. The article helped us to understand the strengths and potential pitfalls of using MRI to provide a diagnosis and thus prognosis and plan for our owners. The article centered around the ability of MRI to detect spinal cord neoplasia, inflammation or stroke from dogs with degenerative myelopathy (normal MRI group). To be included, a microscopic post mortem exam of the spinal cord was required.

Sensitivity – the ability of a test to detect an individual with disease

Specificity – the ability of a test to detect an individual without disease

Imagine a flea in is in the hair of a dog and moving a flea comb through the dog’s hair is a test. The sensitivity would be the ability of the comb to identify the flea. If every little dark piece of crud was called a flea then all dogs would have fleas when some actually might have dirt. If you called all dark pieces of crud a flea then the test would be 100% sensitive. However, if the test was interpreted in this manner then it would have low specificity because it would identify some dogs as flea infested when actually they are just dirty.

How sensitive is MRI for identifying disease in dogs (meaning how good is it at saying a dog has spinal cord disease relative to degenerative myelopathy)?
A) 40% – owners are wasting their money at BVNS
B) 70% – pretty good, but about 1 in 3 dogs have disease which is failed to be recognized on MRI
C) 98% – MRI is amazing for finding spinal cord disease
D)100% – MRI is perfect and even a monkey could do my job

Here are some interesting facts:

  • MRI is 98% sensitive in finding spinal cord disease
  • MRI is only 50% sensitive for inflammatory disease but this improves to 80% when the results of the CSF tap is included – this is why we do so many CSF taps in conjunction with MRI
  • MRI was 69% specific – meaning it correctly identified the disease as normal, neoplasia, vascular or inflammatory about 2/3 times
  • Specificity improved to 82% when the reader knew the patient’s history
  • Vascular lesions often masqueraded as normal – the sensitivity was only 25%, so do not be surprised if you think a dog with a spinal stroke has a normal MRI

Atlanta, GA
(678) 400-0042
Leesburg, VA
(703) 669-2829
Richmond, VA
(804) 716-4716
Rockville, MD
(301) 637-4248
Springfield, VA
(703) 451-3709