Limitations of CT in assessment of traumatic brain injury

This issue has come up in  Question 19 from the second paper of 2005. Why would it be a bad idea to scan your neurosurgical patient's head? Well, turns out there are numerous issues, which can be grouped into four major categories. There are logistical problems (getting to the scanner and back in one piece), machine problems (eg. contrast and radiation exposure), interpretation problems (how well can you read the scan is important to the utility of that scan) and concerns regarding relevance (i.e. was the CT really necessary, will it change management?)

Logistical limitations:

  • The investigation requires transport to and from the CT scanner
  • It involves a degree of cooperation, and may require sedation
  • Not all patients will fit into the 84cm aperture or on the table which has a 220kg fully extended loading maximum

Machine limitations:

  • It involves radiation exposure
  • It may involve contrast exposure
  • The images of posterior fossa structures are usually poor
  • Diffuse axonal injury will not be visualised
  • On a non-contrast study, you will not see arterial dissection or vascular insufficiency
  • Infarcted regions cannot be visualised early in the infarct
  • Hypoxic injury cannot be seen early in the injury
  • Artifact from metallic implants (eg. crowns, aneurysm clips, scalp staples) will obscure the view
  • CT may miss small amounts of blood which occupy the space between slices, because of image averaging
  • Where it comes to subtle neuronal and axonal injury, or to petechii, CT misses 10-20% of pathology seen on MRI

Interpreter limitations:

  • The CT needs to be interpreted by a competent radiologist for the findings to be valid
  • Subtle features and posterior fossa injuries may be missed

Relevance to clinical setting

  • The CT is an assessment of structure, rather than function
  • Early CT may underestimate the extent of an evoving injury
  • It may not be possible to get the CT done because some sort of damage control surgery takes precedence.
  • CT findings may not correlate with ICP, unless you are coning.
    • This is debated. Some people swear by their grey-white junction findings and their basal cistern size.
    • Certainly, a non-raised-ICP-looking CT should not prevent you from inserting an EVD if there are clinical indications for it.


Lee, Bruce, and Andrew Newberg. "Neuroimaging in traumatic brain imaging." NeuroRx 2.2 (2005): 372-383.

Miller, M. Todd, et al. "Initial head computed tomographic scan characteristics have a linear relationship with initial intracranial pressure after trauma." Journal of Trauma-Injury, Infection, and Critical Care 56.5 (2004): 967-973.