CT head (an approach)

Last revised by Daniel J Bell on 13 May 2023

The approach taken to interpreting a CT scan of the head is no doubt different depending on the circumstances and the reading clinician, however, most radiologists will go through the same steps, although order may vary. What follows is merely a suggested approach to interpreting a CT of the head

An important aspect of this approach is the use of windowing to maximize pickup rate. With PACS, windowing the scan to optimize detection of certain pathology is simple and should be routine. For example, reviewing every scan using not only a brain window but also a bone window, subdural window and stroke window is strongly recommended and will result in reduced miss rate.

People have surprisingly strong opinions on this. It is important to look at all the images and ensuring careful review of slices at the very top or bottom. It has been argued by many that starting at the top can be helpful in that it gives you the first few images to decide which side has mass-effect and then work your way down to the "busy" slides at the bottom.

The reality is that with PACS you are likely to scroll up and down a number of times: a quick scroll through the whole volume will give you a sense of any obvious major pathology.

Additionally, most institutions will not just have axial images but also routinely generate sagittal and coronal images as, increasingly, scans have been obtained volumetrically (see CT head technique). Alternatively, you may prefer to review the whole volume in a multiplanar viewer allowing adjustment of the plane according to the structure under review.

What follows is merely a suggestion, but one that will ensure that you cover everything. Performing the same search patterns every time until it is a habit is a powerful tool for anyone looking at imaging and reduces the chances of forgetting to look at part of the scan. This is particularly important when interrupted midway through review. 

It is good practice to get into the habit of looking at the scout radiograph(s) before reviewing the main imaging stack(s). Occasionally abnormal skull morphology, fractures and previous neurosurgical procedures are easier to appreciate on the scout than the main stack, especially if the center still scans using a step and shoot protocol precluding MPRs.

  • window: window width (W): 80 HU window level (L): 40 HU

Start at the vertex. Look at the brain matter and gyral pattern. Are there any masses, abnormal attenuation or mass effect?

Without changing the window setting, review the basal cisterns and sulci. Are they visible and commensurate with the patient's age and ventricular size? Do they extend all the way to the bone? If not is there a subacute (isodense) subdural hemorrhage? Is there any subarachnoid blood? Look in the dependent areas for any small volume subarachnoid hemorrhage particularly in the interpeduncular fossa or occipital horns of the lateral ventricles. Next look at the vessels. Is there a hyperdense MCA or basilar tip hyperdensity suggesting thromboembolism? 

  • window: W:8 HU L:32 HU or W:40 HU L:40 HU

Now search for subtle cortical hypoattenuation by assessing grey-white differentiation. The cortical grey matter should be readily differentiated from the white matter. The deep grey matter should be distinct from the white matter also. Pay particular attention to the insular cortex and basal ganglia as they are lost early in middle cerebral artery infarction.

  • window: W:130-300 HU L:50-100 HU

Next, concentrate specifically on reviewing for small collections of subdural or extradural blood that may have blended with the skull on routine windowing. Coronal reformats if available are particularly useful for this. Look for tiny volume subdural hemorrhage around the tentorium and circumferentially as thin hyperdense crescents.

  • window: W:350–400 HU L:20–60 HU

Do not forget that a CT of the head does not just look at the brain. Soften the window to something suitable for soft tissues and especially in the anterior face including the globe, sinuses, palate and pharynx. Check the scalp for focal lesions or hematomas. 

  • window: W:2800 HU L:600 HU

Now, and lastly, review the bones. This should always be performed, even when a bony algorithm has not been provided or where slice thickness is suboptimal. Note that if there is a history of trauma, then dedicated thin bony images are helpful to detect undisplaced fractures. 

Review the skull vault for any fractures or destructive lesions. Spend some time checking the base of the skull as the increased complexity of this region can make identification of abnormalities more difficult. Do not forget to ensure that both TMJs are normally aligned. 

Review the paranasal sinuses for evidence of fluid that may represent acute sinusitis or, in the correct setting, fractures. 

Review areas are different for everybody and you will develop your own list over time, usually as the result of misses, but some common review areas for a CT head include:

  • tentorium

  • orbit and globe

  • retropharyngeal space

  • foramen magnum

  • middle ear

  • paranasal sinuses

Something that is worth remembering is that just because you looked at it does not mean you have to include it in the report. Exactly what your standard normal CT head ends up being will be dependent not only on your training and personality but also on your referrers and their clinical question(s). See CT head standard report.

ADVERTISEMENT: Supporters see fewer/no ads

Updating… Please wait.

 Unable to process the form. Check for errors and try again.

 Thank you for updating your details.