All that Glitters is not Gold

Due to the length of time it takes to acquire images and the magnetic manipulation of local fields, there are a lot of artifacts that can arise. It is important to recognise these and separate them from true findings.

 

Motion artifact

MRI takes a long time to acquire with each sequence taking several minutes. During this time the patient has to stay absolutely still. You can imagine that for some patients this will be difficult e.g. confused, seizing, unable to follow commands. A small amount of motion can be corrected for but significant motion can make images uninterpretable.

mod1_motion_artifact

Pulsation artifact

Another artifact related to motion is the pulsation artifact in which pulsating structures such as vessels and CSF can create abnormal signal.

mod1_pulsation_artifact

Pulsation artifact and flow void can cause apparent lesions in the CSF that are dark on T2WI and bright on FLAIR. This is much more of a problem in the spine but can be seen intracranially in the prepontine cistern.

mod1_artifact_T2_FLAIR

Metal artifact

Metal distorts the local magnetic field and no signal can be returned from the surrounding area. This is particularly a problem with orthodontic braces and dentures as well as metallic surgical implants. Also, make up containing metals, such as some mascaras or permanent make up, can cause artifact.

In the below image, the patient has a metal skull implant from previous surgery. You can see how there is loss of signal in the anterior skull. The worst affected sequence is DWI.

mod1_metal_artifact

Other local field artifacts

Air can act to disrupt the local field and, particularly at air-bone interfaces, it can create significant artifact. DWI is particularly susceptible to this artifact.

mod1_DWI_artifact

Partial voluming

This is related to the slice thickness acquired. If the slice is thick an object can partly, but not fully, be imaged on an adjacent slice. The signal is averaged out and it can look like there’s a lesion. The best way to overcome this is to view adjacent slides. Is there a structure on one slice in approximately the same location? If yes, do you think this could be what has been partly imaged?

mod1_partial_voluming

Tips

  • if a lesion crosses multiple structures – e.g. subarachnoid space, parenchyma, bone – it is likely artifactual as barely anything does this in real life.
  • scrolling up and down will help to reduce the partial voluming effect.
  • metal, as well as being dangerous, can cause significant artifacts.