I am given different CT perfusion maps like cerebral blood flow (CBF) or mean transit time (MTT). Given these 3D images, I have to produce predictions using machine learning algorithms.

Can you explain how these perfusion maps are produced (input and output)? I just need a general overview from a medical POV.

Is my understanding correct?

  1. inject contrast material in the patient
  2. perform CT scan, the output is a 4D dimensional image with x, y, z and time axis. The time axis shows how the blood flows during the k seconds of the scan.
  3. reduce the 4D image into a 3D image with some computations to produce MTT, CBF, ...
  4. The medical practitioner analyzes the different maps (MTT, CBF, ...) and looks at different window settings (stroke window, brain window, ...)
  • 1
    Don't they do perfusion maps mostly with radio-pharmaceuticals rather than contrast materials, and PET/SPECT scans?
    – einpoklum
    Apr 16, 2021 at 9:22
  • According to some public dataset "To assess cerebral perfusion, a contrast agent (CA) is administered to the patient and its temporal change is captured in dynamic scans acquired 1-2 sec apart. " (see isles-challenge.org). Wikipedia says "injection of a bolus of iodinated contrast material as it travels through the vasculature". But I have actually no expertise in this area (how contrast material works etc.). Any comments are welcome :) Apr 16, 2021 at 18:19
  • 1
    If you do a CT, then yes, you use a contrast material, not one chosen for its radioactivity - because CT is an active scan (i.e. camera emits at object). If your scan is passive (object emits, camera only receives) then a contrast material is meaningless.
    – einpoklum
    Apr 16, 2021 at 21:27


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.