Print

Print


Hi,

Using FAST to segment small pathologies is difficult since FAST is based on histograms and these are fairly badly insensitive to small volumes.  You can try changing the MRF options and the number of iterations in FAST and maybe also using priors, but I'm not sure that it will fix the problems.  In general we would recommend that you use BIANCA for such a problem, as it was designed to find small pathologies.  However, it does require a training set of data, where manual masks have been drawn for the images.

All the best,
Mark


On 14 Aug 2017, at 20:29, Rockers, Elijah D. <[log in to unmask]> wrote:

Hello FSL,

I am looking at hypointensities on T1 structural images with white matter disease. I have used Freesurfer/fslmaths to come up with a nice mask that includes all the white matter and ventricles, (basically everything inside Freesurfer's white matter surface) but excludes the striatum (intentionally)

So the masked region contains normal white matter, hypointesities, and CSF. I am interested in quantifying the volume of hypointensities. If I use the masked region as an input to FAST, it seems to do a relatively good job segmenting the hypointensities, but voxels near the cortex are clustered with the hypointense class, and some of the more hypointense parts of the tissue are classified with the CSF.

Is there some way to adjust or improve this useage of FAST, so the classification is more accurate?

Thanks,
Eli

Houston Methodist. Leading Medicine.

U.S. News & World Report has recognized Houston Methodist as one of the nation’s top 20 hospitals, placing it for the second time on the magazine’s prestigious Honor Roll. It is also designated as a Magnet hospital for excellence in nursing. Visit us at houstonmethodist.org. Follow us at twitter.com/MethodistHosp and facebook.com/HoustonMethodist.

***CONFIDENTIALITY NOTICE*** This e-mail is the property of Houston Methodist and/or its relevant affiliates and may contain restricted and privileged material for the sole use of the intended recipient(s). Any review, use, distribution or disclosure by others is strictly prohibited. If you are not the intended recipient (or authorized to receive for the recipient), please contact the sender and delete all copies of the message. Thank you.