Dear Michael,
1, I'm interested in only the normalized brain volume, and I thought that the two class segmentation is more robust:)
This idea was based on an article "http://www.unboundmedicine.com/evidence/ub/citation/18777529/Agreement_between_different_input_image_types_in_brain_atrophy_measurement_in_multiple_sclerosis_using_SIENAX_and_SIENA_" , which demponstrates that contrast enhanced spin-echo is as good as native spine-echo T1-weighted images to assess normalized brain volume. The authors also used a 2-class segmentation in the article.
2, How can I use sparial priors in SIENAX?
3, Betting the T2 image is really hard and finding an appropriate skull image is even harder. That's why I'm trying to use the post-contrast MPRAGE.
4, If I use sienax with 3 classes most of the veins are segmented as CSF. I suspect that the same thing happens using a native MPRAGE.
5, Do you know any article using SIENAX on postcontrast MPRAGE?
Thanks,
Gabor
|