Dear Pravesh,
I really like CAT12 for T1w-based segmentation. If you need a
bias-corrected T2/T2FLAIR you may try the SPM12 routines, such as
multimodal segment etc.
In the past I remember some nice results from this (at least for
whole-volume intensity correction and normalization). Tissue
segmentation based on T2w/T2FLAIR will always be inferior compared to
T1w as Christian already stated.
Cheers,
Barbara
On 09/02/2018 17:30, Christian Gaser wrote:
> Dear Pravesh,
>
> T2 or Flair images will be always internally intensity-corrected in CAT12 to get the right intensities w.r.t. a T1 image. The internal CAT12 segmentation approach can only deal with T1 images and there is no way to skip that step. I would strongly recommend to prefer T1 images, because the contrast-to-noise ratio (CNR) is better and results in more accurate segmentations.
>
> Best,
>
> Christian
>
> On Thu, 8 Feb 2018 11:18:22 +0530, Pravesh Parekh <[log in to unmask]> wrote:
>
>> Dear SPM/CAT Users,
>>
>> We tried segmenting some T2 weighted images in CAT12.1 and noticed that
>> after normalization, the tissue contrast gets flipped back to T1 like. Why
>> should this happen and can this be corrected within CAT? Perhaps CAT is
>> making use of some T1 template in the processing pipeline? Attached is a
>> snapshot showing the raw approximately ACPC aligned T2 data in the top pane
>> and the the normalized modulated [wm file] data in the lower pane.
>>
>> In general, CAT seems optimized for processing T1 images. Is it okay to
>> consider segmentation of other contrasts in CAT or is it not recommended in
>> general? We tried processing a T1w and T2w data (for the same subject from
>> the same session) in CAT and SPM. Below are the volumetric estimates:
>>
>> *For T1 image:*
>>
>> *TIV*: 1530.40 (CAT); 1439.78 (SPM)
>> *GM*: 664.81 (CAT); 682.94 (SPM)
>> *WM*: 490.48 (CAT); 418.38 (SPM)
>> *CSF*: 375.11 (CAT); 338.46 (SPM)
>>
>> *For T2 image:*
>>
>> *TIV*: 1510.65 (CAT); 1427.68 (SPM)
>> *GM*: 673.87 (CAT); 632.48 (SPM)
>> *WM*: 547.05 (CAT); 441.97 (SPM)
>> *CSF*: 288.94 (CAT); 353.23 (SPM)
>> *WMH:* 0.79 (CAT)
>>
>> We also noted that CAT reported WMH consistently for a bunch of T2w data
>> (all from the same healthy subject) with estimates of WMH ranging from 0.12
>> to 120.15!
>>
>> Thank you in advance for your insights and thoughts on these.
>>
>>
>> Regards
>> Pravesh Parekh
>>
|