 |
 |
Likewise, we have had some success in using the LST toolbox for the same non-MS hyperintensity volumetric analysis, but as Dr. Gaser notes one needs to tailor the LST settings to fit the sequence and data characteristics. We did this iteratively to find the optimal settings for our data by comparing against the kappa obtained in the dice-coefficient output for each setting variation. Warm regards, Jeff > On Apr 28, 2020, at 2:19 AM, Christian Gaser <[log in to unmask]> wrote: > > Dear Aurelia, > > The LST toolbox was primarily designed to detect lesions in the white matter in MS patients. However, we have also successfully used LST to detect white matter hyperintensity (WMH) in healthy controls and AD patients. If the LGA approach with multiple thresholds is not suitable, you can also try the LPA approach, which only works with FLAIR images. However, it also depends on your sequence and data (e.g. the severity of the lesions) whether the LST approach works well with data from dementia patients. > > Best, > > Christian > > On Mon, 27 Apr 2020 17:40:29 +0100, Aurelia Morabito <[log in to unmask]> wrote: > >> Dear SPM users, I am trying to use the Lesion Segmentation Tool (Schmidt et al, 2012) to segment lesions in FLAIR on patients with dementia. I have tried to use the LGA (Lesion Growth Algorithm) setting several different thresholds, but the results are not as good as they are in segmenting lesions on patients with multiple sclerosis. I would like to know if any of you have ever used this toolbox for segmenting this kind of lesions, and in this case, if some pre-processing of the image is needed in order to have a better result. Otherwise, I would like to know if anyone knows other toolbox that may be useful for my aim. Thank you very much, Aurelia Morabito