This title appears in the Scientific Report :
2016
TIP: A diagnostic tool to create 3D deviation maps and brain surface views from F18-AV1451-PET data
TIP: A diagnostic tool to create 3D deviation maps and brain surface views from F18-AV1451-PET data
AbstractIntroduction Assessing quantity and distribution of pathological cerebral tau protein aggregation with PET in suspected tau associated neurodegeneration is currently in the scientific focus and may also become more common in clinical practice, particularly with regard to novel therapies dire...
Saved in:
Personal Name(s): | Hammes, J. (Corresponding author) |
---|---|
Bischof, Gerard Nisal / Giehl, K. / Drzezga, A. / van Eimeren, Thilo | |
Contributing Institute: |
Kognitive Neurowissenschaften; INM-3 |
Imprint: |
2016
|
Conference: | SNMMI 2016 Annual Meeting, San Diego, CA (United States), 2016-06-11 - 2016-06-15 |
Document Type: |
Abstract |
Research Program: |
(Dys-)function and Plasticity |
Publikationsportal JuSER |
AbstractIntroduction Assessing quantity and distribution of pathological cerebral tau protein aggregation with PET in suspected tau associated neurodegeneration is currently in the scientific focus and may also become more common in clinical practice, particularly with regard to novel therapies directed against tau pathology (Dani, Brooks, and Edison 2015). For [18F]FDG PET, semi-automated preprocessing for a standardized evaluation by means of 3-dimensional stereotactical surface projections (3DSSP) and z-score comparison to healthy control databases is in use for more than 20 years now (Minoshima et al. 1995). In analogy, we developed a handy preprocessing tool (Tau Image Processing, TIP) for F18-AV1451-PET data that provides an easy way to gain an overview of intensity and spatial distribution of pathological tau protein accumulation in an individual in comparison to a cohort of healthy controls. Methods Image data: PET scans were performed at the Department of Nuclear Medicine, University Hospital Cologne, Germany (Biograph mCT Flow 128 Edge scanner, Siemens, Knoxville, TN). Attenuation correction was based on a low dose CT, acquired prior to PET. 15 minutes of data acquisition in list mode started 90 minutes after application of 200 MBq F18-AV1451. Scans were iteratively reconstructed on a 128x128 matrix using a 3D OSEM algorithm of 4 iterations and 12 subsets (Smoothing: Gaussian filter of 5 mm FWHM). Avid Radiopharmaceuticals, Inc., Philadelphia, has provided a preexisting dataset of F18-AV1451 PET scans of 19 healthy controls. Both datasets were spatially normalized to MNI space (norm cohort). Images were intensity standardized to the cerebellum according to AAL-atlas (Tzourio-Mazoyer et al. 2002). Algorithmic design of PTI (Figure 1): Data are imported via a DICOM listener connected to a PACS. Files are converted to NIFTI format and spatially normalized to a mean image of the healthy controls using SPM 8 (Wellcome Trust Centre for Neuroimaging 2015) resulting in file dimensions of 79x95x68 isometric voxels. Images are intensity standardized to the mean cerebellar uptake. Then z-transformed voxelwise deviations from the norm cohort (3D deviation maps) are calculated. 3DSSPs are created from lateral, mesial and top and bottom views by creating maximum intensity projections of the first 10 voxels of brain-tissue from the respective viewing direction. A rainbow color table is applied for creation of RGB images from monochrome data. 3DSSPs with two different z-value thresholds and axial slices superimposed on a MNI standard MRI are saved in bitmap format. TIP runs in a Matlab R2015a environment (The MathWorks, Inc., Natick, MA). Spatially normalized and intensity standardized images as well as 3D deviation maps are stored in NIFTI format and are accessible for further research questions. Results The method requires only little user interaction and can be carried out by medical technologists in a standardized manner, potentially allowing the integration in daily routine work-up. Automated data processing proved to run stable without errors or interruptions in 50 test runs. Results obtained in 15 patients with different clinical diagnoses and suspected pathology were plausible and clearly correspond with the findings of unprocessed images as judged visually in transaxial, coronal and sagittal slices. Conclusions TIP is a helpful supportive tool for evaluating F18-AV1451-PETs in patients with suspected tau protein associated pathology. By providing an easy way for physicians and researchers to assess the spatial distribution and intensity of F18-AV1451 uptake in comparison to healthy controls our tool simplifies the diagnostic process and automates the data preprocessing for further scientific use of the PET data. |