waveletsinpetmodelling-turkupetcentre.pptVIP

Applications of wavelets in PET modelling  - a literature survey Hammersmith, London Turkheimer et al. 1999-2000 3 articles of the subject theoretical framework of wavelets image reconstruction and image processing Dyadic wavelet transform DWT and translation invariant DWT DWT-TI techniques applications to dynamic PET-SPECT studies Parametric images: kinetic and statistical modelling in wavelet space Denoising images with wavelets Wavelet transform of dynamic image Thresholding e.g. SURE, universal, Bonferroni Inverse transform Studies Low-resolution SPECT study Hoffman brain phantom DWT vs. DWT-TI in 2D High-resolution PET images FDG brain study universal thresholding vs. SURE thresholding Very high-resolution PET images [11C]PK11195 study of peripheral benzodiazepine receptors in the brain high signal-to-noise ratio Studies Synthetic dynamic study artificial data set with nonstationary noise field FDG dynamic PET study Patlak plot [11C]raclopride PET dynamic study D2-receptor distribution in normal brain Logan plot vs. WT Studies Randomization study null dataset with PET H215O activation/rest - two groups of simulated datasets Parametric study of cerebral blood flow response to word recognition 5 right-handed normal subjects Measuring the effect of depression on brain serotonin receptors with [11C]WAY-100635 WT vs. statistical parametric mapping Karolinska hospital, Stockholm Cselenyi et al. 2002 Binding potential BP study with [11C]FLB 457 4 methods traditional ROI analysis reference data pixel-based analysis 2 variants of wavelet-aided analyses Aim is to decrease the noise-sensitivity of a parameter estimation procedure with wavelet approach 10 healthy male subjects ROI analysis BP was estimated using the reference region version of Logan’s graphical analysis reference region cerebellar cortex Pixel-by-Pixel analysis radioactivity of a pixel area under the curve AUC of the corresponding TAC Fitting is done with the same Logan analysis