SPM technique for DECIDE (Diagnostic Enhancement of Confidence by an International Distributed Environment ) Isabella Castiglioni – Institute of Molecular Bioimaging and Physiology (IBFM) National Research Council (CNR), Italy DECIDE coordinator of JRA
Assisted diagnosis assist physicians in the interpretation of medical images Diagnostic Enhancement of Confidence
Computer-Aided Assisted (CAD) diagnosis -Probability of a specific disease CAD -biomarker value -characteristic pattern -…
Computer-Aided Assisted (CAD) diagnosis SPM: Statistical Parametric Mapping Maps of statistically significant hypo/hyper metabolic regions obtained by functional neuroimaging studies (es. 18 FDG-PET) -characteristic metabolic pattern associated to a -probability e.g.“subject has a 95% chance of having that pattern.” • developed by the Wellcome Trust Center for Neuroimaging (London, UK) in 1995 Friston KJ Holmes AP Worsley KJ Poline JB Frith CD and Frackowiak RSJ. (1995b) Statistical Parametric Maps in functional imaging: A general linear approach Hum. Brain Mapp. 2:189-210 • free available (software package) to the scientific community since 1996 (spm96-spm2008)
SPM: Statistical Parametric Mapping
Analysis: voxel by voxel
SPM and 18 FDG PET for neurodegenerative studies Comparison of N different conditions of a single subject 1. Health subject 1 1. health NORMATIVE GROUP 2. health 2. Health subject 2 3. health 3. Health subject 3 4. health 4. Health subject 4 . . Hypothesis real . . . . . CONDITIONS CONDITIONS . . . N . . N-1. health N-1. health subject N-1 N. pathological N. pathological
SPM and 18 FDG PET for neurodegenerative studies Covariates of interests: Normal aging hypometabolism AGE SEX Smaller brain, global lower metabolism
SPM and 18 FDG PET for neurodegenerative studies t-test: single subject (suspected Alzheimer’s disease) vs group of normal subjects vs Hypo-metabolism Normal metabolism Analysis: voxel by voxel Results: Maps of statistically significant hypometabolic regions General Linear p<0.05 Model Statistical inference
SPM and 18 FDG PET for neurodegenerative studies Coregistration to Original images Standard atlas standard atlas (Elestic deformation + affine traslation) Spatial rialignement +normalization
SPM and 18 FDG PET for neurodegenerative studies Spatially-normalized images Smoothed images Smoothing
SPM and 18 FDG PET for neurodegenerative studies Statistical design matrix Patient images Statistical parameter maps Genera l Linear Statistical inference Spatial Smoothing Model normalization p<0.05 standard atlas parameter estimates
18 F-FDG PET study: SPM Diagnostic report for a patient with suspected AD. The coloured regions identified by SPM (retrosplenial cortex and bilateral temporal- parietal) are typical of AD pattern.
ISSUES SPM Requirements - Matlab licence - Normal subject database - Expertise in the use of SPM - Interaction with the software interface during image processing
THE SOLUTION SPM e-services for the assisted diagnosis of neurodegenerative diseases: the FP7-Eu project DECIDE (Diagnostic Enhancement of Confidence by an International Distributed Environment )
SPM SPM e-services in DECIDE (GridSPM) Requirements Requirements -Centralized/Grid porting - Matlab licence -Collected by DECIDE clinical partners - Normal subject database - Expertise in the use of SPM -Process is simpified/Users are trained - Interaction with the software -Avoided interface during image processing
SPM in DECIDE (GridSPM)
SPM in DECIDE (GridSPM)
SPM in DECIDE (GridSPM)
Results file : Page 1
Results file: Page 2
Results file : Page 3
Results file : Page 4
Recommend
More recommend
