Abstract #4733
Volume-based vs. voxel-based brain morphometry in Alzheimer's disease prediction
Alexis Roche 1,2 , Daniel Schmitter 1,3 , Bndicte Marchal 1 , Delphine Ribes 1 , Ahmed Abdulkadir 4 , Meritxell Bach-Cuadra 2,5 , Alessandro Daducci 5 , Cristina Granziera 1,6 , Stefan Klppel 4 , Philippe Maeder 2 , Reto Meuli 2 , and Gunnar Krueger 1
1
Advanced Clinical Imaging Technology,
Siemens Healthcare IM BM PI, Lausanne, Switzerland,
2
Department
of Radiology, University Hospital (CHUV), Lausanne,
Switzerland,
3
Biomedical
Imaging Group, EPFL, Lausanne, Switzerland,
4
Group
of Pattern Recognition and Image Processing, University
of Freiburg, Germany,
5
Signal
Processing Laboratory 5, EPFL, Lausanne, Switzerland,
6
Service
of Neurology, University Hospital (CHUV), Lausanne,
Switzerland
This study compares different MR T1-based brain
morphometry methods for automated classification of
Alzheimer patients, mild cognitively impaired patients
and elderly controls on a standardized analysis set of
818 scans from the ADNI NIH-funded project. The methods
under investigation are standard voxel-based morphometry,
as implemented by the SPM software, and volume-based
morphometry as implemented in respectively different
ways by FreeSurfer and Siemens prototype MorphoBox. Our
results show that classification using volume-based
morphometry is at least as accurate as voxel-based
morphometry, therefore proving volume-based morphometry
to be a valuable methodology to assist the diagnosis of
Alzheimer's disease and mild cognitive impairment.
This abstract and the presentation materials are available to members only;
a login is required.
Join Here
