Background: Whereas a large body of evidence suggests the usefulness of volumetric measurement of cerebral atrophy for diagnosing Alzheimer's disease (AD), the clinical applicability of cerebral volumetry for early detection of AD across multiple clinical sites is not well known. In the current study, we assessed the precision of volumetric measurement of the brain based on magnetic resonance imaging (MRI) in a multicenter setting.
Methods: The reliability of MRI was assessed by a phantom test of the American College of Radiology and voxel-based morphometry applied to the images obtained from a single subject tested at 11 centers of the German Dementia Competence Network.
Result: Nine of the 11 centers tested met the reliability criteria of the phantom test. Across all centers, a bias was found in the measurements of slice thickness and length. For voxel-based morphometry, the coefficient of variation yielded 5.02% for gray matter volume and 12.81% (SD 9.06%) for gray matter signal intensity in voxels. Power analysis showed that a sample size of 150 subjects is sufficient for statistically valid detection of reduced gray matter volume in patients with mild cognitive impairment.
Conclusion: The reliability of measurements from multiple centers is sufficient to allow statistically valid analysis of MRI data.