Reliability of Changes in Brain Volume Determined by Longitudinal Voxel-Based Morphometry.

Abstract

Longitudinal magnetic resonance imaging (MRI) studies have become increasingly important to assess the changes in brain morphology during normal aging and neurodegenerative disorders. However, the reliability of longitudinal morphometric changes has not been fully evaluated. To examine the reliability of longitudinal (2-year) changes in brain morphology determined by longitudinal voxel-based morphometry (VBM) in healthy elderly subjects, patients with mild cognitive impairment (MCI), and patients with Alzheimer's disease (AD). Retrospective analysis. Twenty-four healthy elderly subjects, 28 MCI patients, and 16 AD patients. A 1.5 T, magnetization-prepared rapid gradient-echo. Longitudinal (2-year) changes in gray matter volume determined by longitudinal VBM processing, and visual assessment of image quality. Intraclass correlation coefficient (ICC) and Kruskal-Wallis test. The ICC maps differed among the three groups. The mean ICC was 0.81 overall (0.86 for healthy elderly subjects, 0.75 for MCI patients, and 0.76 for AD patients). The reliability was good to excellent (ICC, 0.60-1.00) for 92% of voxels (99% for healthy elderly subjects, 83% for MCI patients, and 83% for AD patients). The image quality differed significantly among the three groups (P< 0.05). These results indicate that the reliability of longitudinal gray matter volume changes by VBM is good to excellent for most voxels. However, reliability may be affected by the disease, possibly due to differences in head motion during imaging. Evidence Level 3 Technical Efficacy Stage 1.