IC-P2-117: Comparing voxel-based morphometry (VBM) to region of interest (ROI) measures while assessing hippocampal atrophy in patients with mild cognitive impairment

Abstract

atrophy rates with 90% power. Results: The FTLD cohort exhibited significantly greater whole-brain atrophy rates than controls using both manual segmentation (p 0.029) and the automated BSI (p 0.001). The BBSI and VBSI techniques significantly reduced variability of results compared with manual segmentation and were used as the primary outcome for the following analyses. Pairwise comparisons of all three FTLD subgroups with controls demonstrated significantly higher whole-brain atrophy rates (bvFTLD, p 0.040; SD, p 0.001; and PNFA, p 0.023). Ventricular expansion rates were also significantly greater than controls for the whole FTLD cohort (p 0.001) and for the SD (p 0.001) and PNFA (p 0.001) subgroups, but not bvFTLD (p 0.226), potentially due to the pathogenetic heterogeneity of bvFTLD patients. In future multi-centre clinical trials, using the BBSI would require 221 FTLD patients (with similar numbers per subgroup) to detect a 30% reduction in atrophy rate with 90% power. Conclusions: Whole-brain and ventricular BSI measurements show promise as biomarkers of disease progression in FTLD and could be utilised as outcome measures of treatment effects in future clinical trials of disease-modifying therapies.