A tensor based morphometry study of longitudinal gray matter contraction in FTD.

Frontotemporal dementia (FTD) is a neurodegenerative disease characterized by progressive behavioural abnormalities and frontotemporal atrophy. Here we used tensor based morphometry (TBM) to identify regions of longitudinal progression of gray matter atrophy in FTD compared to controls. T1-weighted MRI images were acquired at presentation and 1-year follow-up from 12 patients with mild to moderate FTD and 12 healthy controls. Using TBM as implemented in SPM2, a voxel-wise estimation of regional tissue volume change was derived from the deformation field required to warp a subject's late to early anatomical images. A whole brain analysis was performed, in which a level of significance of p<0.05 corrected for multiple comparisons (family wise error-FWE) was accepted. Based on prior studies, a region of interest (ROI) analysis was also performed, including in the search area bilateral medial and orbital frontal regions, anterior cingulate gyrus, insula, amygdala and hippocampus. Within this ROI a level of significance of p<0.001 uncorrected was accepted. In the whole brain analysis, the anterior cingulate/paracingulate gyri were the only regions that showed significant atrophy change over 1 year. In the ROI analysis, the left ventro-medial frontal cortex, right medial superior frontal gyrus, anterior insulae and left amygdala/hippocampus showed significant longitudinal changes. In conclusion, limbic and paralimbic regions showed detectable gray matter contraction over 1 year in FTD, confirming the susceptibility of these regions to the disease and the consistency with their putative role in causing typical presenting behaviours. These results suggest that TBM might be useful in tracking progression of regional atrophy in FTD.