Abnormal metabolic brain network in behavioral variant of frontotemporal dementia

Abstract

AbstractBackgroundThe increasing number of patients with neurodegenerative dementia is a major public health issue. While clinical features of neurodegenerative disorders causing dementia are overlapping, accurate diagnosis is a prerequisite for an effective drug development. Besides the need for a better biomarker of Alzheimer’s disease (AD), development of non‐AD biomarkers is necessary. Behavioral variant of frontotemporal dementia (bvFTD) is the second most common neurodegenerative dementia in patients younger than 65 years. The aim of our study was to identify and validate bvFTD specific metabolic brain pattern using voxel‐based spatial covariance analysis according to the scaled subprofile model (SSM/PCA) on a new cohort of European patients and to study its biologic significance.Method20 patients with clinical diagnosis of bvFTD and 20 age‐matched normal controls (NC) underwent metabolic imaging with FDG PET. CSF results were available for 10/20 bvFTD patients and potential AD was excluded based on the biomarker profile. Using SSM/PCA, we identified a significant bvFTD‐related metabolic pattern termed bFDRP‐Slov. The stability of voxel weights on this pattern was evaluated using bootstrap resampling. For validation, pattern expression values were computed in an independent testing set comprised of 25 bvFTD and 22 NC subjects. The specificity of the pattern for bvFTD was assessed by computing values in patients with other causes of dementia: 26 with AD and 16 with sporadic Creutzfeldt‐Jakob disease (sCJD). bFDRP expression was correlated with disease duration and with MMSE.ResultThe bFDRP was characterized by pronounced metabolic reduction in the medial frontal cortex, middle and inferior frontal gyri, anterior cingulum, insula, bilateral caudate and thalamus. bvFDRP expression values discriminated bvFTD patients from their counterparts with AD, sCJD and NC (p<0.0001; one‐way ANOVA). The expression of this pattern was abnormally elevated in bvFTD patients in the identification and validation cohorts (p<0.0001). bvFTD patients’ pattern expression of values did not correlate disease duration or MMSE.ConclusionbFDRP is a promising metabolic network biomarker of the disorder, consistent with previously identified disease pattern. Further research is needed to study its clinical significance. Identification of multiple neurodegenerative brain networks may be critical as the combination of different patterns can improve diagnostic accuracy.