[P3–377]: BRAIN MRI FINDINGS WHICH MAY PREDICT PROGRESSION OF MILD COGNITIVE IMPAIRMENT

Background: 19% of people older than 65 develop mild cognitive impairment (MCI), and 46% of these patients will progress to dementia within 3 years. Not many prognostic indicators are available. Metabolic syndrome has been suggested as a possible risk factor for dementia. Methods: In a cohort of 96 total subjects with MCI with and without progression to dementia and normal controls, the earliest available axial FLAIR images for each subject were analyzed using Olea Sphere software (Olea Medical, France) to calculate white matter hyperintensity volume and total brain parenchyma volume by a blinded observer. Additionally, a subjective categorization of the distribution of white matter hyperintensity was assigned to each case by a blinded observer into one of three categories (central, peripheral, or mixed). Statistical analysis was performed to identify correlation between patterns of distribution, volume of hyperintensity, and presence/absence of disease and disease progression. Results: There was significant differences of the volume of hyperintensity with respect to distribution of disease within the normal (p = 0.002) and MCI (p = 0.046) subjects. There was a significant interaction (P =0.020) between the assigned pattern of white matter disease distribution and the subject's disease burden. There was no association (p = 0.221) between the disease status (normal controls, MCI with and without progression) and total volume of white matter hyperintensity, although this strengthened after adjusting for total parenchymal volume (p = 0.131). Conclusions: The correlation between the pattern of brain white matter changes distribution and the disease status may help in determining which patients with MCI will be more likely to progress to dementia, therefore allowing directed therapy or better targeting of therapeutic trials. Additional characterization with a larger cohort including volumetric analysis of the white matter hyperintensity pattern and correlating with clinical diagnosis of metabolic syndrome could yield more specific resolution as to whether or not this can predict progression of MCI to dementia.