Michiel Sala, Presenter: Nothing to Disclose

Albert De Roos MD, Abstract Co-Author: Nothing to Disclose

Gerard-Jan Blauw, Abstract Co-Author: Nothing to Disclose

Huub A. Middelkoop PhD, Abstract Co-Author: Nothing to Disclose

Mark A. Van Buchem MD, PhD, Abstract Co-Author: Research Consultant, Medis Medical Imaging Systems, Inc

Ton de Craen, Abstract Co-Author: Nothing to Disclose

Jeroen Van Der Grond, Abstract Co-Author: Nothing to Disclose

To investigate whether magnetic resonance imaging markers of brain microstructure predict cognitive decline in elderly subjects at risk of vascular disease.

520 nondemented elderly subjects (296 men, mean age 74 ± 3 years) at risk of vascular disease were included. We used linear regression analysis to assess cross-sectional associations between mean magnetization transfer ratio (MTR), normalized MTR histogram peak height (PH), and z scores of cognition. Models were adjusted for age, gender, depressive symptoms, education level, individual white matter lesion volume, and brain atrophy. In 193 subjects with complete imaging and cognition data at baseline and follow up, we compared difference in cognition at baseline versus 3.3-year follow-up across tertiles of baseline MTR PH using one-way analysis of variance.

In the adjusted models, high MTR PH was associated with better cross-sectional performance on the STROOP-III test (mean test score, lower versus upper tertile, 61.9 versus 52.3 seconds, p=0.002). Likewise, both PH and mean MTR were positively associated with the delayed Picture Word Learning (PWL) test (p=0.008 and p=0.032, respectively) and the Letter Digit Coding (LDC) test (p=0.001 and p=0.007, respectively). Subjects with low baseline MTR PH demonstrated worse performance at follow-up as compared to baseline on the STROOP-III test (p=0.008), immediate PWL test (p=0.044), and delayed PWL test (p=0.003). On the contrary, in subjects with intermediate or high MTR PH height, cognitive performance was maintained after 3.3 years. 

In the elderly, MTR measures appear to be not only a general marker for cognition cross-sectionally, but especially MTR-peak height values seem also predictive for cognitive changes over time.

MTR histogram peak height may be considered as MRI marker of cognition in future clinical intervention studies. 

Sala, M, De Roos, A, Blauw, G, Middelkoop, H, Van Buchem, M, de Craen, T, Van Der Grond, J, Changes in Brain Microstructure Predict Cognitive Decline in Elderly Subjects at Risk of Vascular Disease.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14003164.html