Premium
[O1–12–06]: RELATIONSHIP BETWEEN LONGITUDINAL STRUCTURAL ATROPHY AND AV1451 TAU UPTAKE IN AMYLOID‐POSITIVE INDIVIDUALS
Author(s) -
Das Sandhitsu R.,
Xie Long,
Wisse Laura E.M.,
Ittyerah Ranjit,
Tustison Nicholas J.,
Yushkevich Paul A.,
Wolk David A.
Publication year - 2017
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2017.07.106
Subject(s) - atrophy , temporal lobe , frontal lobe , pittsburgh compound b , psychology , orbitofrontal cortex , posterior cingulate , neuroimaging , cortex (anatomy) , magnetic resonance imaging , parietal lobe , neuroscience , nuclear medicine , medicine , pathology , prefrontal cortex , cognitive impairment , cognition , radiology , epilepsy
Background: In clinically normal (CN) older individuals, interrelationships between cortical thickness and b-amyloidosis on cognition have not been fully explored using continuous distributions of the biomarkers. Methods:We studied 1200 persons, ages 50-95 years, in the Mayo Clinic Study of Aging who were CN by a consensus process that included neuropsychological testing, a physician examination, and a formal Clinical Dementia Rating involving participant and informant. Participants underwent structural MRI from which cortical thickness measures from an AD signature meta-ROI (“THK”) were extracted using Freesurfer 5.3, and a PIB PET scan from which a global SUVR (“PIB”) was extracted. Cognition was quantified using a global Z-score derived from a 9-test neuropsychological battery. Participants had at least one follow-up visit during which they were assessed cognitively. Relationships between longitudinal cognition (annual change in global cognitive z-score (DCOG) and baseline imaging were assessed using a Gradient Boosting Machine (GBM) model. GBM is a machine learning technique, with subsequent statistical theory, that can model general nonlinear multivariate relationships yet avoid overfitting. Results: In this population based sample, age had a large influence on the range of observed THK, PIB and baseline cognition. The change in z-score (magnified X 10 for presentation purposes) as a function of the baseline biomarkers is shown in the Figure. The Figure shows the 3-dimensional relationships (left), PIB x DCOG for different levels of THK (middle) and THK x DCOG for different levels of PIB (right) for two age groups. DCOG worsened across THK, but more so at lower THK levels ( w2.0. However, higher incremental levels of PIB were associated with less incremental DCOG. Conclusions: Consistent with its hypothesized role as a neurodegeneration biomarker linked mechanistically to cognition, THK was related to future declines in cognition with incrementally greater changes at lower levels of cortical thinning. While PIB contributed to cognitive decline, its decelerating relationship to cognition at higher b-amyloid levels suggested a more remote mechanistic relationship. Support: NIH grants U01AG06786, R01AG41851, R0 AG11378, Elsie and Marvin Dekelboum Family Foundation, Robert H. and Clarice Smith and Abigail Van Buren Alzheimer’s Disease Research Program. O1-12-06 RELATIONSHIP BETWEEN LONGITUDINAL STRUCTURAL ATROPHY AND AV1451 TAU UPTAKE IN AMYLOIDPOSITIVE INDIVIDUALS Sandhitsu R. Das, Long Xie, Laura E. M. Wisse, Ranjit Ittyerah, Nicholas J. Tustison, Paul A. Yushkevich, David A. Wolk, Penn Image Computing and Science Laboratory (PICSL), University of Pennsylvania, Philadelphia, PA, USA; University of Pennsylvania, Philadelphia, PA, USA; University of Virginia, Charlottesville, VA, USA; Penn Memory Center, University of Pennsylvania, Philadelphia, PA, USA. Contact e-mail: sudas@seas.upenn.edu