Detecting Dynamic (4D) Profiles of Degenerative Rates in Alzheimer's Disease Patients, Using High-Resolution Tensor Mapping and a
Brain Atlas Encoding Atrophic Rates in a Population

¹Paul Thompson PhD, ²Greig de Zubicaray PhD, ²Andrew L. Janke, ²Stephen E. Rose PhD, ¹Michael Mega MD PhD, ³James Semple PhD, ¹Michael Hong, ¹Jeffrey L. Cummings MD, ²David M. Doddrell PhD, ¹Arthur W. Toga PhD

¹Laboratory of Neuro Imaging, Brain Mapping Division, and UCLA Alzheimer's Disease Center,
Dept. Neurology, UCLA School of Medicine, Los Angeles CA 90095, USA

²Centre for Magnetic Resonance, University of Queensland, Brisbane 4072, Australia

³SmithKline Beecham Pharmaceuticals plc, UK

Objective:        

We report a new strategy to create detailed maps of degenerative rates in Alzheimer's Disease (AD). We compare individual patients' maps with normative data stored in a population-based brain atlas.

Background:     

Early detection of Alzheimer’s Disease requires tools with unprecedented sensitivity for mapping the dynamic progression of disease and therapeutic response. By contrast with conventional volumetric approaches, we developed a tensor mapping technique to uncover spatially complex profiles of atrophic rates and gray matter loss throughout the brain, in individuals and large patient populations.

Design/Methods:   

A population-based brain atlas was first created from an archive of T1-weighted 256x256x170 3D MRI scans of 43 AD patients (age: 68.7+/-1.7 yrs.; 24 females/19 males; MMSE score: 20.0+/-0.8) and 34 controls matched for age, education, gender and handedness (all right-handed). Anatomic variability patterns, based on 3D models of 6840 structures, were encoded as a Gaussian random tensor field, and used to detect pathology in new patients.

Results:

Dynamic maps of atrophic rates, with millions of degrees of freedom, were then generated for 17 AD patients and 14 demographically-matched controls scanned repeatedly over a 4-year period (interscan interval: 2.6±0.3 yrs.; final age: 71.3±1.8 yrs.). 4D maps of annual atrophic rates were elastically aligned across subjects, averaged, and confidence limits were computed for tissue loss at each anatomical point throughout the brain. Profiles of local atrophic rates were visualized. Left faster than right hippocampal tissue loss was detected in controls (L:-3.8±1.6%/yr.; R:-0.5±1.2%/yr.; p<0.05). Significantly faster loss rates were found bilaterally in AD (L: -5.9%±1.7%/yr.; R:-7.1±3.2%/yr.; p<0.03), and their 3D profiles were visualized. In controls, these loss rates peaked at a localized region of the medial surface of the left hippocampal head. In AD, an anterior to posterior shift was detected in the region of peak loss, which broadened to encompass the entire hippocampus, bilaterally. Local atrophic rates were significantly linked to the rate of cognitive decline (r=0.7; p<0.05). Severest decline in MMSE score was observed when average atrophic rates exceeded 3%/yr. Individual rates were unrelated to educational level (r<0.1). Specialized approaches for group averaging of cortical and hippocampal anatomy were then used to compare atrophic rates, shape, and gray matter loss across groups, revealing a comparative sparing of specific cortical gyri. Regional asymmetries and disease-specific changes appeared in the average maps that were not apparent in individual anatomies.

Conclusions:

By storing information on degenerative rates in a population, these disease-specific 4D brain atlasing systems may be advantageous for charting disease progression in genetic, demographic, and drug studies of dementia.

Grant Support: (PT/AWT): NIMH/NIDA (P20 MH/DA52176), P41 NCRR RR13642, NINDS/NIMH (NS38753), NLM (LM/MH05639), NSF (BIR 93-22434), NCRR (RR05956); (MSM): NIA K08-AG100784.

Related Publications

Contact Information