Quantitative Measurement of MR Cortical Atrophy: MR Brain Surface Intensity Model(BSIM)and Group and Individual Cortical Thinning Studies

摘要

Alzheimer's disease(AD)is caused by pathological changes including cortical thinning occurring throughout the brain. Traditional methods for assessing cortical thickness are challenged by the sub-millimeter accuracy required for clinical conditions and the convoluted nature of brain surface. Furthermore, there is a significant overlap of gray and white matter intensities. A novel Brain Surface Intensity Model(BSIM)has been developed for use as a potential imaging biomarker for neurodegenerative diseases. BSIM technique extracts MR intensity profiles perpendicular to a mathematically defined gray matter iso-intensity layer(GMIIL)at predefined reference points, fits that profile to BSIM, and computes cortical thickness. A 3D visualization tool has been developed to evaluate intensity extraction and model calculation. 29 normal subjects aged between 70 to 80 years from ADNI database were used to generate normal references and measure individual Z-score cortical thinning. 30 age-matched AD subjects were used to study thinning patterns. Significant cortical thinning(p＜0.0001)was found for AD group. 95% confidence interval of the cortical thinning in AD patients was from 0.17 to 0.23 mm. The cortical thinning of the AD patients showed distinct features that differentiate AD patients from normal controls. The thickness measurements of 29 normal controls were validated by comparing with results from literature(p = 0.94). BSIM technique avoids complicated 3D segmentation of brain gray and white matters, and simplifies the thickness calculation. Moreover, it is less affected by the image noise, inhomogeneity, partial volume effects, and the intensity overlap of the white and gray matters.