[A quantitative study of brain atrophy on computed tomography--multivariate analysis for comparison between the...

Area of Science:

• Neuroimaging
• Gerontology
• Biostatistics

Background:

• Cerebral atrophy assessment traditionally uses quantitative pixel counting or less precise linear measurements.
• Previous data suggested cerebrospinal fluid volume increases significantly after age 50.
• A need exists for a more accessible yet quantitative method to measure brain atrophy.

Purpose of the Study:

• To develop and validate a simplified, quantitative method for measuring cerebral atrophy.
• To compare a novel linear measurement formula against a pixel count method.
• To establish a reliable index for brain atrophy using accessible parameters.

Main Methods:

• Seventy-four subjects were analyzed using both pixel count and linear measurement techniques.
• Stepwise multiple regression identified key linear parameters: caudate nuclei distance (B), lateral ventricle distance (E), and visible sulci count (G).
• A formula was derived using B, E, and G, with B and E normalized by skull width (H), to estimate the CSF-cranial ratio (CCR).

Main Results:

• A multivariant analysis yielded a formula: y = 42.66 X B/H + 12.52 X E/H + 0.232 X G - 2.92, predicting CCR.
• The formula achieved a high multiple correlation coefficient of 0.76 (p < 0.001).
• The developed formula demonstrated statistical significance in predicting brain atrophy.

Conclusions:

• The derived formula provides an easy and quantitative method to predict the CSF-cranial ratio (CCR) as an index of brain atrophy.
• This non-computerized approach offers a practical alternative to the labor-intensive pixel count method.
• The findings facilitate accessible assessment of age-related brain changes and cerebral atrophy.