Fractal analysis of cerebral blood flow distribution in Alzheimer's disease

This study describes a method for quantifying cerebral blood flow (CBF) dis tribution in Alzheimer's disease (AD) from SPECT images obtained with Tc-99 m-hexamethylpropyleneamine oxime (HMPAO) by 3-dimensional fractal analysis (3D-FA). Methods: HMPAO SPECT was performed on 68 patients with probable AD and 14 healthy control subjects. We delineated the CBF images using 4 cuto ff levels (35%, 40%, 45%, and 50% of the maximal voxel radioactivity) and m easured the total number of voxels in the areas surrounded by the contours obtained with each cutoff level. We calculated fractal dimensions from the relationship between the total number of voxels and the cutoff levels trans formed into natural logarithms. Results: The fractal dimensions (mean SD) f or patients with probable AD and healthy subjects were 0.74 +/- 0.33 and 0. 52 +/- 0.09, respectively. A significant difference in the fractal dimensio n was found between groups (P = 0.001). Statistically significant correlati ons were obtained between the fractal dimension and the Mini-Mental State E xamination score (r = -0.598; P < 0.0001) and between the fractal dimension and the AD Assessment Scale (r = 0.670; P < 0.0001). The fractal dimension s for subjects with clinical dementia rates (CDRs) of 0, 1, 2, and 3 were 0 .52 +/- 0.09, 0.63 +/- 0.21, 0.77 +/- 0.23 (P < 0.05 vs. the group with a C DR of 1), and 1.43 +/- 0.49 (P < 0.0001 vs. the group with a CDR of 2), res pectively. Conclusion: Quantification of CBF distribution on SPECT images i n AD was possible using 3D-FA. The fractal dimension was well correlated wi th the cognitive impairment, as assessed in neuropsychologic tests. 3D-FA m ay be a useful method for objectively evaluating the progression of AD.