Quantification of the heterogeneity of cerebral blood flow in vascular dementia.

Abstract

In vascular dementia (VaD), assessment of cerebral blood flow by single photon emission computed tomography (CBF SPECT) has been used to detect a patchy decrease of blood flow or a frontal reduction. In addition to reduced blood flow, the heterogeneous distribution of cerebral blood flow is often observed in VaD. However, no objective method to evaluate the heterogeneity has been established. In this study, we applied three-dimensional fractal analysis (3D-FA) to CBF SPECT images as a method for assessing the heterogeneity of the cerebral blood flow distribution in VaD. The subjects included 18 patients with a diagnosis of VaD (aged 69.7 +/- 8.3) based on neuropsychological testing and imaging findings and 18 age-matched controls (aged 66.9 +/- 10.3). CBF SPECT images were obtained with (99m)Tc-hexamethyl propyleneamine oxime. On the reconstructed images, we obtained a linear regression equation between the cut-off values (from 35 to 50 %) and the number of voxels with a radioactivity exceeding the cut-off value transformed into natural logarithms, and then calculated the fractal dimension from the slope of the regression line thus obtained. The Mini-Mental State Examination (MMSE) was used to evaluate cognitive function. The fractal dimensions were 1.084 +/- 0.153 and 0.853 +/- 0.062 (mean +/- SD) in the VaD and control groups, respectively. The fractal dimension was significantly greater in the VaD group than in the control group (p < 0.0001). A significant negative correlation was observed between the fractal dimension and the MMSE score in the VaD group (r = 0.871, p < 0.0001). Because the CBF SPECT images of VaD patients showed a higher fractal dimension, these images were quantitatively more heterogeneous than those of age-matched controls. In the VaD group, cognitive function was shown to decline as the fractal dimension increased and images became more heterogeneous.