[Visualization of intracranial arteriovenous malformation using multivariate analysis]

Nihon Igaku Hoshasen Gakkai Zasshi. 2003 Jul;63(6):322-8.
[Article in Japanese]

Abstract

Intracranial arteriovenous malformations(AVMs) are composed of native deformed vessels where arteries and veins are directly connected in the absence of capillary vessels, and have a complicated three-dimensional(3D) shape. The purpose of this study was to define precisely the location of AVM areas for stereotactic irradiation. The proposed method used a sequence of dynamic X-ray CT images with contrast and opaque media. We analyzed these images by a principal component analysis(PCA) method and extracted component images: artery, vein, and background images. Then we made a fusion image with these three images and displayed the AVM areas. Results showed that the proposed method could identify the physiological location of these AVMs areas. Furthermore, we reconstructed 3D fusion images using two-dimensional component images extracted by the PCA method.

MeSH terms

  • Angiography, Digital Subtraction
  • Humans
  • Image Processing, Computer-Assisted
  • Imaging, Three-Dimensional*
  • Intracranial Arteriovenous Malformations / diagnostic imaging*
  • Multivariate Analysis
  • Tomography, X-Ray Computed / methods*