Masashi Yamauchi BS, Presenter: Nothing to Disclose

Yoshikazu Uchiyama PhD, Abstract Co-Author: Nothing to Disclose

Takahiko Asano, Abstract Co-Author: Nothing to Disclose

Hiroki Kato MD, Abstract Co-Author: Nothing to Disclose

Takeshi Hara PhD, Abstract Co-Author: Nothing to Disclose

Hiroshi Fujita PhD, Abstract Co-Author: Nothing to Disclose

Hiroaki Hoshi MD, Abstract Co-Author: Nothing to Disclose

Toru Iwama, Abstract Co-Author: Nothing to Disclose

et al, Abstract Co-Author: Nothing to Disclose

et al, Abstract Co-Author: Nothing to Disclose

Cerebrovascular diseases are the 3rd leading cause of death in Japan. Therefore, a screening system for early detection of cerebral and cerebrovascular diseases, which is named Brain Dock, is widely performed in Japan. In this system, occlusion is often detected using MRA images. The purpose of this study was to develop a CAD scheme for detection of occlusions in order to assist radiologists’ interpretation as a “second opinion.”

Our database consisted of 100 MRA studies obtained from 100 patients. These images were acquired on a 1.5 T MR scanner in Gifu University Hospital and Gero Hot Springs Hospital by using a 3D time-of -flight. All 3D MRA images were converted to isotropic volume data by using a linear interpolation, where each of the volume data had 400 x 400 x 200 voxels with a voxel size of 0.5 mm. Detection of abnormality was based on comparison with a reference normal MRA study with all vessels known. Vessel regions in a 3D target MRA study was first segmented by using thresholding and region growing techniques. Image registration was then performed so as to maximize the overlapping of the vessel regions in the target image and the reference image. The segmented vessel regions were classified into eight arteries based on comparison of the target image and the reference image. Relative lengths of the eight arteries were used as eight features in classifying the normal and arterial occlusion cases. Classifier based on the distance of a case from the center of distribution of normal cases was employed for distinguishing between normal cases and abnormal cases.

The sensitivity and specificity for the detection of abnormal cases with arterial occlusion were 80.0% and 95.3%, respectively. We also found that our method based on relative length of eight arteries can identify the artery with occlusion.

The results indicated that relative lengths of the eight arteries were useful for the detection of abnormal cases with arterial occlusions in MRA images.

Our computerized scheme for the detection of abnormal cases with arterial occlusion in MRA images may be useful for a screening system in detecting of cerebrovascular diseases.

Yamauchi, M, Uchiyama, Y, Asano, T, Kato, H, Hara, T, Fujita, H, Hoshi, H, Iwama, T, et al, , et al, , Computer-aided Diagnosis for Cerebral Diseases: Detection of Arterial Occlusions in MRA Images.  Radiological Society of North America 2007 Scientific Assembly and Annual Meeting, November 25 - November 30, 2007 ,Chicago IL. http://archive.rsna.org/2007/5005825.html