COMPUTER-AIDED DIAGNOSIS FOR PULMONARY NODULES BASED ON HELICAL CT IMAGES

In this paper, we present a computer-assisted automatic diagnostic sys tem for lung cancer that detects nodule candidates at an early stage f rom helical CT images of the thorax. Our diagnostic system consists of analytical and diagnostic procedures. In the analytical procedure, fi rst we extract the lung and the pulmonary blood vessel regions using t he fuzzy clustering algorithm, then we analyze the features of these r egions using image-processing techniques. In the diagnostic procedure, we define diagnostic rules utilizing the extracted features which sup port the determination of the candidate nodule locations. We show the effectiveness of our system by giving the results from its application to image data for mass screening of 450 patients. (C) 1948 Elsevier S cience Ltd. All rights reserved.