Abstract
Medical imaging devices have been widely used in clinical medicine to support diagnosis. Doctors need a lot of time and experience for the interpretation of a huge number of images before a screen or monitor to find the lesion candidates without any overlook. In order to reduce the reading time, effective image analysis algorithms, and software tools are expected by Picture Archiving and Communication System (PACS) that commonly used for image interpretations. Computer-aided diagnosis (CAD) system has been expected as a sufficient solution by detecting the lesion candidates, however, the performance and accuracy of CAD are still not good enough for practical applications. In the other hand, the fundamental part of CAD can provide the robust recognition results of the anatomical structures on CT images. Those information maybe useful but still not evaluated on PACS. In this study, we use the anatomical structures that recognized automatically by a CAD to improve the efficiency of torso CT image interpretations on a PACS. We confirmed that our improvements can simplify the operations of PACS and reduce the reading time of CT image interpretations without any loss on accuracy.
