With the progress of medical imaging devices such as CT and MRI, the number of images per examination increases yearly. We have developed computer-aided diagnosis/detection (CAD) software to reduce the burden on radiologists in image interpretation and improve the accuracy of diagnosis. We have developed CAD software, including lung nodule detection in chest CT images, cerebral aneurysm detection in head MRA images, and lung mass detection in chest radiographs. We have also been improved the performance of CAD software by using deep learning.

The development of CAD software involves a cycle of algorithm development, software implementation, clinical use, and refinement of the algorithm and the software based on clinical evaluation. This cycle is expected to accelerate the development of CAD software. CIRCUS (Clinical Infrastructure for Radiologic Computation of United Solutions) is an integrated CAD development platform developed by The University of Tokyo Hospital, and related research projects aimed to realize the cycles. Dr. Nomura is participating as a core member of the project.