Sensitivity improvement of automatic pulmonary nodules detection in chest X-ray CT images

In this paper, we develop an automatic detection method of non-isolated pulmonary nodules as a part of computer-aided diagnosis (CAD) system for lung cancers in chest X-ray CT images. An essential core of the method is to separate non-isolated nodules from connecting structures like chest wall and blood vessels. The isolated nodules can be detected more easily by CAD syetems developed previously. To this end, we propose a binarization technique by using two thresholds as a preprocessing for nodule candidates. We evaluate the performance using the receiver operating characteristic (ROC) analysis in clinical chest CT images. The results suggest that detection ability of nonisolated nodules by the proposed method is superior to that by the conventional preprocessing methods.