Computer Aided Diagnosis for Pulmonary Nodule on Low-Dose Computed Tomography (LDCT) Using Density Features

Low dose CT (LDCT) is imperative in pursuing early detection of lung cancer. From October 2008 to September 2010, 25 patients with 31 nodules were found in the health examination and underwent surgical operations at China Medical University Hospital . Of which, the lesions is classified into atypical adenomatous hyperplasia (AAH), bronchioloalveolar carcinoma (BAC), malignant, or benign other than AAH in the pathologic results. In this study, objective and standardized features were defined based on the CT number analysis in order to describe lesions and construct a computer aided diagnosis system. The diagnostic performance of constructed system is 70.97% in the accuracy index.