[Evaluation of computer-assisted diagnosis of small pulmonary nodules using contrast-enhanced dynamic helical CT]

Dynamic helical computed tomography(DH-CT) was performed to examine 45 lesions(23 lung cancers, 2 pulmonary metastasis, and 20 benign nodules) as an application of computer-assisted qualitative analysis of small pulmonary nodules. Based on the three-dimensional(3D) CT image data, the internal structure of the nodule was assessed quantitatively and temporal changes were evaluated. The entire lesion was examined by helical scanning with a beam width of 2 mm(pitch of 1) before contrast enhancement and 2 and 4 minutes after contrast enhancement. Using the digital CT data, the pixels inside the nodules were quantified based on a combination of CT values and 3D curvature in order to differentiate between benign(BN) and malignant nodules (MN). (The numeral score abovementioned was calculated.) The average scores for BN 2 and 4 minutes were -5.72 and -12.8, respectively, and those for MN were 5.51 and 12.0, respectively(p < 0.01). Assuming that a score of 0 or higher indicates a MN based on the CT data 4 minutes. In conclusion, 3D computer-assisted analysis of the internal structure of small pulmonary nodules using contrast-enhanced DH-CT was found to be effective for differentiating between benign and malignant nodules.