To facilitate the accurate diagnosis of small peripheral lung cancers (less than 30 mm in
diameter), we evaluated the use of three-dimensional computed tomography (3D-CT).
Conventional CT images (10 mm thick), thin-slice CT images (2 mm thick) and 3-D-CT
images were prospectively studied in 29 patients with peripheral lung cancer (23
adenocarcinoma, 3 squamous cell carcinoma, 1 large cell carcinoma, 1 adeno-squamous
cell carcinoma, 1 small cell carcinoma) who underwent thoracotomy. Patients were 20
males and nine females with a median age of 65 (range 38-75) years. Conventional CT
and thin-slice CT images were examined with an extended scale (window level: -600,
window width: 1900) using high resolution images. 3D-CT images were made using
TCT-900S (Toshiba, Tokyo) software. In resected cases, surgical specimens were sliced
transversely to correlate with the CT images. Comparison of diagnoses based on CT
findings and on pathological findings showed that 3D-CT was 96% accurate for detecting
involved vessels, while conventional CT was 47% and thin-slice CT was 87% accurate.
Thus 3D-CT was superior to conventional CT for detecting involved vessel (p < 0.01).
3D-CT images also showed specific features of the lesions, including the appearance of the
margin and pleural indentations, as seen in thin-slice CT images. These results suggest that
3D-CT images may have an important role in diagnosing small peripheral lung cancers by
demonstrating involved vessels and the morphology of the lesions.