Can maximum intensity projection images with multidetector-row computed tomography help to differentiate between the micronodular distribution of focal and diffuse infiltrative lung diseases?

OBJECTIVE

The objective of this study was to evaluate whether maximum intensity projection (MIP) images increased the ability of experienced and resident radiologists to differentiate between the micronodular distribution of focal and diffuse infiltrative lung diseases.

METHODS

The cases used in the study were those of 26 patients with focal or diffuse micronodular lung diseases, including 7 cases of sarcoidosis, 6 of miliary tuberculosis, 3 of pulmonary tuberculosis, 3 of chronic bronchitis, 2 of human T-lymphotropic virus type 1-associated bronchoalveolar disorder, 2 of diffuse aspiration bronchiolitis, 1 of atypical mycobacterial infection, and 1 of lymphangitic carcinomatosis. Scans of the entire lung during a single breath hold at 1.25-2.5 mm thickness and a pitch of 6 were performed using a multidetector-row computed tomography (MDCT) apparatus with a 4-row detector. Additional MIP image slabs were produced from the initial axial images on all study patients on a workstation according to a protocol that incorporated a slab thickness of 10 mm, a reconstructed interval of 10 mm, and a window width of -1500 Hounsfield units. The ability of 10 radiologists (5 board-certified radiologists and 5 radiology residents) to interpret contiguous thin-section CT scans and additional MIP images was then studied in an observer performance study. The results of both sets of observer performances were compared using receiver operating characteristic analysis.

RESULTS

In the resident observers, the mean area under the receiver operating characteristic curve (Az) value increased significantly from 0.654 without the MIP images to 0.753 with the MIP images (P < 0.01). In the board-certified radiologists, however, the mean Az values remained unchanged at 0.867 without the MIP images and 0.846 with the MIP images.

CONCLUSIONS

This study showed that MIP images may help radiologists in training to differentiate between the micronodular distribution of focal and diffuse infiltrative lung diseases.