Quantitative pulmonary imaging: Spatial and temporal considerations in high-resolution CT

Abstract

The authors performed this study to determine how scanner temporal resolution affects image quality in high-resolution computed tomography (CT)-based quantitative analysis of lung parenchyma. A 37-kg, anesthetized white pig, in which 1.5-mm-diameter radiopaque markers were bronchoscopically implanted, was scanned supine with electron-beam CT and spiral CT scanners. Images were first acquired during apnea at functional residual capacity with electron-beam CT at a single level with multiple scan apertures. Without altering body posture, the animal was transported to the spiral CT scanner and scanned at the same lung volume and level. The animal was then euthanized and rescanned with spiral CT. All images were reconstructed with high-spatial-frequency algorithms resident on the respective scanners. High-resolution CT images of the live animal in the spiral scanner showed substantial cardiogenic motion artifacts for markers both near and distant from the heart. The matched postmortem images showed no motion artifacts. While line-pair phantom scans showed reduced spatial resolution of electron-beam CT compared with spiral CT scans, the electron-beam CT images of the markers in the live animal were free of artifacts, even with scan apertures of up to 700 msec. Motion artifacts may be accentuated by differences in high-resolution CT implementations. Applications such as lung parenchyma texture analysis, which regionally quantifies the subtle tissue density variations, will likely benefit from short scanning apertures.