Abstract

RATIONALE AND OBJECTIVES:

To improve effective scan time and image quality in cone-beam computed tomography (CT), Parker's weighting function (half scan [HS]-Feldkamp-Davis-Kress [FDK]) extended to a larger range up to 2pi was proposed as new half-scan algorithm (NHS-FDK). We conducted a practical physical evaluation of NHS-FDK and HS-FDK using 256-detector row CT.

MATERIALS AND METHODS:

Three types of weighting function (full-scan [FS-FDK], HS-FDK, and NHS-FDK) were evaluated by using 256-detector row CT for five variables, ie, point spread function, image noise, CT number uniformity, Feldkamp artifact, temporal resolution, and clinical evaluation.

RESULTS:

Image noise, Feldkamp artifact, and temporal resolution were dependent on weighting function. Image noise magnitude was independent of projection angle for all regions of interest with FS-FDK, but showed a symmetric pattern with projection angle with HS-FDK and NHS-FDK. With regard to temporal resolution, NHS-FDK did not remove the motion artifact in the heart except in such slower motion organs as the pulmonary vessels, whereas HS-FDK reduced the motion artifact in the heart. HS-FDK had an even more incomplete data region in the Radon space than FS-FDK, suggesting that it would provide poor image quality distant to the midplane in the longitudinal direction. In practical testing in human subjects, HS-FDK showed inferior performance in all variables except temporal resolution.

CONCLUSION:

Despite its inferiority to FS-FDK for static objects, HS-FDK may be useful in chest imaging. Contrary to previous findings using static images, NHS-FDK failed to show advantages over HS-FDK or FS-FDK in a moving phantom and human subjects.