Performance of evaluation of computed tomography scanners using a phantom model.

A phantom was designed to test computed tomography (CT) scanner performance relative to spatial resolution, contrast resolution, field uniformity, spatial linearity, artifact resistance, and radiation dose. Translate/rotate, stationary detector with rotating x-ray tube, and rotate-only scanners were tested under conditions that produced the best possible results for each machine within the dose ranges used in clinical practice. Machines can be ranked according to results for each parameters tested. The range of performance exhibited on machines tested was unrelated to class of scanner. Further, machine dedication to a particular body region did not necessarily secure superior results for that region. Spatial resolution at low contrast levels (poor on most machines) seems to be the most sensitive test of performance.