Spatial variation of resolution and noise in multi-detector row spiral CT.

RATIONALE AND OBJECTIVES The authors performed this study to evaluate an approach for measuring the variations of three-dimensional spatial resolution and image noise throughout a field of view imaged with multi-detector row spiral computed tomographic (CT) scanners. MATERIALS AND METHODS The authors designed a phantom (diameter, 320 mm) that contained 37 metallic spheres (diameter, approximately 0.8 mm) positioned between two disks made of a material with attenuation being that of water. One sphere was located at the isocenter of the phantom, and the rest were evenly spaced in three concentric rings with diameters of 100, 200, and 300 mm, respectively. The phantom was imaged with two widely used multi-detector row CT scanners by using a standard protocol and four variations of that protocol. Because a recently developed theory holds that image resolution should be proportional to the square root of the trace of the covariance matrix of a point spread function, the authors developed a software package to segment high-attenuation spheres from the CT image volume and compute point spread functions from blurred images of the spheres. Three-dimensional spatial resolution and image noise were calculated as a function of radial distance within the field of view. RESULTS Resolution and noise were quantified in the resultant CT image volumes and found to be nonisotropic, with worse resolution and less noise occurring at the periphery of the field of view. CONCLUSION The method enabled measurement of variations in spatial resolution and of their distribution on images obtained with multi-detector row CT scanners. These findings may contribute to the development of an improved algorithm for image reconstruction.