Effects of axial spatial resolution and sampling on object detectability and contrast for multiplanar position emission tomography.

Abstract

A multiplanar positron emission tomography (PET) system is simulated using Gaussian curves to model the axial point spread functions (PSFs) of the planes to study the effects of resolution and sampling. Poor spatial resolution or insufficient sampling may cause deleterious data losses or artifacts in the reconstructed image. For a multiplanar PET system with an axial full width at half-maximum (FWHM) of 6 mm and a 12 to 13 mm ring separation, a ripple in sensitivity of 9% is observed. A 1 mm object placed at the central direct plane results in detection of 59% of the signal in that plane. The theoretical observed contrast of a 3 mm object positioned at the center plane is 25% of the true contrast and decreases to 24% when the object is positioned between the central direct and cross planes. A PET system with an axial FWHM of 12 mm and a ring separation of 5-6 mm has a uniform sensitivity. A 1 mm object placed at the central direct plane detects 14% of the object signal in that plane. The theoretical observed contrast for a 3 mm object is 13% of true contrast when the object is positioned between the central direct and cross planes. It should be noted that all dimensions refer to the z direction through the center of the gantry in the simulated multiplanar system. The uniform sensitivity due to wider axial FWHMs decreases the amount of data loss for inter-ring gaps; however, the blurring associated with wider FWHMs decreases observed contrast.(ABSTRACT TRUNCATED AT 250 WORDS)