Differences due to collimator blurring in cardiac images with use of circular and elliptic camera orbits

Background. In cardiac imaging systems, an elliptic acquisition orbit about the patient can be used to enhance resolution of single photon emission co mputed tomography (SPECT) images by minimizing the distance between the obj ect imaged and the rotating detector system. In this study artifacts from i mages acquired with the standard circular acquisition are compared with tho se acquired with various elliptic acquisitions. Methods and Results. With the use of elliptic camera orbits of different ec centricities, simulated projection data were generated from a slice through the left ventricle (LV). The projection data included a simulation of the degradation due to the depth-dependent response of the collimator. As is co mmon in many clinical systems, SPECT images were reconstructed with the sta ndard filtered backprojection algorithm without correction for the collimat or response. When the ratio of the major-to-minor axis of the acquisition a rc is changed from 1 (circular) to 1.5 (elliptic), reconstructed SPECT imag es show an additional loss of counts (about 10%) in the apical region of th e LV. The severity of the apical defect is also dependent on the starting a ngle of the acquisition arc. When the starting angle is changed from 0 degr ees (detector parallel to the major axis of the LV) to 60 degrees, the rati o between the minimum count in the apical region and the maximum count in t he left ventricular myocardial wall decreases by as much as 20%. Conclusions. SPECT image artifacts from elliptic acquisitions are significa ntly more severe than those from circular acquisitions. Because of the sign ificant difference in images reconstructed from circular and elliptic acqui sitions, standardized normal files acquired from circular acquisitions shou ld not be used for comparisons with patient data acquired from elliptic acq uisitions.