Preliminary assessment of multislit-slat collimation for cardiac SPECT
Cardiac imaging accounts for over half of all SPECT scans. Improvements in this area have a large potential clinical impact. Slit-slat imaging has shown advantages over parallel-beam (PB) and fan-beam (FB) for mid-size objects. That advantage can be extended by using multiple non-multiplexing slits. Herein, the relative sensitivity for cardiac imaging on a general-purpose SPECT scanner is determined in comparison with PB and FB for equal average resolutions. A single transverse slice of a digital phantom is used to represent the location of the myocardium within a male torso. The transverse and axial resolutions averaged over the myocardium are computed for PB and FB, based on manufacturer specifications from Siemens. These average resolutions are then matched with a multislitslat collimator by adjusting the focal length, slit width, and the ratio of slat gap to height while maximizing average sensitivity over the myocardium. The orbit of the collimators extends from LPO to RAO with a 1 cm clearance, resulting in torso truncation for multislit-slat. However, myocardium truncation did not occur. Likewise, FB truncation was possible for large torsos, but myocardium truncation did not occur. The sensitivity for the different collimations is averaged over the myocardium. The ratios of these sensitivities depend on the torso size, which is adjusted to determine the collimators' performance for different patient types. Multislit-slat is 85% and 34% more sensitive than PB and FB, respectively, for a torso width of 40 cm. These ratios are more favorable for multislit-slat as the torso diameter decreases. In addition, multislit-slat has improved sampling completeness compared with FB in the myocardium region for this orbit. Multislit-slat collimation on a general-purpose scanner offers favorable sensitivity compared with PB and FB for matched average resolutions and for realistic torso sizes. © 2007 IEEE.
