The Impact of Additional Sampling on Hot-Spot Contrast for DE-SPECT

DE-SPECT is a CZT-based system for assessing Peripheral Arterial Disease [1], [2]. Each CZT module is 4 cm x 4 cm x 1 cm. They are configured in a half-populated 4 x 4 array per panel - with 6 panels in the scanner - such that every other slot is populated. This configuration reduces cost, but also provides the opportunity to have two different collimator configurations, with one in front of the detectors at a time. An axial actuator switches between the two modes. Herein, the tradeoffs between sampling and sensitivity when utilizing that actuator for additional sampling positions are considered by using contrast to noise as the metric for evaluation. A metric of complete sampling was evaluated as a function of axial shift between acquisitions. In addition, a hot-spot phantom with hot spots configured at 27 positions on a 3x3x3 grid with 4 cm pitch was used to evaluate contrast. Three phantoms were considered: 0.6, 0.8, and 1.0 cm hot spots. Ensembles of 25 noise instances for shifts of 0.0, 0.5, 1.0, and 1.5 cm were considered. Three noise levels were considered: 640k, 2560k, and 10240k total count equivalents, where a count equivalent is a count in the nominal position, but considered the reduced efficiency for shifted projections; thus, equivalent duration was used. The optimal sampling occurs at 1.5 cm displacement. Optimal contrast occurred at 0.5 cm axial shift, for the shifts considered, for the 0.6 cm and 0.8 cm spheres. The advantage was more ambiguous for 1.0 cm spheres. Sensitivity drops off quickly with axial shift, limiting the advantage of additional sampling. The optimal tradeoff occurs between 0.0 cm and 1.0 cm. This study considered steps of 0.5 cm so far. In conclusion, hot-spot contrast can be improved by the addition of additional sampling provided by small axial shifts from the nominal detector position.