The effect of object size on the sensitivity of single photon emission computed tomography: comparison of two CZT cardiac cameras and an Anger scintillation camera.

BACKGROUND

Heart sizes vary greatly across the spectrum of patients referred for myocardial perfusion imaging. We therefore performed a phantom study to explore under controlled circumstances how count rates change when different volumes containing the same amount of activity are scanned. Two dedicated cadmium-zinc-telluride cameras, the D-SPECT (Spectrum Dynamics, Caesarea, Israel) and Discovery 530c (D530c, GE Healthcare, Haifa, Israel), and the conventional SPECT Anger (A-SPECT, GE Healthcare, Haifa, Israel) camera are included in the study.

METHODS

Different heart sizes were represented by syringes of various column heights mimicking a range of cardiac diameters. Syringes with fixed activity were scanned at five different volumes by successively adding non-radioactive water to the syringes. This procedure was repeated five times on each of the three cameras. Raw count rates were recorded for each scan to determine whether count rates changed with syringe column height.

RESULTS

Using mixed-effect regression modeling, a linear relationship was found between count rate and water column height. For the D-SPECT, D530c, and A-SPECT, the changes in count rate for each centimeter increase in water column height were -1.75, +0.28, and -0.022 kilocounts per min per MBq, respectively (95% confidence intervals -1.89 to -1.61, 0.19 to 0.36, and -0.035 to -0.009); all effects are significantly different from each other and significantly different from zero. Average coefficients of variation were 0.080, 0.028, and 0.009.

CONCLUSIONS

The D-SPECT demonstrated a significant progressive increase in count rate related to decreasing size of the imaged object. D530c count rate increased slightly with increasing column height. The Anger SPECT showed minimally increased count rates with decreasing column height, an order of magnitude smaller than the D-SPECT based on their relative coefficients of variation.