Evaluation of differences in attenuation pattern for myocardial perfusion scintigraphy between CZT and conventional gamma cameras

The new CZT camera has a different technology compared to conventional scintillation gamma cameras. The CZT camera have pinholes instead of parallel hole collimators, a static gantry instead of a rotating one and detection material made of a semiconductor in- stead of a scintillation crystal. The images arising from the CZT camera have a difference in appearance compared to the images produced by the conventional technique. The extent of the difference in the attenuation pattern has not yet been investigated. In this master thesis the differences of the attenuation pattern between the CZT and conventional camera used in myocardial perfusion SPECT (MPS) imaging were investigated by quantifying the size, depth and distribution of the attenuation artifact. Phantom and patient measurements were performed with a CZT camera (GE NM 530c) and a conventional gamma camera (GE Ventri). A cardiac phantom including a lung in- sert (Data Spectrum Corporation) was used in the phantom measurements. Twenty-two patients underwent a 2-day stress-rest protocol (4MBq/kg, stress and rest) and were exam- ined with both cameras approximately 1.5 hours after administration of 99mTc-tetrofosmin. All images were corrected for attenuation by an externally acquired CT image. Recon- struction parameters recommended by the manufacturer were used. By using the software Segment v 2.0 the extent of the attenuation artifact for each camera was defined by a delin- eation in the ratio between the non-attenuation corrected (NAC) and attenuation corrected (AC) phantom images. The ratio image NAC/AC was used to quantify the depth, size and distribution of the attenuation artifact, in both phantom and patients. The result showed that the localization of the attenuation artifact was shifted counter- clockwise from the inferolateral wall to the lateral wall for the CZT camera compared to conventional cameras in both patient and phantom measurements. Based on the man- ual delineation of the attenuation artifact in the phantom ratio NAC/AC, the attenuation thresholds were 0.90 and 0.85 for the CZT and conventional cameras, respectively. These threshold values were then applied on the ratio NAC/AC for the phantom and patient mea- surements to obtain the relative depth, size and distribution of the attenuation artifact. For the phantom the relative depth was 0.73 and 0.67 for CZT and Ventri, respectively. For the patients the depth difference was statistically significant (p<0.05), with values of 0.72 and 0.68 CZT and Ventri, respectively. The relative size of the attenuation artifact was larger for the CZT compared to the Ventri camera, 28 % and 19 % for phantom and 23 % and 15 % for patients (p<0.05). There was a statistical difference between the cameras with respect to the distribution of the attenuation artifact for both phantom and patients (p