Attenuation correction in myocardial SPECT using gadolinium-153 line sources or low-dose CT: Comparison of phantom and patient examinations

Abstract

Background: The aim was to compare, both in phantoms and in patients, attenuation correction (AC) in myocardial SPECT using either 153Gd transmission line sources or low-dose CT. Methods: SPECT was performed using a cardiac phantom (Capintec Inc, NJ), an anthropomorphic phantom with a medium sized latex breast (Data Spectrum Corp, NC) and in 10 patients (mean age: 57 years, 9 men) using 99mTc-sestamibi and 1) a Millennium MG (GEMS) dual head camera with 153Gd transmission line sources for AC (interleaved), and 2) a Hawkeye (GEMS) with low-dose CT used for AC (sequential). Phantom studies were performed with normal myocardium, with defects and with simulation of a hot gall bladder. In both systems scattered photons were acquired in a window below the 99mTc peak. All images were processed using the same iterative algorithm with scatter subtraction and AC. Images obtained with Gd attenuation maps and CT attenuation maps were compared with regard to uniformity and effective contrast (phantom studies) and with regard to impact on defects in 5 regions of the left ventricle (anterior, lateral, inferior, septal and apical, consensus reading without knowledge of clinical or angiographic data). Results: Phantom studies were comparable with respect to uniformity (normal myocardium & hot gall bladder exp.) and effective contrast (ischemic defects) using the two systems. In all 10 patients, interpretation of non-corrected images was similar on both systems. In 7 of the patients patterns were identical following AC with Gd or low-dose CT: AC did not change interpretation in 3 cases, in 4 cases a total of 3 inferior, 1 anterior and 1 apical defect were eliminated by AC. In the remaining 3 studies, there were differences between the 2 AC systems: Using Gd-based AC, 1 of 2 inferior defects disappeared compared to none with low-dose CT AC. Furthermore, an anterior and 1 apical defect was eliminated by Gd AC but not by low-dose CT AC. Finally, 1 anterior defect was introduced with low-dose CT AC whereas no new defects were introduced by the Gd-based AC. Conclusions: Both Gd and CT AC have great impact on myocardial SPECT. Although the systems perform similar in phantoms, when compared on a head-to-head basis in patients the two systems in some cases create important differences in the corrected images.