Prediction of angiographic coronary disease and mortality with a cadmium-zinc-telluride camera: a comparison of upright and supine ejection fractions and left ventricular volumes.

Abstract

Introduction: Perfusion imaging strongly predicts coronary artery disease (CAD), whereas cardiac volumes and left ventricular ejection fraction (LVEF) strongly predict mortality. Compared to conventional Anger single-photon emission computed tomography (SPECT) cameras, cadmium-zinc-telluride (CZT) cameras provide higher resolution, resulting in different left ventricular volumes. The cadmium-zinc-telluride D-SPECT camera is commonly used to image in the upright position, which introduces changes in left ventricular loading conditions and potentially alters left ventricular volumes. However, little or no data exist on the predictive value of left ventricular volumes and ejection fraction when acquired in the upright position. We investigated models for the prediction of CAD and mortality, comparing upright and supine imaging.

Methods: A retrospective study of patients with upright/supine stress and rest imaging and coronary angiography within 3 months was performed. Univariate and multivariable analyses were performed to predict abnormal angiograms and all-cause mortality.

Results: Of the 392 patients, 210 (53.6%) had significant angiographic CAD; 78 (19.9%) patients died over 75 months. The best multivariable model for CAD included the supine summed stress score and supine stress LVEF, with an area under the receiver operating characteristic of 0.862, a sensitivity of 76.7%, and a specificity of 82.4%, but this model was not statistically superior to the best upright model. The best multivariable models for mortality included age, diabetes, history of cardiovascular disease, and end-systolic volume, with the upright and supine models being equivalent.

Discussion: Angiographic CAD was best predicted by the supine summed stress score and LVEF but was not statistically superior to the next-best upright model. Mortality was best predicted by end-systolic volume in combination with age, diabetes status, and cardiovascular disease status, with equivalent results from the upright and supine images.