[Three-dimensional display and its quantification of exercise stress myocardial tomography using thallium-201]

Abstract

For accurate and stereoscopic delineation of the location and extent of perfusion abnormality by exercise stress myocardial emission tomography using thallium-201, three-dimensional myocardial images (3D image) were reconstructed from ordinary tomograms. We also quantitated perfusion abnormality, with myocardial thickness taken into consideration. We evaluated the usefulness of these 2 methods in patients with coronary artery disease (CAD). Sixty-one patients with 75% or more narrowing of at least one of their major coronary arteries were studied. Myocardial imaging was performed with thallium-201 immediately after exercise stress, and 3 hours thereafter (redistribution) using a rotating gamma camera. We reconstructed 3 routine oblique images and bull's eye maps which included myocardial thallium-201 washout rate maps. In addition to visual interpretation, washout rate abnormality (< 30%) was used as a criterion for exercise-induced myocardial ischemia. For reconstruction of 3D image, we used short-axis images. To identify the cardiac surface, an appropriate count threshold level was determined and image elements exceeding the threshold level were considered the cardiac surface. The heart was observed from 16 points around it and the brightness of the cardiac surface was adjusted in accordance with the distance between the observation point and the cardiac surface. Gradient shading was added and a stereoscopic 3D image was obtained. For the quantitative analysis of the perfusion abnormality, we selected short-axis images. By approximating each short-axis image using 2 circles which contacted the epicardial and endocardial surfaces, we readily calculated the total left ventricular myocardial voxel numbers and the perfusion abnormality voxel numbers. The ratio between these 2 parameters was expressed as % defect. The sensitivity of the 3D image for detecting CAD was 84%, which was similar to that of routine oblique images and the bull's eye method. We also detected the location and extent of perfusion abnormalities and their changes between exercise stress and redistribution in real size and stereoscopically. In patients who had initial myocardial infarction and one-vessel disease but no exercise-induced additional perfusion abnormalities, % defect correlated linearly with the left ventricular ejection fraction (r = -0.85, p < 0.005) and the peak level of the serum cardiac myosin light chain I in the acute phase of myocardial infarction (r = 0.81, p < 0.01). In addition, the relationship between % defect and the number and/or location of coronary artery stenosis in 22 patients with CAD, in whom exercise-induced perfusion defects had completely resolved at redistribution, showed that % defect is a useful indicator for quantitating perfusion abnormalities. In conclusion, the extent of perfusion abnormalities can be expressed in the unit of gram with this method.