Computed Masks in Coronary Subtraction Imaging

In this paper we propose a method to overcome the effects of cardiac and respiratory motion in coronary subtraction imaging. We present the ideas of retrospective gating of masks, where both cardiac and respiratory phases are measured for a set of masks, and are subsequently used in a functional decomposition of motion. Through retrospective gating of masks, we are able to select appropriate images and to perform temporal and spatial processing on them to produce computed masks, subtraction of which should lead to motion-artifact-free images. The computed masks are built from two components: the first includes the time-variant structures related to respiration, such as ribs and soft tissues of the chest wall, and the second incorporates the time-variant structures related to cardiac motion. A preliminary study of the method in the area of digital subtraction angiography produced images which are comparable to but not better than those produced by techniques in current clinical practice: we discuss the reasons for this.