Automated coronary artery tracking of low-axial resolution multi slice CT

Vessel tracking is a topic that has had extensive coverage when dealing with high axial resolution data, obtained from modalities requiring patients to be exposed to greater levels of radiation dose, possibly to the detriment of their health[10]. The use of low axial resolution data such as multi-slice CT(MSCT) used in coronary artery disease screening is becoming widely used clinically to provide the same imaging capabilities of previous modalities but with faster acquisition, often within a single breath hold[2]. Such modalities must counteract the loss of image clarity, detail and partial volume effects with benefits to the patient such as faster acquisition and lower dose exposure. Tracking of the arteries can be beneficial for pinpointing objects within such as calcified plaques, stents and other foreign bodies or pathology. In addition to differentiating them from similar, difficult to discern objects in neighbouring regions. To this end, a fully automated artery tracking system is presented to be validated against re-sampled expert extracted vessel centrelines. Implementations include an automated heart isolation algorithm using active contours and region voxel analysis, fully automated coronary artery root detection and segmentation followed by proximal, lumen and distal tracking of the main coronary arteries. Early results show a good degree of tracking ability despite the poor axial resolution.