Analysis of the Existing Data Fusion Methodologies for 3D Coronary Imaging

Background

Technology in medical imaging has made substantial progress. Miniaturization of medical devices and advances in signal processing has allowed the development of numerous invasive imaging modalities which address the limitations of coronary angiography and provide detailed intravascular coronary imaging. The first methodology available for research and clinical purposes was IVUS and subsequently other modalities ensued including angioscopy, thermography, OCT and most recently NIRS.

However, although these new imaging techniques provide a plethora of data which add substantially to our understanding of the mechanisms involved in the progression of atherosclerosis, none of them allows complete representation of vessel morphology and geometry. For example IVUS does not describe arterial geometry and has limited capability in identifying the type of plaque, while its moderate resolution restricts its ability to detect plaque features associated with increased vulnerability. Some of these limitations are addressed by OCT (e.g. detailed visualization of luminal morphology, accurate characterization of the type of plaque and identification of microstructures [macrophages, microcalcifications, fibrous cap thickness] seen in high risk plaques) which has a higher resolution but again does define vessel’s geometry. In addition, its poor penetration often does not allow imaging of the whole vessel wall (Kubo et al., 2007), (Kawasaki et al., 2006). Angioscopy is restricted to the luminal surface and cannot provide data regarding the vessel wall morphology and plaque burden while the other two available imaging techniques (NIRS and thermography) may be useful for the detection of the high risk plaques as the first allows identification of the lipid pools and the second determination of the plaque heating but are unable to portray vessel’s geometry, anatomy and histology (Takumi et al., 2007), (Moreno et al., 2002).

Taking the best from each modality and integrating the information seems the obvious solution. This can be achieved either through the development of micro-devices that allow simultaneous multi-coronary imaging (e.g. combination of IVUS and NIRS) or through the development of efficient methodologies that permit offline integration of the information provided by the different imaging techniques (e.g. coronary angiography and IVUS, IVUS and computed tomography [CT] etc). A major European project currently in progress called ARTreat implements the fusion of the aforementioned image modalities. Although simultaneous imaging is superior, as it does not require additional processing time and overcomes the problem of accurate co-registration, it is off-line co-registration that is currently the most frequently used approach in hybrid imaging.