Wahle A, Prause GPM, DeJong SC, Sonka M:


Accurate 3-D Fusion of Angiographic and Intravascular Ultrasound Data.

CAR '98, in:

Lemke HU, Vannier MW, Inamura K, Farman AG (eds):

Computer Assisted Radiology and Surgery (CAR '98).

Proceedings of the 12th International Symposium and Exhibition on Computer Assisted Radiology and Surgery, Tokyo JP, Excerpta Medica International Congress Series, Elsevier

Volume 1165, Page 164-169, 1998

This paper describes the geometrically correct fusion of intravascular ultrasound (IVUS) imaging and 3-D reconstruction from biplane angiograms, resulting in the spatially correct assignment of plaque with respect to vessel curvature (Continued) (Erratum) (Paper) (Images) (Links)

Abstract: Biplane coronary angiography and intravascular ultrasound are complementary methods to assess coronary artery disease. The first one delivers accurate information about the vessel topology and shape, but only restricted data concerning the vessel cross-section. The latter one provides detailed information about the shape and composition of vessel wall and plaque, but fails to consider the geometrical relationships between adjacent images. Thus, a combination of both methods enables judgement of both the longitudinal and cross-sectional geometry of the coronary arteries. In this paper, we present an accurate method for fusion of biplane angiography and intracoronary ultrasound. For decades, selective coronary contrast angiography has been the definite catheter-based imaging technique for the diagnosis of coronary artery disease. Apart from several well-established systems for quantitative analysis of individual coronary stenoses (QCA), methods for reconstruction of the spatial morphology of vessels from biplane projections have been developed. However, the major limitation of X-ray angiography is its restriction to the assessment of the inner lumen. For example, the wall thickness and the composition of the plaque cannot be determined. Recently, intravascular ultrasound of the coronary arteries (IVUS) evolved as a complementary method in cardiovascular diagnosis. Using a catheter with an ultrasonic transducer in its tip, pulled back during imaging, the luminal cross-sections can be determined as well as the wall thickness, and even the composition and orientation of the plaque. A major drawback of IVUS is its in-ability to reconstruct the vessel segment geometrically correct, i.e. considering the vessel curvature when assigning the detected plaque to specific locations.

Up to now, 3-D reconstruction of vessels assessed by IVUS has been performed by simply stacking the images up to a rectangular - but straight - volume. These methods cannot reflect the real curvature of the vessel, which has to be considered since vessels are completely straight only in a minority of cases. Although most algorithms have proven high accuracy in phantom studies using straight objects, they usually fail as soon as the vessel curvature gets strong enough. Our solution for these problems is to employ spatial fusion of the coronary angiography and the intravascular ultrasound data. The 3-D vessel shape and topology are determined by accurate reconstruction from biplane angiograms of known imaging geometry. The vessel cross-sections are extracted from IVUS data acquired in the same session as the angiographic images. For both tasks, our previously developed and validated systems are utilized. The aim of our new approach is the exact assignment of the cross-sectional data to the vessel segment in both location and orientation.

Erratum: In reference [4], a wrong book title were given; please find the correct bibliographic entry here.