In this investigation, we studied the feasibility of 3D-reconstruction from 2D cross-sectional intracoronary ultrasound images. A computer-aided, automated 3D-reconstruction was used to generate cylindrical and sagittal format of vessels in vitro (n = 9) and in vivo (n = 48). Ultrasound 2D-images were acquired with a 20 MHz mechanical intracoronary ultrasound catheter (Boston Scientific). A slow pullback (about 1 mm/s) of the catheter was performed during continuous recording of the ultrasound images. The recorded 2D-images were then fed to an image processing computer. Correction and scaling of the axial vessel dimensions was performed by the assessment of the catheter position in the simultaneously recorded biplane fluoroscopy. Digitized data were then processed to 3D-images with use of voxel space modeling. The 3D-views from any plane can be reconstructed. The in-vitro study demonstrated that the reconstructed images were able to correctly portray the pathological changes of the vessel wall in all specimen as proved by pathologic examination. In the in-vivo study, 3D-reconstruction provided not only a spatial visualization of the coronary arterial disorders (such as coronary aneurysm, coronary dissection, spontaneous plaque rupture etc.) but also provides the potential to quantify the mass of lesions. By combining sagittal and cylindrical views, 3D-reconstruction enables longitudinal and orthogonal imaging of the both the vessel lumen and vessel wall, therefore, it has the advantage of cross-sectional viewing as obtained from intracoronary ultrasound and the longitudinal viewing as derived from coronary angiography. These preliminary results of the study indicate that 3D-reconstruction of coronary segments is a promising technique for studying coronary artery disease. Analysis can be based not only on single or multiple cross-sectional images but also on vessel segments facilitating serial studies as for instance studies assessing the progression and regression of atherosclerosis.