Optimal data acquisition for volumetric intracoronary ultrasound.

Abstract

Three-dimensional analysis using intracoronary ultrasound (ICUS) pull-back data provides the unique ability to quantitate lumen and atherosclerotic plaque volumes. Optimal data acquisition parameters for volumetric acquisition were established using simulations on computer phantoms of stenotic arteries. Eleven computer phantoms were generated using cross-sectional area data from quantitative angiography of stenotic coronary arteries. Three methods of data acquisition were simulated: conventional manual pull-back; motorized pull-back; and manual pull-back with measured displacement. Effects of pull-back velocity and cardiac gating on cross-sectional area profiles and volumes were studied. Cardiac gating eliminated errors introduced by vessel deformation within a cardiac cycle. With cardiac gating, pull-backs with mean velocities up to 1.2 mm/sec allowed reconstruction of cross-sectional area profiles within 5% RMS error. With faster pull-backs, cardiac gating resulted in sparse spatial sampling and significant errors in cross-sectional area profiles. The accuracy of both motorized and measured required equal displacements of the catheter proximal and distal ends. This assumption was validated with in vitro experiments where X-ray fluoroscopy was used to measure the displacement of the imaging tip. Excellent correlation was found between the two displacements (r = 0.99). Finally, slow pull-backs were performed by 3 operators, and pull-back velocities were measured. It was found that mean pull-back velocities as low as 0.8 mm/sec were achievable. From our simulations, we predict that accurate volumetric analysis requires cardiac gated, calibrated, slow (< 1.2 mm/sec) pull-backs.