In-vivo validation of videodensitometric coronary cross-sectional area measurement using dual-energy digital subtraction angiography.

Previous studies indicate that conventional geometric edge detection techniques, used in quantitative coronary arteriography (QCA), have significant limitations in quantitating coronary cross-sectional area of small diameter (D) vessels (D < 1.00 mm) and lesions with complex cross-section. As a solution to this problem, we have previously reported on an in-vitro validation of a videodensitometric technique that quantitates the absolute cross-sectional area including small vessel diameter (D < 1.00 mm) and any complex shape of the vessel cross-section. For in-vivo validation, plastic tubing (5-8 mm long) with different shape complex cross-section with known cross-sectional area (A = 0.8-4.5 mm2) were percutaneously wedged in the coronary arteries of anesthetized pigs (40-50 kg). Contrast material injections (6-10 ml at 2-4 ml/sec) were made into the left main coronary artery during image acquisition using a motion immune dual-energy subtraction technique, where low and high X-ray energy and filtration were switched at 30 Hz. A comparison was made between the actual and measured cross-sectional area using the videodensitometry and edge detection techniques in tissue suppressed energy subtracted images. In eighteen comparisons the videodensitometry technique produced significantly improved results (slope = 0.87, intercept = 0.24 mm2, r = 0.94) when compared to the edge detection technique (slope = 0.42, intercept = 1.99 mm2, r = 0.39). Also, a cylindrical vessel phantom (D = 1.00-4.75 mm) was used to test the ability to calculate and correct for the effect of the out of plane angle of the arterial segment on the cross-sectional area estimation of the videodensitometry technique. After corrections were made for the out of plane angle using two different projections, there was a good correlation between the actual and the measured cross-sectional area using the videodensitometry technique (slope = 0.91, intercept = 0.11 mm2, r = 0.99). These data suggest that it is possible to quantitate absolute cross-sectional area without any assumption regarding the arterial shape using videodensitometry in conjunction with the motion immune dual-energy subtraction technique.