Cardiac phase-related variability of border detection or densitometric quantitation of postangioplasty lumens

Abstract

We applied an automated computer program capable of simultaneous geometric (through border detection) and densitometric quantitation of digital angiograms for evaluation of the results of percutaneous transluminal coronary angioplasty (PTCA) in different phases of the same cardiac cycle. Digital subtraction coronary angiograms (DSA) of 28 patients who had undergone PTCA to a total of 30 lesions, were analyzed in diastole, in systole, and in the middle of the cardiac cycle to test the variability in coronary quantitation resulting from random frame selection relative to cardiac phase. Before PTCA there was a low degree of variation between measurements obtained from the same lesion in different phases of the cardiac cycle, in both geometric (coefficient of variation between cardiac phases = 4.2%) and densitometric (coefficient of variation between cardiac phases = 5.1%) quantitation. After PTCA, however, there was a wider variation of values in different cardiac phases, which predominated in the densitometric measurements (coefficient of variation between cardiac phases = 33.6%, compared to 20.6% for geometric measurements). There was less agreement between different post-PTCA phases in densitometry, and discrepancies as large as 47% could occur in densitometric evaluation of the stenotic areas when different phases of the cycle were used. We conclude that border detection or densitometric quantitation of the postangioplasty lumens is subject to greater variation resulting from random frame selection relative to cardiac phase, as compared to preangioplasty assessment. This variation predominates in densitometric quantitation, which seems to be dependent not only on the radiographic projection but also on the cardiac phase. The usefulness of densitometric techniques for the evaluation of PTCA results appears to be questionable.