Cinevideodensitometric analysis of the effect of coronary angioplasty on coronary stenotic dimensions.

The accuracy and reproducibility of caliper and cinevideodensitometric measurements of coronary stenotic dimensions were compared in radiographic phantom models and in coronary arteriograms of 28 patients undergoing coronary angioplasty. Projected, single-plane coronary cine frames were analyzed by a computer-assisted videodensitometric method, which measures stenotic cross-sectional area without assumptions about lesion geometry. The accuracy (2.4%) and precision (+/- 1.9%) of cinevideodensitometry for measuring percent area stenosis in Plexiglas models of eccentric stenotic lesions was superior to the accuracy (24.7%) and precision (+/- 5.4%) of caliper measurements. Interobserver variability was significantly (p less than 0.05) better for cinevideodensitometric (r = 0.98; SEE = 6.4%) than for caliper measurements (r = 0.87; SEE = 13.1%). After angioplasty, percent diameter stenosis measured by calipers fell from 70 +/- 12% to 30 +/- 15%. Mean percent area reduction measured by cinevideodensitometry fell from 89.1 +/- 8% to 40.1 +/- 22% and stenotic area increased five-fold, from 0.59 +/- 0.5 to 3.47 +/- 1.6 mm2. Pre and post PTCA gradients did not correlate with lesion dimensions. Cinevideodensitometric measurements of absolute stenotic dimensions were more reproducible than relative measurements expressed as a percentage, due to the tapered caliber of normal arterial segments. Thus, cinevideodensitometric measurements were more accurate and reproducible than caliper measurements. The angiographic effects of coronary angioplasty are best measured by cinevideodensitometry, because residual lesions post PTCA are often eccentric, have indistinct margins, and are better characterized by changes in area than by changes in diameter.