Assessment of stenoses in coronary angioplasty. Inter- and intraobserver variability.

Three observers twice assessed coronary artery lesions of 10 patients in three oblique views using the films obtained before, immediately after, and six months after percutaneous transluminal coronary angioplasty. In order to improve accuracy of the interpretation, the stenoses were traced from the cine projector and the diameters measured with a calibrated magnifying glass. Interobserver and intraobserver variability of the commonly used method assessing only the projection showing the most severe stenosis was compared to that of a method assessing the mean of three projections. The interobserver coefficients of variation were 7.0% for using the one projection showing the most severe stenosis and 6.4% for using the mean of three projections (not significant). The intraobserver coefficient of variation was significantly reduced from 16.0 to 10.5% (P less than 0.0001) by using the mean of three projections. The assessment of three projections instead of one justifies the additional time needed by significantly increasing assessment reliability which is of great importance in evaluating and comparing anatomical results of percutaneous transluminal coronary angioplasty.