Interobserver and intermethod variation in evaluation of regional wall motion of the left ventricle

To determine the variation in the repeated evaluation of quantitative regional wall motion of the left ventricle (LV) from right anterior oblique (RAO) silhouettes, an intra- and inter-observer test was performed in 17 patients without wall motion abnormalities, using nine different methods of regional wall motion (RWM) analysis. The results were: 1. Around the cardiac apex and near the valve plane there are regions where the subjective contour delineation yields regional mean discrepancies up to 4-5 mm (2-3 mm, respectively, near the valves); while in the other regions the mean discrepancies are usually smaller than 1 mm. 2. Absolute regional contour discrepancies, calculated in mm, show only small differences in the various methods, applied for RWM analysis. 3. If these discrepancies instead are calculated as relative changes (in relative percent) of hemiaxial or segmental shortening or area diminution, the method applied is of great significance, since the length of the axes can vary considerably from method to method. 4. From all methods applied for RWM analysis, that of Mathes [10] (radial method) had the smallest average discrepancy. These results suggest, that the LV silhouette may be best defined in the anterior and inferior wall area, while in other regions subjective impressions of contour tracing may produce a considerable variation, which can be further enhanced by the method applied for RWM analysis. Methods that employ relatively large axes of as similar as possible length are therefore preferred.