To establish sensitivity and specificity of cross-sectional echocardiography for detection of global and regional left ventricular function, a prospective study was performed in 80 patients suspected of having coronary artery disease; 24 h before heart catheterization, cross-sectional echocardiograms were recorded in the apical RAO[right anterior oblique position]-equivalent view; high quality echocardiograms were obtained in 71/80 patients (89%). M-mode echocardiograms of the left ventricle could be performed in 56/71 patients (79%). Based on normal values sensitivity and specificity of cross-sectional echocardiography were calculated in comparison to cineventriculography. For fractional shortening (normal value > 25%) sensitivity measured 46% and specificity 93%. When the E-point septal separation was measured (normal value < 7 mm), sensitivity increased to 73% and specificity was 84%. For detection of increased end-diastolic volume (< 155 ml) sensitivity reached 84% and specificity 98% and for increased end-systolic (< 70 ml) volume it was 86 and 97%, respectively. A depressed left ventricular ejection fraction (> 49%) was found with a sensitivity of 93% and specificity of 100%. Regional left ventricular wall motion, analyzed by an area method, revealed for anterior wall motion a sensitivity of 68% and a specificity of 94%; posterior wall motion sensitivity reached 80% and specificity 96%. Regression equation between the number of pathological segments and left ventricular ejection fraction for cineventriculography was given by Y = -4.06 X + 73.4, r = 0.93 and for cross-sectional echocardiography by Y = -3.69 X + 62.6, r = 0.85. Cross-sectional echocardiography can be used as a screening method to detect impaired left ventricular function. It is superior to M-mode echocardiography. Reduced regional function of the posterior wall can be found with high sensitivity, but depressed motion of the anterior wall may be overlooked.