Quantification of Left Ventricular Function by Applying Signal Averaging to Echocardiographic Automated Border Detection

myocardium may be different from the infarcted myocardium; however, clin­ ical studies are still rare. In order to investigate the relationship of texture properties of the infarcted myocardium and patency of infarct-related artery (IRA). we have developed the computer program for tissue characterisation of echocardiographic images. The program uses spatial gray level cooccurrence method for texture char­ acterization to compare tissue samples taken from normal and infarcted ar­ eas of myocardium. The program is menu driven and has two main operating modes: auto-diagnose mode and user-diagnose mode. In the auto-diagnose mode, after the decision making procedure that incorporates human expe­ rience and knowledge, the program gives the prediction of future recovery. In the user-diagnose mode, the program provides numerical and graphical presentation of obtained results, and may be used as an analysis tool during echocardiographic examinations. The program runs under MS-DOS or Microsoft Windows on IBM 386/486 compatibles. It accepts ultrasound images stored on disk by commercial frame grabbers. As an inputto the algorithm, the user selects two areas to be analyzed: a normal one and an infarcted one. Output data can be transferred to Windows based spreadsheet or database programs using Clipboard fea­ ture. The validation study compared two groups of pts after acute myocardial infarction examined before hospital discharge: the group with occluded and the group with patent IRA. It demonstrated myocardial texture changes in the infarct zone in pts with occluded IRA, whereas no difference between infarcted and normal tissue was detected in pts with successful reperfusion. These data indicate that the program may help in prediction of successful reperfusion in the early post-infarction period.