THE ASSESSMENT OF LEFT-VENTRICULAR FILLING DYNAMICS USING AN ONLINE AUTOMATIC BORDER DETECTION ALGORITHM - COMPARISON WITH CINEVENTRICULOGRAPHY

Citation
Pm. Sapin et al., THE ASSESSMENT OF LEFT-VENTRICULAR FILLING DYNAMICS USING AN ONLINE AUTOMATIC BORDER DETECTION ALGORITHM - COMPARISON WITH CINEVENTRICULOGRAPHY, Echocardiography, 12(6), 1995, pp. 559-569
Citations number
NO
Categorie Soggetti
Cardiac & Cardiovascular System
Journal title
ISSN journal
07422822
Volume
12
Issue
6
Year of publication
1995
Pages
559 - 569
Database
ISI
SICI code
0742-2822(1995)12:6<559:TAOLFD>2.0.ZU;2-G
Abstract
An echocardiographic system has been developed that performs automatic endocardial border detection. and instantaneously calculates and disp lays a waveform of left ventricular cavity area versus time. The purpo se of this study was to compare measurements of left ventricular filli ng dynamics from automatic border detection. echocardiography with sim ilar measurements from cineventriculography. Thirty-three patients und ergoing cardiac catheterization. had automatic border detection echoca rdiography performed within 45 minutes of cineventriculography. Ten pa tients had normal catheterization findings and 23 had cardiac disease. The automatic border detection waveforms generated from two echocardi ographic views were measured to determine the fraction of filling occu rring during the early diastolic rapid filling phase and during the fi lling phase resulting from atrial contraction. Similar fractions were derived from curves generated from frame-by-frame measurements of cine angiographic volumes. Results were analyzed by correlating echocardiog raphic and cineventriculographic results, and by a (limits of agreemen t analysis were +/- 2 standard deviations of the mean difference betwe en echocardiography and cineventriculography). There were significant correlations between echocardiography and cineventriculography for eac h of the parameters studied. The best results were obtained for the ap ical four-chamber view (rapid filling fraction r = 0.72, P < 0.0001, a trial filling fraction r = 0.56, P < 0.001). Differences in. filling p atterns between normal and abnormal patient groups detected by cineven triculography were also detected by automatic border detection. echoca rdiography. However, broad limits of agreement were observed, that may limit the ability of the automatic border detection system to reliabl y predict cineventriculographic results in an individual patient. Auto matic border detection echocardiography can provide information about left ventricular filling dynamics that is similar to that obtained fro m frame-by-frame analysis of cineventriculograms. However the variabil ity in the results may limit the application of the technique in indiv idual patients.