| Popis: |
Objective: This work evaluates a new approach for calculating the left-ventricular (LV) early filling propagation velocity (VP) from color M-Mode (CMM) echocardiograms using wavelet analysis. Unlike current approaches, the method requires no assumptions, user inputs, or heuristic conventions. Background: Current methods for measuring VP using CMM echocardiography do not account for the spatiotemporal variation of the filling wave propagation velocity. They are instead confined by empirical assumptions and user inputs that significantly hinder the accuracy of VP, subsequently limiting its clinical utility. Methods: We evaluated three methods for measuring LV early filling VP: conventional VP, the strength of propagation (VS), and VP determined from the most energetically important wave (Peak-VW), using 125 patients (Group A) with normal filling (n=50), impaired relaxation (n=25), pseudonormal filling (n=25) and restrictive filling (n=25), and in 69 patients (Group B) with normal (n=32), dilated (n=15), and hypertrophic ventricles (n=22). Results: Peak-VW most accurately distinguished normal left ventricles from diseased ones. Specifically, for Group A which spanned the stages of diastolic function, using receiver operator characteristics (ROC) and measuring their corresponding area under the curve (AUC), the AUC for Peak-VW was 0.92, versus 0.62 for conventional VP, 0.63 for VS and 0.58 for intraventricular pressure difference (IVPD). These correspond to a 50-70% improvement in classification ability. Similar improvements were found using Group B. Conclusion: A new determination of the LV early filling using wavelet analysis, Peak-VW, may provide a more accurate evaluation of diastolic function than the standard method of determining Vp and enable better diagnostic classification of patients suffering with diastolic dysfunction. |
