M.Sc Thesis

M.Sc StudentRappaport Dan
SubjectMyocardial Regional Tissue Strain Estimation by Local Tissue
Tracking in B-Mode Echocardiograms
DepartmentDepartment of Biomedical Engineering


The assessment of regional motion of myocardial tissue carries meaningful clinical information such as strain and strain rate. These measures are considered to reflect well the left ventricle performance, which is routinely studied as part of many cardiac evaluation tests. To date, the only established method for myocardial regional motion analysis is based on the Doppler effect (TDI). However, it has inherent limitations such as angle dependency and spatial vs. temporal resolution tradeoff that restrict its accuracy and use. This study was designed to test the feasibility of accurate regional motion assessment using a time domain method (SAD tracking) by applying a statistical approach to the tracking results. Acoustic markers are characterized, and those stable for tracking are selected. A weighting index for each marker tracking and a spatial curve fitting are calculated. Here, to prove the feasibility of the approach, a simplified problem was addressed, that of tracking of an approximate rectangle ROI that preserve its shape in time. The septum (in four chambers view) is a good example for such a region. The method requires a video stream as input, of at least 40 frames/sec, applicable to any commercial system. After being tested and tuned in simulations, the method was applied to in-vivo datasets. The longitudinal peak systolic strain of 4 normals and 4 patients suffering from dysfunction in the apical segment in septum were compared. The calculated peak systolic strain values were 0.15+-0.023 and 0.099+-0.008, respectively (p<0.05). The successful results obtained, which showed significant advantages over TDI, encourage further development.