Estimation of septal wall thickness by processing sequential echocardiographic images

Non-invasive quantitative analysis of the heart wall thickness is a fundamental step in diagnosis and discrimination of heart disease. Thickness measurements in 2D echocardiographic images have many applications in research and clinic for assessment of wall stress, wall thickening and viability parameters. The measurement of interventricular septum wall thickness by conventional manual method is more dependent on sonographer's experiment.This encouraged researchers to develop a semi-automatic computer algorithm to access interventricular septum segments thickness. We proposed and developed a computerized algorithm for wall thickness measurements in 2D echocardiographic image frames. In this program, wall thickness measurement is based on intensity profile function and adaptive bilateral thresholding operation. For validation, thicknesses of septum base and mid segments were estimated in constituent image frames using proposed technique followed by comparing them with conventional manual results from same images of the cardiac cycle by statistical methods. In our sample image frames [240 corresponding segments; with different range of image quality], a bias of 0.10 mm and 0.12 mm with SD differences of +/- 0.81 mm and +/- 0.72 mm and correlation coefficients of 0.87 and 0.89 were found in base and mid segments, respectively. Interobserver variability using the computer-assisted method [CAM] and conventional manual technique [CMM] were 4.0% and 4.7% for the basal and 2.8% and 3.9% for the middle segments. The method introduced in the present study permits precise thickness assessment of base and mid segments of the interventricular septum wall with high concordance with CMM