Objective. The aim of this work was to develop a fast and robust (semi)automatic segmentation technique of the aortic valve\narea (AVA) MDCT datasets. Methods.The algorithm starts with detection and cropping of Sinus of Valsalva on MPR image. The\ncropped image is then binarized and seed points aremanually selected to create an initial contour.Thecontourmoves automatically\ntowards the edge of aortic AVA to obtain a segmentation of the AVA. AVA was segmented semiautomatically and manually by two\nobservers in multiphase cardiac CT scans of 25 patients. Validation of the algorithm was obtained by comparing to Transthoracic\nEchocardiography (TTE). Intra- and interobserver variability were calculated by relative differences. Differences between TTE\nand MDCT manual and semiautomatic measurements were assessed by Bland-Altman analysis. Time required for manual and\nsemiautomatic segmentations was recorded. Results.Mean differences fromTTE were âË?â??0.19 (95% CI: âË?â??0.74 to 0.34) cm2 formanual\nand âË?â??0.10 (95% CI: âË?â??0.45 to 0.25) cm2 for semiautomatic measurements. Intra- and interobserver variabilitywere 8.4 Ã?± 7.1% and 27.6\nÃ?± 16.0% for manual, and 5.8 Ã?± 4.5% and 16.8 Ã?± 12.7% for semiautomatic measurements, respectively. Conclusion. Newly developed\nsemiautomatic segmentation provides an accurate, more reproducible, and faster AVA segmentation result.
Loading....