Measuring size dynamics of embolism in aortic aneurysm

Rapture of an aortic aneurysm may be fatal, yet no measurement or index is accepted as a good predictor of such an occurrence. Ultrasound imaging is the most common tool used for diagnosing the aneurysm and monitoring its progress, specifically at the descending aorta. Since in most cases an embolism fills the aneurysm, a hypothesis has been presented whether the embolism reduces the danger of rupture. Here a boundary detection method is presented, which was developed for helping verify the hypothesis. The method allows segmentation of echo ultrasound transverse cross sectional images of an artery, either normal or with an aneurysm, where segmentation of the lumen, embolism and arterial wall is required. A user-defined point must be marked (on one image of the sequence), as an approximate location of the center of the lumen. Polar coordinates are defined, and a search is performed for minimizing a cost function, which includes a function that describes the boundary (based on a limited series of sine and cosine functions) and information from the image intensity gradients along the radii. Results of several patients demonstrate dynamic changes of +-2.5% confirm the applicability of this approach for rapid segmentation of such images, with minimal operator intervention.