Quantitative image analysis for planning of aortic valve replacement

Aortic stenosis is the most common and frequent cause of sudden death among all valvular heart diseases. Symptomatic aortic stenosis is considered to be a fatal disease if left untreated. Aortic valve replacement is the mainstay of treatment of symptomatic aortic stenosis. Traditional treatment of severe aortic valve stenosis is aortic valve replacement (AVR) by open-heart surgery. Various novel techniques have been introduced for less invasive AVR, including trans-catheter aortic valve implantation (TAVI) and minimally invasive aortic valve replacement (mini-AVR). For both techniques, careful patient selection and preoperative planning play an important role in the successful deployment of the new prosthetic and reducing the AVR complications. Despite the progress made in the preoperative planning, there is still need for further improvement. In this thesis, we proposed a fully automated aortic root segmentation technique to extract the aortic root surface and landmarks from TAVI patients CTA images which facilitates the automated pre-procedural sizing. Then, we studied the dynamics of the aortic annulus by evaluating the annular dimensions using 4D CTA specifically for TAVI population. Furthermore, an automatic planning tool of the mini-AVR surgery was proposed and a clinical feasibility study is presented that validates our tool retrospectively by analyzing patients who underwent the mini-AVR surgery and compared the tool measurements with the patients on-site difficulty scores. In another study, we reported the comparison results between Non-Contrast MRA and Contrast-Enhanced MRA of the aortic image quality by evaluating the vessel sharpness.