Automatic Planning And Simulation For Minimally Invasive Aortic Valve Surgery
Vito Giovanni Ruggieri1, Hui Li2, Miguel Castro2, Jean Philippe Verhoye3, Pascal Haigron2.
1Reims University Hospital, Reims, France, 2Rennes University, Rennes, France, 3Rennes University Hospital, Rennes, France.
OBJECTIVE: During the last decades different less invasive approaches to the aortic valve have been developed. However their global diffusion is still limited because minimally invasive aortic valve surgery remains more complex and technically challenging compared to the conventional full sternotomy surgery. Preoperative planning could reduce the risk of anatomical unknowns, the rate of conversion and finally patient’s risk. We developed an automatic tool to identify, optimize and simulate the patient-specific minimally invasive approach to the aortic valve.
METHODS: Angio-CT images were segmented to visualize chest bones and heart structures 3D-meshes. The automatic detection of sternum right bordering and centerline related to the ascending aorta position allowed to identify the best minimally invasive approach (J- or T-shaped mini-sternotomy or right mini-thoracotomy). The intercostal spaces were automatically detected and related to the aortic valve position to choose the best intercostal space. The aortic valve plane and its normal vector were computed and visualized into the 3D mesh including the patient skin.
RESULTS: The automatic tool was tested on 50 datasets coming from patients affected by severe aortic stenosis for validation purposes. The quality of automatic results was verified by 2 surgeons that felt comfortable in minimally invasive setting: 93% of intercostal spaces, 96% of right sternum bordering and 93% of sternum centerlines were judged as perfectly detected by both surgeons. In remaining cases the error was minimal allowing anyway a minimally invasive planning result. The possibility to modify the skin incision length and the sternal/ribs retraction width was also implemented to optimize the access. A warning concerning technical difficulty to anatomical features such as deep ascending aorta or very low aortic valve was included.
CONCLUSIONS: The 3D reconstruction allows to simulate the real surgical vision during minimally invasive approach to the aortic valve in order to choose the best access for each patient and to optimize it. This automatic tool could be helpful especially for surgeons starting their experience in minimally invasive aortic valve surgery.
Back to 2017 Display Posters