Computer Vision News - June 2022

56 Congrats, Doctor! Total hip arthroplasty (THA) has become one of the most common surgeries worldwide as it reduces the pain and restores the hipmobility of around 1million patients every year. However, it is still at risks of complications, such as dislocation, with an incidence varying from 1% to 5%. Dislocation usually occurs by lever arm effect when the femoral stem impinges the acetabular cup and is directly related to the correct positioning of the implants. The goal of my thesis was to propose a global approach that considers patient-specific functional parameters to optimize the implants positioning in order to minimize the risks of prosthetic instability. Determine a Functional Orientation for the Acetabular Cup In order to reduce the risks of dislocation, the acetabular cup must be oriented regarding the pelvis inclination. However, the standard recommendations do not take into account the mobility of the patient: the hip range of motion (ROM) and the pelvis tilting during activities of daily life. Therefore, I developed a functional safe zone, adaptable to the pelvic tilts of each patient in different positions, that would ensure the absence of prosthetic impingement and thus reduce the risks of dislocation (Figure 1). This safe zone is based on the 3Dmodelling of the implants motion and is simple and fast to compute. The pelvic tilts are acquired using a navigated ultrasound device developed in our lab, which avoid ionizing imaging techniques. Moreover, it can be used to analyze the impact of prosthetic parameters regarding the patient-specific pelvic tilts and ROM, in order to personalize the pre-operative planning and estimate the margin of error for the prosthesis implantation. Acquire the Patient Reference Plane During Surgery Once the pre-operative planning is determined, the surgeon must be able to easily position the implants. One challenge is to acquire the patient-specific Anterior Pelvic Plane (APP), which is the plane of reference for the cup orientation, with the patient already installed for surgery. However, most THAs are performed with the patient lying on his side, which makes the APP acquisition difficult as one anatomical landmark is found against the operating table and covered by the patient abdominal apron. In my thesis, I proposed a new method to determine the APP from anatomical landmarks of only one side of the pelvis acquired using navigated ultrasound (Figure 2). To do that, I built a statistical shape model of the pelvis, able to reconstruct the patient’s whole pelvis from a few anatomical landmarks. Once the model is fitted to the target landmarks, the APP can be automatically determined. This approach is quick, low-cost and compact, which makes it convenient to use in the operating room. I amnow looking for postdoc opportunities to continueworking onmedical and translational research. Feel free to contact me at azilizgp@gmail.com. Ken ar c'hentañ !* * “See you soon!” in Breton Aziliz Guezou-Philippe has recently completed her PhD at the Laboratory of Medical Information Processing, University of Western Brittany, Brest, France. Her research focuses on biomechanical and statistical shape modelling, statistical analysis and point cloud registration. She aims to help surgeons improve their surgery by developing computational tools that are easy to integrate in clinical routine. She’s on the market for a postdoc position and she’s awesome. Congrats, doctor Aziliz!