Developing a software system to improve the management and outcome of knee replacement surgery by combining radiographic image information with clinical and patient data (360G-Wellcome-223267_Z_21_Z)

Knee replacement surgery (KRS) is commonly performed (~100,000/year in UK) in patients with severe osteoarthritis. Despite a good outcome for most patients, 20% of patients experience chronic pain afterwards and 4.4% require revision surgery within 10 years for a variety of reasons. My goal is to develop a software system to predict poor outcome and early revision in KRS. The system will facilitate (i) informed decision-making about surgery, enabling targeted interventions such as preventive pain management; and (ii) the identification of patients that need closer follow-up post-surgery, enabling more remote monitoring of patients. Radiographic knee shape has been shown to be a predictor for the outcome of KRS. I have developed the BoneFinder®-technology, an internationally leading technology to automatically outline structures in radiographs. Using machine learning methods, a system will be developed to automatically measure the knee in pre-/post-operative radiographs, and to predict poor outcome based on the radiographic measurements, clinical and patient-reported data. To facilitate implementation of such a system for patient benefit, it will serve as a use case to create a sustainable toolkit for translating medical imaging solutions in particular and digital healthcare solutions more broadly; aiming to reduce the translational gap between research and clinical practice.