Automated 3D planning of high tibial osteotomies optimised for specific morphotypes of the knee - smart HTO

• Applicants: Privatdozent Dr. Julian Fürmetz; Professor Dr.-Ing. Klaus Radermacher
• Subject Area: Orthopaedics, Traumatology, Reconstructive Surgery ; Medical Physics, Biomedical Technology
• Term: since 2026
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 574977868

Open wedge high tibial osteotomy (owHTO) is an established method for the treatment of unicompartmental (medial) osteoarthritis (OA) in varus knees. The surgical procedure aims to shift the load from the medial to the lateral area of the knee joint in order to slow down the progression of medial OA. While it is of high clinical relevance, the procedure is technically demanding. Even with experienced surgeons, there is a significant risk of over- or undercorrection when planning is based on conventional 2D X-ray diagnostics due to rotational changes in the leg axis, for example. Although detailed 3D planning enables more precise geometric analyses and biomechanical simulations, it has yet to become established in everyday clinical practice due to the significant effort required and the lack of automation. Consequently, patient-specific changes in morphology and biomechanics cannot yet be adequately considered in owHTO planning and execution. This research proposal therefore aims to develop an automated, AI-supported 3D planning workflow that uses an artificial neural network (ANN) for the automatic segmentation and identification of anatomical landmarks in order to determine patient-specific 3D morphotypes. To investigate the relationship between pre- and postoperative morphology, an adapted knee simulation model is used in which different osteotomy cutting strategies are implemented virtually and their effects on leg alignment, joint pressure distribution, and kinematics are quantified. The simulated biomechanical parameters are validated by in-vitro studies and finally compared with prospective clinical data – specifically, patient-reported outcome parameters (PROMs: KOOS, IKDC) – to ensure the predictive accuracy and clinical relevance of our workflow. By combining automated AI-assisted 3D morphotype classification, osteotomy definition, and biomechanical simulation for the first time, we will create a standardised, patient-centric planning process that minimises miscorrections, increases surgical safety, and optimises functional outcomes. This reproducible workflow will not only improve the efficiency and precision of owHTO, but it could also be transferred to other knee operations to improve planning quality and clinical outcomes sustainably.

• DFG Programme: Research Grants