Endoprosthetic implants are used to replace hips or knees and have become increasingly common in recent decades.In addition to the initial surgery to insert the prosthesis, further operations (revisions) become necessary when the prosthesis shows signs of wear or has loosened over time. However, these revisions have proven to be a cause for concern as they require more time than the initial surgery itself, increasing the risk of complications due to physical and mental exhaustion. Another major concern is bone fragmentation, which can occur if the bone cement holding the existing prosthesis in place cannot be successfully removed to assess and replace the prosthesis. Currently, medical chisels or drill wires are used to remove the bone cement and separate the prosthesis from the bone, placing a significant burden on the surgeon.
The OrthoJet project aims to develop an innovative cyber-physical system for bone cement removal using a pulsed water jet that meets the unique and stringent requirements of its medical application.
The pulsed water jet enables selective, precise and efficient removal of bone cement with less stress on the bone and tissue surrounding the prosthesis. This directly reduces the time required per revision, thereby reducing the risk of complications and the need for bone fragmentation.
To assist the operator during the revision, the OrthoJet uses structure-borne sound sensors to measure the vibrations induced by the water jet. These measurements allow classification of the material being removed and its effectiveness using machine learning. The neural networks are trained to classify whether the water jet has no effect, plastically deforms the bone cement, effectively removes it or instead removes bone and tissue. This information is fed back directly to the operator, who can adjust the target accordingly. In addition, the system introduces a second layer of safety by automatically turning off the OrthoJet if it is pointed at tissue or bone for too long. The operator's workload is further reduced by automatically adjusting the jet pressure and/or pulse rate for optimal bone cement removal.
The research and development is carried out by a consortium of three industrial partners (endocon GmbH, Oskar Moser technische Edelsteine GmbH and Hotho Data GmbH) and three research institutes (Heidelberg University Hospital, TU Chemnitz and Institute for Information Processing (TNT)).
TNT is responsible for the development of the sensor signal pre-processing algorithms and the deep learning approach for assessing the current state of the removal process.
This project is supported by the Federal Ministry of Education and Research (BMBF), Germany (grant no. 13GW0586F).