Intraoperative decision-making for tumour resection is based on the surgeon’s experience, supported by macroscopic, endoscopic, microscopic or robotic and navigation-assisted examination of the tumour. However, this approach does not allow for a precise definition of the tumour borders and can lead to incomplete tumour resection and thus poor survival of the patient. Intraoperative navigation is based on preoperative imaging. Due to the continuously changing situation caused by the operation, the data is increasingly outdated and thus inaccurate. Preliminary work by the project partners shows that multimodal marker-free imaging is suitable for accurately determining tumour boundaries. This project aims to use multimodal imaging and artificial intelligence (AI)-based real-time analysis to continuously record the current tumour boundary in the operating theatre and visualise it to the surgeons, and also to use the analysis to provide haptic feedback so that the surgeons can immediately use the information to make decisions.
In this complex project, sensorised surgery means two things: on the one hand, sensors are used for more precise tissue analysis and, on the other hand, for optimised information acquisition and transmission for surgical tactics. In this way, sensorised surgery can not only realise personalised surgery precisely for the patient to be operated on, but above all increase the rate of complete tumour resections while sparing healthy tissue as much as possible. The realisation of an intuitive sensory interaction between machine and surgeon as a breakthrough will increase the acceptance of such new life sciences technology of personalised medicine by surgeon and patient, which in turn will accelerate the transfer into everyday care. The challenges lie in the reduction of potential model bias due to low training data and the provision of information about the uncertainty in the analysis, which should be visualised to the surgeons in an appropriate way. Real-time operation also requires model compression without compromising the quality of the results.