Providing real time feedback to surgeons is of major importance. Some issues using imaging modalities for real time feedback are the low frame rate, low compatibility, and prolonged exposure to radiation. These issues limit the quality of real-time feedback given to surgeons. Low frame rate and low compatibility make the procedure harder and will increase the intervention time and decrease the accuracy. The increased intervention time causes prolonged exposure to ionizing radiation in case of X-ray based image guidance. With the use of real data it is proven to be possible to train a model that generates images of the liver that follow the pattern of breathing of the real data till a point where it doesn't need any real time input. One major drawback of using this method is that it can't provide real time feedback. These generated images of the liver can't show what is happening real time (for example feedback of the needle insertion).
This research is focussed on developing a needle insertion model that can predict liver tissue deformation based on the cutting of the needle trough the liver tissue. First it is important to investigate a model that can predict the liver tissue deformation accurately. This will be done using finite element modelling (FEM) in a 4-step model. The model starts with no interaction between needle and liver tissue. Next the needle starts pressuring in the liver tissue, but no cutting occurs. Next the the needle will cut through the liver tissue and the liver tissue jumps back and the final step is cutting of the needle trough the liver tissue to the desired location.
Using this model and operation specific parameters, it is possible to generate a data set that can be used to determine the tissue deformation at all locations. This will be done using a Kriging-based model. The model uses a limited data set to gain estimates of values over a continuous spatial field. The benefit of using kriging-based modelling is that it can be trained offline and when using it real time it reduces computational expensiveness. When known exactly when the needle will hit the liver tissue the Kriging-based model can show the liver deformation in the generated images of the moving liver to optimise the provided feedback for the surgeons
