Description
Cancer treatment with internal radiotherapy, which is how cervix cancer is treated, was traditionally planned using CT images, but those are being replaced with MR images due to the fact that they allow for a more accurate radiotherapy. The problem with MR images is that MR devices are either too expensive or are overused in most hospitals. Because of that the goal of our research was to translate as much information, gained from MR images, to CT images, as possible - the result of the research work is a procedure which transforms the cancerous tumor from the MR image to the CT image using the applicator's coordinate system. We chose this approach because the planning of the initial radiation phase, which is planned using MR devices, could support further radiation phases, planned using CT devices. The error of the developed procedure is significantly smaller than the image voxel size and its time complexity is low enough for real-time use. The procedure can be used to assist radiotherapy planning, it can be integrated into a larger radiotherapy system, but most of all it is intended as a platform for researching dependencies between MR and CT cancer images. An analysis of intensity dependencies was performed and the findings are pointing to the fact that intensity features indeed exist in the surrounding areas of the tumor. This analysis represents the first step towards discovering deeper intensity connections between MR and CT cancer images, thus presenting an incentive for further research in the field. The research work is based on images of cervix cancer, but the knowledge gained can be applied to other forms of cancer as well.
Primary authors
Dr
Peter Rogelj
(University of Primorska)
Mr
Simon Mezgec
(Jožef Stefan Institute)
