Development And Precision Test Of Navigationsystem For Particle Implantation

Computer-assisted surgical navigation technology taking advantage of many kinds of medical images not only can help surgeon doing the preoperative plan and minimizing the wound, but also can bring the surgery high quality and convenience.Surgical navigation system of particle implantation is using surgical navigation technology to treat the cancer. The system can provide surgeons with more intuitive information. According to the information of patient tumor in the computer screen, surgeon can decide the amount of particles and the positions of them to finish the preoperative plan. This system can help doctors implant particle accurately in target region and distribute the equal- dose evenly, which leads to the minimal trauma to the normal tissue.The system has the following technical difficulties: the segmentation of tumor, the 3D visualization of tumor, rendering the equal-dose surface of particles, and calculating volume of irregular polyhedron tumor etc.The main contribution of this paper includes the following aspects:Firstly, we designed and implemented the surgical navigation system of particle implantation. We analyzed the system process and divided system into the following modules: the patient data module, segmentation reconstruction module, surgical planning module, equipment module and navigation module. Secondly, we studied the calculation of particle implantation and realized the equal-dose surface rendering. After studied the effects of several existed algorithms, we used the AAPM dose calculation methods.Thirdly, we implemented the algorithm for automatic calculation of tumor volume. We applied the algorithm of arbitrary polyhedral subdivision into tetrahedron and solved the problem of calculation of irregular polyhedron volume.Fourthly, we introduced and implemented the other key technologies of the navigation system: image segmentation, 3D visualization of medical data, path planning, registration, intraoperative alarm. These technologies were essential to our system.Finally, this paper described the related experiments. Through the experiment of heart, head and spine's CT image data, we concluded that the mean error of equal-dose surface rendering was 1.2±0.5mm, and the error between the theory and experiment was within 5%, which could basically meet the clinical requirements. The error of calculation of tumor volume was less than 10%, which showed the validity of those methods.