Surgical navigation system for brachytherapy based on mixed reality using a novel stereo registration method

In this study, we present a novel mixed reality navigation system to facilitate brachytherapy, which is an effective method for curing cancer. The accuracy of needle positioning is a vital problem in brachytherapy that can influence the treatment effect. The purpose of the developed system is to help doctors more quickly and easily position the needles and improve seed accuracy in brachytherapy surgery. Based on mixed reality and a multi-information fusion method, a successful fusion of medical images and a preoperative plan for real patients was achieved, allowing doctors to gain an intuitive understanding of the tumor. Image recognition and pose estimation were used to track the needle punctures in real time and perform registration processes. After global registration using an iterative closest-point algorithm with a pattern tracker, medical images and volume renderings of organs, needles and seeds were aligned with the patient. Based on a phantom experiment, the average needle location error was 0.961 mm, and the angle error was 1.861°. The accuracy of needle insertion was 1.586 mm, and the angle error was 2.429°. This research presents the design and validation of a surgical navigation system for thoracoabdominal brachytherapy based on mixed reality. The proposed system was validated through both phantom and animal experiments. The results indicated that the proposed system achieves clinically acceptable accuracy and can aid doctors in performing surgery based on a visualized plan.