Information loss challenges in surgical navigation systems: From information fusion to AI-based approaches

Surgical navigation technology provides minimally invasive surgery (MIS) with the relative pose relationships amongst medical images, surgical instruments, and lesions. On the other hand, traditional operation procedures depend heavily on direct surgical field exposure. Consequently, introducing surgical navigation can enable surgeons to operate more accurately and efficiently. A tracking system is a core enabling technology of a surgical navigation system. In this paper, after reviewing the tracking technologies, we compare and analyze their pros and cons, and find that information loss is a common challenge. The information loss problem is an inherent drawback in mono-modality surgical navigation systems. It is characterized by physical constraints, attenuation, breakdown of signal, and accuracy instability of the tracking algorithms. This review focuses on the information loss problem in tracking technologies for surgical navigation systems. Furthermore, we survey the existing solutions that aim at tackling the information loss problem, especially in the information fusion of surgical tracking technologies, and we also summarize their key improvements and limitations. Particular attention has been given to the modalities, approaches, objectives, and surgical application scenarios, which can improve the accuracy, precision, and stability of surgical navigation systems. Finally, future research trends directed at improving the information loss problem are discussed, i.e., tight integration of sensing technology, augmented reality for visualization in surgical tracking, stable high-speed 5G networks for telesurgery, strong intelligence and affordable service.