Surgical navigation system for liver surgery based on HoloLens(R) 2

A significant challenge in surgical oncology is the accurate intraoperative identification of liver metastases, with disappearing liver metastases (DLM) being particularly problematic to locate during procedures. This paper presents the development and validation of a surgical navigation system for liver surgery based on HoloLens^® 2 augmented reality technology. The system integrates three key components: real-time tool tracking using the HoloLens^® 2's AHAT camera, soft tissue deformation correction through a combination of rigid and non-rigid registration methods, and dynamic 3D model visualization within a mixed reality environment. The tool tracking framework achieved a mean point measurement accuracy of 3.1±1.6 mm when validated using a multi-level block object, with a standard deviation of 2.54 mm in pivot testing. Surface measurement consistency was evaluated using a tissue-mimicking liver phantom, yielding a Modified Hausdorff Distance of 4.55 mm between the measured surface data and ground truth model. The system successfully implements a server-client architecture for real-time model transfer and visualization, allowing seamless switching between rigid and non-rigid registration results through voice commands. This work is an initial feasibility study using HoloLens^® 2 as a standalone surgical navigation platform, and offers potential advantages in workflow integration and intraoperative guidance for liver surgery.