腹腔镜下精准切除肝脏肿瘤的增强现实导航关键技术研究

Laparoscopic resection of liver tumors is one of the main trends of minimally invasive surgery in hepatobiliary surgery. However, laparoscopic surgery has some clinical problems, such as difficulty in directly observing complex anatomy of liver, limited flexibility of operation and long learning curve. This project focuses on naked-eye stereo-augmented reality technology to guide laparoscopic surgery. This technology can not only reduce the learning curve of doctors through intuitive intrahepatic anatomical structure perception, but also improve the prognosis by reducing the surgical trauma of patients. The key technologies of augmented reality navigation for laparoscopic precise hepatectomy mainly include the following aspects: First, the liver segmentation based on depth learning will be explored. Second, laparoscopic high-precision and real-time positioning technology based on multi-source information fusion such as optical tracking, inertial navigation and laparoscopic image sequence will be developed. Finally, labelessness laparoscopic image based registration technology will also be developed and the feasibility of the new navigation method will be validated and evaluated by model experiments and animal experiments. Finally, the automatic segmentation of liver in laparoscopic images, high-precision laparoscopic uninterrupted self-localization, as well as new theoretical and technological breakthroughs of naked-eye three-dimensional augmented reality navigation in hepatectomy can be achieved to improve the level of minimally invasive diagnosis and treatment of laparoscopic surgery.

通过腹腔镜切除肝脏肿瘤是当前肝胆外科微创手术发展的主要趋势之一，但是腹腔镜手术存在难以直接观察肝脏内复杂解剖，操作灵活性受限，学习曲线长等临床问题。本项目利用裸眼立体增强现实技术引导腹腔镜下肝脏肿瘤切除手术，既能通过直观的肝脏内解剖结构感知降低医生的学习曲线，又能通过降低患者手术创伤改善预后。开展的腹腔镜下精准切除肝脏肿瘤的增强现实导航关键技术研究主要包括以下方面：基于深度学习的腹腔镜三维重建图像中的肝脏分割研究；基于光学跟踪、惯性导航和腹腔镜图像序列等多源信息融合的腹腔镜高精度无间断实时定位方法研究；基于腹腔镜图像无标记配准的裸眼立体增强现实手术导航方法研究；以及通过模型实验和动物实验验证与评价该新型导航方法的可行性。最终实现腹腔镜图像中肝脏的自动分割，腹腔镜高精度无间断自定位，以及裸眼三维立体增强现实手术导航在肝脏肿瘤切除手术中的新型理论与技术突破，提高腹腔镜手术微创诊疗水平。

本项目围绕腹腔镜图像中的肝脏难以自动三维重建和分割，腹腔镜操作灵活性受限易跟踪中断，腹腔镜手术难以直接观察肝脏内复杂解剖，以及腹腔镜手术学习曲线长等关键问题，开展腹腔镜下精准切除肝脏肿瘤的增强现实导航关键技术研究，最终实现了立体腹腔镜图像中的高精度肝脏三维重建和分割，腹腔镜的高精度无间断实时定位，以及裸眼三维立体增强现实手术导航的新型理论与技术突破。本项目主要与临床合作单位的肝胆外科、影像科等多科室沟通合作，通过研究内窥镜图像处理，无间断定位，以及增强现实手术导航技术，为临床的诊断与治疗提供新型的增强现实导航方式。同时，本项目产生的相关学术成果在国内外会议和期刊上获得了发表，并获得了多项发明专利授权。本项目根据临床应用需求和相关难点，通过大量的模型、离体动物实验等方式，在贴近临床的场景中实现了本项目所提出方法及系统的可行性、精确性与安全性，为后续开展临床实验与成果转化推广提供了基础。总之，该项目对提高腹腔镜手术的微创诊疗水平具有推动作用，同时对其他疾病的精准诊疗也有借鉴价值。