Satoshi Ieiri, MD PhD, Munenori Uemura, Kouzou Konishi, MD PhD, Takanori Nakatsuji, MD PhD, Mayumi Higashi, MD PhD, Junko Akiyoshi, MD PhD, Ryota Souzaki, MD PhD, Yoshiaki Kinoshita, MD PhD, Morimasa Tomikawa, MD PhD FACS, Kazuo Tanoue, MD PhD FACS. Department of Pediatric Surgery, Kyushu University
Purpose: In minimally invasive endoscopic surgery, because the surgeon is likely to have less tactile feedback than in the open surgical approach, image assistance can be increasingly helpful for three dimensional anatomical understanding of the surgical target. But in general surgery, the advantages of the 3D image data were reduced by organ shift and tissue deformation caused by motion and pneumoperitneu. Therefore an intraoperative navigation system is strongly recommended. We developed an augmented reality (AR) navigation system based on preoperative CT imaging. The purpose of this study is to evaluate the usefulness, feasibility, and accuracy of this system using laparoscopic splenectomy in children.
Methods: Volume images were reconstructed by 3D viewer application. We used an optical tracking system for registration between volume image and body surface markers. AR visualization was superimposed preoperative three-dimensional CT images onto captured laparoscopic live images. This system was applied for 6 cases of laparoscopic splenectomy in children. Five patients were hereditaly spherocytosis (HS) and one patient was idiopathic thrombocytpnea purpura (ITP). To evaluate registration accuracy, distances from the marker position to the volume data were calculated
Results: Developed our AR navigation procedure was successfully introduced in the clinical setting in all cases. This system was able to navigate and superimpose the virtually created images and real-time images with acceptable speed. Overlay images were followed according to the movement of the scope with about 10 fps by using optical tracking system. Typical overlay image was shown in Figure 1. Splenic artery, splenic vein, and pancreas were fused on to laparoscopic live images. The operator recognized the hidden vascular anatomy of the isolated accessory spleen in the fat tissue (Fig. 2a), the splenic artery and vein (Fig. 2b and c), and the pancreatic tail (Fig. 2d) by overlaying an image onto a laparoscopic live image. Preoperative imaging revealed the isolated accessory spleen in one case in which intra-operative localization using conventional laparoscopic means would undoubtedly have been time consuming (Fig. 2a). Operator could confirm the hidden pancreas hidden under the huge spleen (Fig.2e). Finally the staple line was confirmed by AR guidance to preserve the pancreatic tail without complications (Fig.2f). None of the cases required conversion to open surgery.
The registration accuracy (FRE: mm) of 6 cases was 18.8 ± 3.56, 5.3 ± 0.08, 5.71 ± 1.70, 10.1 ± 0.60, 4.06 ± 1.71, and 7.05 ± 4.71. The deviations were corrected using the registration of surface profile of the spleen. In only 3cases, Target registration error (TRE) was measured. The registration accuracy (TRE: mm) of 3 cases was 7.00, 4.94, and 3.15. The accuracy in superimposition of the images was sufficient and acceptable level to enable the surgeon to detect the precise 3D orientation.
Conclusion: This navigation system provides real-time anatomical information which cannot be otherwise visualized without navigation. The registration accuracy was acceptable level in clinical laparoscopic operation. In the near future, we will attempt to increase accuracy of our present system, and develop a “clinically approved” multimodal matching method for capturing intra-operative organ deformations. 
Session: SS17
Program Number: S102