Stereotactic Navigation for Rectal Surgery: Comparison of 3-Dimensional C-Arm-Based Registration to Paired-Point Registration.

Abstract

Stereotactic navigation enables surgeons to use the preoperative CT or MRI images as a real-time "navigation map." Although stereotactic navigation has been established in neurosurgery and orthopedic surgery, whether this technology is applicable to GI tract surgery remains challenging because of tissue deformation and organ motion. A critical component of this technology is the registration that links the patient's actual body to the preoperative imaging data. The objective was to assess the applicability of stereotactic navigation in rectal surgery, focusing on the registration method. This study was based on a prospective case series. The study was conducted in a single university hospital. Four patients who underwent laparoscopic rectal surgery were included. Paired-point registration was performed for 2 cases, whereas 3-dimensional C-arm-based registration was performed for the other 2 cases. In addition, 3-dimensional C-arm-based registration was performed twice during the operation. Navigation accuracy was evaluated by measuring target registration error at 8 anatomical landmarks. Target registration error of the 3-dimensional C-arm-based registration group was significantly smaller than that of the paired-point registration group (median, 19.5 mm vs 54.1 mm; p < 0.001). In particular, the error of Z-axis (cranial-to-caudal direction) was significantly smaller in 3-dimensional C-arm-based registration (median, 12.4 mm vs 48.8 mm; p < 0.001). In one case in the 3-dimensional C-arm-based registration group, target registration error of the second registration became significantly smaller than that of the first registration (p = 0.008). This was an observational study with small sample size. Three-dimensional C-arm-based registration could be performed with the patient in a lithotomy position with head down and lateral tilt without being affected by positional changes. Three-dimensional C-arm-based registration resulted in significantly higher navigation accuracy than paired-point registration, and its accuracy could be further improved by intraoperative re-registration.