The robot-assisted system for transforaminal PELS mainly consists of three systems: preoperative planning system, navigation system, and foraminoplasty system. Figure 1 shows the processes of development and operation of this system.
Preoperative planning system
The surgeon inputs the patient’s preoperative Computed tomography (CT) and Magnetic resonance imaging (MRI) data into the preoperative planning system, and then this system constructs fusion images using the multimodal image fusion technique. The fusion images including the vertebra, intervertebral disc, nerves, blood vessels, ligamentum flavum, spinal cord, and spinal dura can be displayed on a 2D displayer or 3D using Microsoft HoloLens head-mounted hybrid reality display technology. In this system, the surgeon can clearly observe the anatomical structures of the surgical segment and surrounding tissues preoperatively and intraoperatively. In the preoperative planning system, the working channel planning is reasonably carried out for the navigation system using the mouse or keyboard device to make sure the landing point is close to the target and complications are prevented. In addition, the surgeon can mark the range of foraminoplasty for the foraminoplasty system. Our team has made some progress on multimodal image fusion of CT and MRI which is the key for the success of this system.
Our team has carried out lots of research on robot navigation and proposed novel “eye‑in‑body” integrated surgery robotic system . The prototype system (Fig. 2a) has been built using a robot arm (VS060A3, Denso Co. Ltd., Japan), a motion tracking system with two cameras (BFLY-U3-28S4C, FLIR Integrated Imaging Solutions, Inc, Canada), a stepper motor (42BYGH47-1684B-ZK6, Liko Inc, China), and two position sensors (EE-SX672WR, Omron Co. Ltd, Japan). Eachcamera has a resolution of 1928 × 1448 and lens focal length of 8 mm. Connecting parts are manufactured by 3D printing. The robot controller software is built using Qt, OpenCV, Eigen, and other software packages. The relationship of the tracking system field of view and robot working range is shown in Fig. 2b. This prototype system was mainly designed for the navigation of pedicle screw implantation. This prototype system can perform stereotactic surgery without the intraoperative hand-eye calibration and manual registration, and can achieve an acceptable position and orientation accuracy while tolerating the errors in the hand-eye coordinate transformation error and the robot kinematics model error.
Based on previous studies conducted by our team, we will develop a new robotic navigation system for transforaminal PELS. Firstly, the surgeon fixes the external marker (which can be developed in C-arm fluoroscopy as well as in the visual system) to the surgical segment or adjacent segments after the patient is placed prone on a radiolucent table. Next, the C-arm is used to take anterior-posterior and lateral radiographic images of the lumbar surgical areas. The navigation system can capture the patient's position and the external marker through visual receptors. According to the external marker and preoperative planning, the robot will automatically adjust the robotic platform and robot arm to the appropriate positions. The above technologies will ensure that the robot arm can work at a small robot working range, has enough flexibility to complete the surgical operations, and realize the rapid placement of the robotic platform and robot arm.
After the robot’s placement, the robot will proceed with automatic registration and tracking of the robot arm. We will use the “eye‑in‑body” integrated surgery robot system similarly. The position information obtained by the visual system can be directly mapped to the robotic arm, which provides great convenience for automatic registration and tracking of the robot arm. The robot can perform automatic registration of the surgical vertebra by establishing the spatial relationship between the external marker and the surgical segment using intraoperative fluoroscopy. After that, the robot will realize the navigation by the robot arm according to the working channel planning, and guide the surgeon to complete the establishment of the working channel.
In addition, as transforaminal PELS is performed under local anesthesia, it is possible that the lumbar vertebra moves, caused by the patient’s movement after vertebral registration during the surgery. To solve this problem, the robot is designed to accurately track the external marker using the visual system. The robot has already established a spatial relationship between the external marker and the surgical target, and the robot can automatically track the moving surgical target through the localization of the external marker intraoperatively. In this way, a fast and highly automatic robot-assisted navigation system can be built to improve the accuracy of the landing point and prevent complications.
With the help of the navigation system, the surgeon establishes the working channel. However, if the surgeon inserts the cannula into the narrow foramen, the cannula may compress the exiting nerve root, causing postoperative dysesthesia. To prevent this complication, the surgeon needs to perform foraminoplasty to enlarge the narrow intervertebral foramen, especially in patients with foraminal stenosis. In addition, if the patient is diagnosed with lateral recess stenosis, the procedure of lumbar ventral facetectomy also needs to be performed simultaneously. Therefore, according to the range of bone resection provided by the preoperative planning system, the robot can perform bone resection automatically using a high speed burr at the end of the 6 DOF robot arm.
However, it is difficult to identify the working status of the high speed burr, such as the range and degree of bone resection. To solve this problem, we will install a force sensor, acceleration sensor, space position tracking marker, and other sensor devices at the end of the robot arm. Through multi-mode sensors such as multidimensional force, position, and acceleration, the robot can accurately identify the working state of the high speed burr and provide feedback to the surgeon. Furthermore, through large-scale emulational and cadaveric tests, various signals such as contact force, tool position, and vertebral vibration will be collected during the working process of bone resection. On this basis, an intelligent algorithm will be constructed to establish the quantitative relationship between the sensor signals and the working state (bone contact state, soft tissue contact state, critical state, etc.) of the foraminoplasty system. This system can provide real-time status monitoring and status feedback to the surgeon. The system can realize the status perception and discrimination of surgical instruments in the process of bone resection, ensure the accuracy of foraminoplasty, and improve the safety and effectiveness of the surgical procedure.
Prototype of the system has been constructed and performance tests are being conducted in simulations and cadavers. In simulation experiments, we will perform the transforaminal PELS. First, lumbar CT and MRI scans are performed to establish multimodal fusion images, and preoperative planning is carried out in the preoperative planning system. The prototype automatically completes the rapid placement and registration of the surgical segment. The operator establishes a working channel with the navigation of the 6 DOF robot arm, and then the robot completes the foraminoplasty using the high speed burr at the end of the robot arm. After the above operations, we will perform dissection of the surgical segment to evaluate the outcome of foraminoplasty. Furthermore, we shall select fresh thawed cadavers for cadaveric tests, and set up experimental and control groups. In the experimental group, we will perform the procedures of navigation and foraminoplasty with the assistance of the prototype. However, we will perform the above procedures without the assistance of the prototype in the control group. When completing the operations, the advantages of the prototype will be evaluated by comparing the outcomes of foraminoplasty, operative time, and radiation exposure between the two groups.