The precise targeting of tumors, lesions and anatomical structures based on CT/MRI imaging is becoming increasingly common for surgical procedures performed through small openings in the skull. However, the 3- to 30-mm diameter of the openings reduces visibility of the target sites and requires the help of support systems for steady, accurate insertion to avoid misplacement of surgical instruments, which can result in hemorrhage and severe neurological complications.Leo Joskowicz, founder and director of the Computer-Aided Surgery and Medical Image Processing Laboratory at The Hebrew University of Jerusalem, has developed a prototype image-guided system for precise automatic targeting and positioning in minimally invasive keyhole brain surgery. Developed in collaboration with researchers at Technion-Israel Institute of Technology in Haifa and at Hadassah Hebrew University Medical Center in Jerusalem, the system provides steady and rigid mechanical guidance without the bulk and cost of larger robots or the pain associated with mounting stereotactic frames on patients.A prototype miniature robotic system (shown attached to the skull) guides the insertion of surgical instruments during brain surgery.The system consists of the parallel robot MARS (MiniAture Robotic System) and its controller; a custom robot-mounting base, targeting guide and registration jig; a 3-D surface scanner and digital video camera; and several software modules for preoperative planning, intraoperative robot positioning, 3-D surface-scan processing and three-way registration. The robot now is called SpineAssist by the company that commercializes it, Mazor Surgical Technologies Ltd. of Caesarea, Israel.The robot is fitted with a mechanical guide for insertion of a needle, probe or catheter, and the device either is mounted on a head-immobilization clamp or is directly affixed to a patient’s skull via pins. The robot automatically positions itself with respect to the surgical targets and can reposition itself during procedures to enable multiple insertion trajectories. The targets are predefined with a CT/MRI image taken before surgery, after anatomical registration with the jig and after a 3-D surface scan of the patient’s upper facial features and ears. Surface scans are captured in 5 s with an accuracy of 0.3 mm using faceScan II, a video scanner manufactured by Breuckmann GmbH of Meersburg, Germany. The system includes several software modules, such as the one shown on the right for preoperative planning. This module inputs CT/MRI images of the patient and geometric models of the robot, enabling the surgeon to define entry and target points and to visualize the resulting needle trajectories. Image courtesy of Computer Aided Surgery.The system’s work volume is approximately 15 to 20 cm3, which requires the surgeon first to manually position the robot roughly near the final site. Coarse positioning is accomplished by aligning the robot with the simulated target position, which has been superimposed on a video image.The team performed in vitro registration experiments to validate the prototype system and to quantify its accuracy. Although the robot itself has an accuracy of 0.1 mm, when all other components of the system are taken into account — registration, surface scanning and imaging — the systemwide accuracy is 1.7 mm, which is close to the clinical accuracy of 1.0 to 1.5 mm required in many keyhole neurosurgical procedures. The next step is to develop a product out of the prototype — which Mazor Surgical Technologies is considering — and to conduct clinical studies. The team is exploring the use of the miniature robot in orthopedic procedures, specifically for targeting screws during pelvic and trauma surgery.