Purpose of the study: Anterior cruciate ligament (ACL) navigation systems are based on two underlying principles: «statistical» anatomic position and isometric anatomic (anatomometric) positioning. The purpose of this study was to demonstrate that an anatometric positioning of the transplant can be achieved, in other words, that the transplant can be positioned in the original anatomic air of the ligament insertion while preserving an optimal isometry without notch impingement. This study was also conducted to compare conventional systems with a computer-assisted system.

Material and methods: This study was conducted on thawed fresh-frozen cadaver knee specimens with > 120° flexion. The computer-assisted protocol for ACL surgery was applied to ten knee specimens. The original anatomic insertions of the ACL were dissected then inserted at the appropriate points into the computer display. The tibial and femoral insertion points of two classical aiming devices were recorded. These points were compared with the original anatomic insertion.

Results: For the tibia: classical aiming methods proposed a point of insertion posterior to the anatomic insertion for eight knees and within the frontiers of the anatomic insertion for two, in line with the anterior border of the posterior cruciate ligament. The computer-designated point of insertion for the tibial fixation was always within the anterior third of the ACL insertion, generally medially. For the femur, the transition (or isometric) line ran across the anatomic femoral insertion in all knees. It was observed that in all cases, the surgeon could choose an anatomic insertion with lesser anisometry by situating the insertion in the distal part of this line: for nine knees, the computer-designated femoral point was anatomic and with lesser anisometry. The Acufex aiming device produced better anisometry (my=4 mm) than the Arthrex device (my=6 mm) but with a less favorable anisometry curve.

Discussion: The notion of anatometry is compatible with computer-assisted surgery. This study demonstrated that the computer-designated tibial point of insertion is more anterior and medial than the conventional aiming points. This is a potential choice if the absence of a notch impingement can be visualized: Howel described a manual fluoroscopic method. In our opinion, at the present time, optimal choice of the femoral point to achieve the desired anisometric curve is strictly operator-dependent.

Purpose of the study: We describe a surgery navigation system based on virtual fluoroscopy images established with a 3D optic localizer. The purpose of this work was to check the accuracy of the system for posterior spinal implants in comparison with conventional surgery. Duration of radiation and duration of surgery were compared.

Material and methods: A 3D optic localizer was used to monitor the position of the instruments in the operative field, as well as the fluoroscopy receptor. The surgeon took two views, ap and lateral, with a total exposure of two seconds. The C arm was then removed. After image correction, the ap and lateral views were displayed on the work station screen where the computer superimposed to tools on each image. Twenty osteosynthesis procedures for implantation of pedicular screws via a posterior approach to the thoracolumbar spine were performed with this virtual fluoroscopy technique (20 patients, 68 screws). During the same study period, twenty other procedures were performed with the conventional technique (ap and lateral x-ray with the C-arm after drilling the pedicle, 20 patients, 72 screws). The position of the spinal implants was compared between the two series on the ap and lateral views and postoperative CT. Similarly time of exposure to x-rays and duration of the surgical procedure were recorded.

Results: The rate of strictly intrapedicular implantation was less than 8% (5/68 screws) in the virtual fluoroscopy series versus 15% (11/72 screws) in the conventional series. Time of exposure to radiation was significantly lower in the virtual fluoroscopy series with a 1 to 3 improvement (3.5 s versus 11.5 s on average) over the conventional method. With training, this method is not more time consuming (10 min per screw for the conventional method versus 11.25 min for virtual fluoroscopy).

Discussion and conclusion: Compared with conventional fluoroscopy, the virtual technique enables real time navigation while significantly reducing the dose of radiation, both for the patient and the surgery team. There are two types of advantages of virtual fluoroscopy over CT-based systems: first virtual fluoroscopy is immediately available without specific preoperative imaging and secondly it provides real non-magnified images acquired once during the procedure, after which the C-arm is removed. 3D virtual fluoroscopy is probably the next step but requires further experience.