Improving accuracy and reducing errors in spinal surgery--a new technique for thoracolumbar-level localization using computer-assisted image guidance.

BACKGROUND CONTEXT The accurate identification of the correct vertebral level during surgery remains problematic and still accounts for a significant percentage of litigation. The ideal technique for spinal-level localization would have the following characteristics: easy availability in the operating theater, lowest-possible radiation exposure for the professional team and the patient, simple technique which is easily reproducible at any time during surgery, usable with all forms of spine surgery, permanently recordable, able to be used throughout the spine, able to be easily checked by nonspecialist members of the team, and accurate. PURPOSE We describe a new technique for thoracolumbar-level localization, based on these principles, which uses computer-assisted image guidance. STUDY DESIGN Surgery technique development in clinical practice. METHODS The technique uses standard image intensifier radiology with FluoroNav Spine or FluoroNav MAST software on the StealthStation computer-assisted surgery system. (Medtronic Navigation, Louisville, ed) Adjacent, contiguous, images are taken in the desired plane from the reference area of the lumbosacral junction to the general area of operative interest. These images can then be displayed simultaneously on the computer screen. Use of the probe extension feature allows counting, external to the skin and drapes, from the reference level to the level of interest at any time without additional radiation exposure. Standard navigation can then be undertaken at the operative level. RESULTS This technique has been used in 17 cases, all of which have been undertaken in the mid- or low-thoracic and lumbar regions where the operative level is not visible on the same image intensifier image as the lumbosacral junction. All cases have undergone postoperative radiology to check the surgery level and no cases of incorrect level of surgery have occurred. No accuracy errors have developed during surgery and no complications from the reference arc have occurred. CONCLUSIONS This technique is indicated for level localization in the spine where the operative level cannot be visualized on the same fluoroscopy field of view as the reference level. It has a relative contraindication in the upper thoracic spine, in the very obese, and in the presence of osteoporosis where fluoroscopic imaging is difficult, although we postulate a technique using preoperative computed tomography (CT) to overcome these difficulties. This technique satisfies a number of criteria for the "ideal technique" and has advantages over current methods. A number of caveats are level localization and the use of this technique are presented.