In the newsletter from June 1998 the treatment of hemangioblastomas was discussed in the article ‘Caution Urged on Stereotactic Radiosurgery’. The authors clearly illustrated that not every VHL patient with a cerebellar hemangioblastoma is a suitable candidate for stereotactic radiosurgery. This observation was supported by a study from Adler, Chang, et al. who demonstrated that only VHL patients who present with small (<3 cm) solid hemangioblastomas without significant mass effects are reasonable candidates for radiosurgery.⁴ Since microsurgical resection remains the treatment of choice for the vast majority of symptomatic cystic hemangioblastomas, we would like to present a new adjunct in neurosurgery, i.e. frameless stereotaxy or neuronavigation.

When a patient is diagnosed with an intracranial or spinal hemangioblastoma and is subsequently scheduled for surgery, it is the task of the neurosurgeon to locate this lesion and remove it. To locate the lesion the surgeon is aided by general knowledge of neuroanatomy and by the MR or CT images that have been made before the surgery. The difficulty of this task depends mainly on the size and location of the lesion. A small and deeply situated lesion will be more difficult to locate than a larger, more superficially located lesion. To assist the surgeon’s orientation, especially in more difficult situations, neuronavigation has been developed.

The first reports of neuronavigation date from the mid-eighties (USA and Japan). It is now used in many neurosurgical centers around the world. The term is actually used for several types of localization instruments, which share the fact that they do not use a head-mounted stereotactic frame. Prior to the development of neuronavigation, these rigid mechanical frames were the only instruments capable of localizing a position shown on the CT or MR images. These frames are particularly useful for performing needle biopsies and inserting drains, electrodes etc. However, they are not applicable in open surgery. Neuronavigation offers a solution in open surgery.

To get help in orientation from the computer system, the surgeon needs to show the system where he is operating at that moment. To do this he uses an instrument that can be localized, or located, by the system. The system will then show the position of that instrument in the MR or CT images. Different localizing techniques have been used in the past to do this. The most popular and widely used technique is based on infrared light. Each instrument to be localized carries 2 or more flashing infrared lightsources (LED’s). These infrared flashes are seen by a camera array consisting of two or more infrared-sensitive cameras. The information of these cameras is combined to generate a stereoscopic image, much like the way we use two eyes to see depth. Since the shape of the instrument is known to the system, the position of the tip of the instrument can be calculated from the positions of the LED’s. Using this instrument, the positions of a number of known points on the patient’s head (fiducials) are entered into the computer. The computer combines these points with their respective counterparts in the CT or MR images. It’s as if the pointer is showing the computer system the position of the patient’s head (this is the so-called registration procedure). From that moment on, each new location of the instrument can be shown in the CT or MR images. Thus, during the operation the neurosurgeon can see the location where he is currently operating, at that moment. Instead of infrared light, other localizing techniques include multi-jointed mechanical arms or the localization of sound bursts or electromagnetic fields. However, these variations are based on the same underlying principles.

Although neuronavigation might seem to be an ideal technique, there are a number of specific drawbacks. These drawbacks are related to the issue of accuracy. A neurosurgeon will only rely on a neuronavigation system for his orientation if the system is fairly accurate. The inaccuracy of neuronavigation is an accumulation of small inaccuracies involved with each part of the technique. These include inaccuracies in the CT or MR images, the localizing technique itself, the registration-procedure and the so-called ‘brainshift’. ‘Brainshift’ denotes the changes in anatomy that occur after the CT or MR images have been made. Since the entire system is based on preoperative CT or MR images, displacement of tissue during the operation will not be accounted for. This effect is exaggerated when a large amount of fluid is drained, like cerebrospinal fluid from the ventricular system or tumor fluid from a cyst.

Spinal surgery presents even more problems. First, the registration-procedure is more difficult to perform since fiducials on the spinal column can only be used after they have been cleared of surrounding tissue. Secondly, the vertebrae can move a little relative to each other, which necessitates the performance of a new registration-procedure for each vertebral level. Thirdly, this registration holds true for the vertebrae themselves, but hardly for the nervous tissue within the spinal canal. In other words, the relationship between the spinal cord and the spinal column is less rigid than the relationship between the brain and the skull. Therefore, the use of neuronavigation has only been reported in spinal operations concerning the vertebrae and not in operations concerning the spinal cord.

With each operative procedure the neurosurgeon will have to decide whether neuronavigation is going to be helpful, considering the level of difficulty of orientation and the effect of neuronavigation in that specific procedure. With respect to hemangioblastomas there are a number of possible situations. To date, the removal of spinal hemangioblastomas can not be improved by neuronavigation due to technical limitations. Brain hemangioblastomas can range from small solid lesions to large, predominantly cystic lesions. The value of neuronavigation will decrease as the lesion gets larger, since it will be easier to find anyway. And its value will decrease as the cystic component gets larger, since intraoperative emptying of the cyst will induce more brainshift.

In conclusion, neuronavigation is considered to be a valuable new adjunct in neurosurgery, known to have its own specific limitations and advantages. The technique has limited value for the large, predominantly cystic, hemangioblastomas that may occur in the cerebellum of VHL patients, but these operations are usually not that difficult. Its value is especially clear when open surgery is intended for deep-seated small intracranial hemangioblastomas. Lesions located in the brainstem still pose the biggest challenge. In these cases stereotactic radiosurgery may play a growing role.

1. Department of Neurosurgery, University Medical Center Utrecht, The Netherlands

2. Department of Internal Medicine, UMC Utrecht

3. Department of Medical Genetics, UMC Utrecht

4. Neurosurgery (1998) 43:1

As printed in the VHL Family Forum 8:1, March 2000. For permission to reprint, please contact VHL Family Alliance, editor@vhl.org. Further information is available from the VHL Family Alliance, info@vhl.org.