Neuronavigation – An Evolution in Fits and Starts

0
268

In every profession, there is one skill that is essential for success. For the cranial surgeon, it is a command of cranial localization and understanding craniometric relationships to internal neuroanatomical structures. Albert Rhoton, MD, FAANS(L), had a mantra for aspiring surgeons and anatomists:

“You need to develop a kind of three dimensional X-ray vision.”

o

His point is valid. Until recent developments in augmented reality, a surgeon could only know what is behind ‘that structure’ with intentional study of normal and pathologic anatomy, coupled with internalization and understanding of pre-op cross sectional imaging. Prior to the era of cross-sectional imaging, pneumoencephalograms and angiograms provided the best hope for interpolating the location of a lesion. Even with a sound understanding of anatomy, neurosurgeons required larger cranial openings to ensure access to a lesion. The advent of high-resolution 3-D CT and MR imaging, coupled with stereotaxy, has transformed neurosurgery. This has enabled the surgeon to peer beneath the surface of the skin, visualizing both the location of the lesion as well as the surrounding critical structures.

The Whole Is Greater Than The Sum of Its Parts

The modern neuronavigational system is an excellent example of how progress is made by “standing on the shoulders of giants.” A neuronavigational computer applies the collected knowledge of:

  • Coordination of based neuroanatomic atlases;
  • The collective experience of more than 60 years of frame-based targeting; and
  • More than 30 years of engineering and development in frameless stereotaxis.

Today’s systems derive from a complex lineage of advances in neuro-imaging, with the classic stories of dead-end developments (ultrasonic and acoustic tracking), simultaneous developments (piezoelectric tracking arms by both the Tokyo and Basel Groups) and concepts that were ahead of their time (electromagnetic tracking predates the development of optical tracking). Today, the use of frameless stereotactic or neuronavigational technology is ubiquitous in neurosurgical oncology and plays an increasingly key role in many vascular, spinal and functional procedures. Though the evidence supporting its use for optimizing extent of resection in neurosurgery is limited, there is little question that neuronavigation is the sine qua non for preoperative planning of cranial access and surgical approach for brain tumors.

Limitations of Current Technology

Accuracy of Reference Image

A central assumption in the application of cranial navigation is that the patient’s anatomy at the time of the operation is closely correlated with the patient’s anatomy when the reference MRI or CT image was obtained. Depending on logistics and hospital radiology workflows, the lag time between the pre-op image acquisition and the intervention creates a window for inaccuracy, as some lesions, especially high-grade gliomas, may be rapidly evolving. A bigger challenge is that as a lesion is accessed, the anatomy of the surgical field deviates farther and farther from the preoperative reference scan, rendering the navigational data increasingly inaccurate as an operation proceeds.

(In)Consistency of Reference Imaging

In addition, a great source of inaccuracy, and often frustration, in the practice of frameless stereotaxy relates to the process of registration based on fiducial points or surface contours. The registration step establishes the relationship of the patient to the navigational reference frame and is one of the main factors that dictates the accuracy and utility of a navigational model during surgery.

Obstructions

Because most navigational systems rely on optical tracking of instruments, a clear line of sight is required for a camera or detector to visualize the instrument as it is being used in the field. In some positions, it is difficult or altogether impossible to integrate the use of navigational probes into the desired working position during a brain tumor operation where a microscope is brought in close to the patient’s head. Additionally, optical instruments are limited to a fixed geometry – and this is why these tools must be validated prior to use. Registering a patient merely tells the computer where the working end of an instrument is relative to the tracking frame.

Electromagnetic (EM)-based navigational systems provide a solution to line of sight issues encountered with optical navigational systems, do not depend on fixed geometry of instruments and obviate the need for cranial immobilization by using small reference sensor(s) affixed directly to the patient (as opposed to the skull clamp). These systems permit direct tracking inside the skull, as opposed to requiring tracking arrays that remain outside the field. One of the advantages of this approach is that it allows tracking of the tip of the instrument, as opposed to the tail as tracked optically. This permits real time position information, especially on malleable or flexible instruments, particularly in patients with complex anatomical considerations. However, these systems have been less widely integrated than optical systems, presumably due to challenges related to maintaining a fixed relationship between the reference sensor and the patient’s head as well as issues related to interference caused by the use of metal retractors and surgical instruments.

Directions for Future Development

Existing methods of registration based on a discrete set of landmarks may ultimately be replaced by optical methods that can use hundreds of thousand points detected on the surface of the patient’s face to provide a rapid method for registration. To date, only one product incorporating single-step multipoint optical scanning of the patient’s face to streamline registration has been FDA-cleared for use. 

As endoscope and exoscope technology gain wider use in neurosurgery, there is an increasing interest in the integration of navigational tracking of scopes to ensure optimal access to deep seated lesions. By integrating the microscopic/endoscopic/exoscopic visualization system with the navigational dataset, it is possible to ensure the surgeon the most effective and safest possible access to deep seated lesions without interrupting the flow of surgery. The irony of this is that one of the first systems we would recognize as a neuronavigation system, developed in the 1980s by David Roberts, MD, FAANS(L), was based entirely on real-time tracking of microscope position. Image injection into the microscope is also an old idea that is new again. Now being promoted as “augmented reality,” some commercially available systems provide real-time 3-D images post-processing and segmentation as overlays in the endoscope image stream.

Navigation as Information Hub

As the field of intraoperative imaging broadens, integrating the burgeoning spectrum of imaging data as an information hub for use during surgery is the promise of neuronavigational systems of tomorrow. To date, investigators have proposed or implemented the integration of DTI, functional MRI, mass spectrometry data, optical spectroscopic as well as ex vivo and in vivo microscopic information into the neuronavigational dataset. As the applications and accuracy of neuronavigational systems of the future grow, they are poised to gain an increasingly central role in helping the neurosurgeon navigate the balancing act of delivering safe and effective care.

The holy grail of neuronavigation is a continuously updated reference image that accounts for brain shift during an operation, coupled with a tracking system free from the limitations of line-of-sight or EM detection. Those who solve this problem will have achieved a major breakthrough and realize the goal set decades ago by Dr. Rhoton for optimal cranial localization. This will have implications in neurosurgery and the broader field of medical imaging, advancing patient care and setting a new industry standard for navigational hardware and software.

[aans_authors]

Print Friendly, PDF & Email
o