1.1 Early Years and the Concept of Co-registration

Before Francis Gall and the much-maligned field of phrenology, the brain was thought to be a holistic organ. Gall, underappreciated as a neuroanatomist, pioneered the concept of localizing cerebral activity to varied brain regions. Fritsch and Hitzig identified circumscribed cortical sties in the brain that elicited contralateral limb movement. Ferrier followed this work by using stimulation and ablation to demarcate the primary sensory and motor cortices, further confirming the legitimacy of this new paradigm of cerebral functional identity. Modern stereotaxy then began with the work of two men, Sir Victor Horsley and Robert Clarke, at what was then known then as the National Hospital for Diseases of the Nervous System including Paralysis and Epilepsy at Queen Square in London, UK.

Horsley used electrical stimulation to identify then excise a cortical epileptic focus but struggled with imprecise cerebellar localization efforts. Clarke, a physiologist, devised a head-mounted instrument for accurate placement of a needle-like probe to minimize brain injury and improve precision. Instruments for cranial measurement and for localization on the scalp or cranium had been developed by others, including Bridges and Morgan, Broca, Kocher, Zernov, and Rossolimo, but “Clarke’s instrument,” as Horsley himself referred to it, was a three-dimensional digitizer, defining a coordinate space and capable of reliably directing a relatively atraumatic probe to target addresses within that space (▶ Fig. 1.1). ^{1 , 2 , 3} Substructures, like the dentate nucleus, could then be identified. Uniquely, the instrument used co-registration of that surgical space, combining the cranial features of the inferior orbital rim and the external auditory canals, with an atlas comprised of anatomic slices of the brain related to the same external landmarks.

From this founding principle, the defining concept of stereotactic surgery was born.

The variability of both the larger human forebrain and its relationship to external cranial features precluded the translation of the Horsley-Clarke instrument to clinical use. Spiegel and Wycis devised a frame that merged loci of the foramen of Monro and pineal gland obtained from pneumoencephalography into coordinates in a stereotactic device. They used it for thalamic ablations as a less invasive option to frontal leukotomy. Their work spawned the development of innumerable, ingenious stereotactic frames. Surgeons such as Leksell, Reichert and Mundinger, Talairach, and Narabayashi, among many others, designed devices. ^{4 , 5} The advent of computed tomography and magnetic resonance imaging brought new clinical capabilities and needs. Tumors could now be readily seen. How could they be safely reached to biopsy or remove? The ability to do so had been restricted to a subset of neurosurgeons who could, for example, interpret Leksell’s spiral diagram to deal with radiologic parallax. By simplifying the co-registration process, the field was widely opened.

1.2 The Computational Era

Computational co-registration underlays the design of a new stereotactic frame, the Brown-Roberts-Wells frame. More significant was the extension of stereotaxy to craniotomy and volumetric resection. Sheldon and Jacques adapted a tulipretractor to a stereotactic frame for tumor resection, but it was Kelly who promoted the computer as a surgical tool by which multiple imaging data sets could be compiled into a single database, which could then be co-registered with the surgical field using an operating microscope attached to the guiding arc of an enlarged stereotactic frame. ^{6 , 7} Whereas early stereotaxy relied upon a surgeon’s calculation of frame coordinates for a point selected within an atlas or tomographic image, the early computers of the 1980s could readily calculate the transformation required to move in either direction between large sets of points in a preoperative image and in the operative field.

Computational resources advanced the field dramatically, introducing what was initially called frameless stereotaxy, then neuronavigation, and now image-guidance. The functions of a stereotactic frame were to define an operative coordinate space, enable co-registration with preoperative coordinate spaces, and render that information useful to guide an instrument to its deep-seated target. Now this could be improved upon computer-based systems. It was now possible to eliminating the need even for the frame. The concept of stereotaxy was now projected beyond intracranial space. Non-contact digitizers based upon sonic, mechanical arm, optical, and electromagnetic technologies defined a coordinate space and tracked instruments within that space (▶ Fig. 1.2). ^{8 , 9 , 10 , 11} Rapid computation of coordinate transformations correlating the surgical field with sets of fiducial points, surfaces, or volumes in imaging studies replaced human-scale point calculations relying upon superstructures attached to frame bases. Graphic displays that could indicate instrument location and project them onto appropriate radiologic image slices or superimpose it into heads-up displays in operating microscopes along with other relevant information as augmented reality radically advanced effector bandwidth and stereotaxic utility. Dozens of such systems created in individual laboratories made for an exciting, disruptive era that soon matured, coalescing into the handful of systems in use today. Without need of a frame, stereotaxy’s co-registration principle now extends to much of general neurosurgery and increasingly to other surgical fields.

Stereotactic principles moved from burr holes with guidance to craniotomies with guidance. Now, the dilemma of intraoperative data degradation needed to be addressed. What was co-registered before surgery could change mid surgery. Intraoperative MRI and CT machines that can acquire new, updated radiologic images have proliferated. Such implementation and the optimization of its surgical application is ongoing. ^{12 , 13 , 14} Whether alternative approaches that rely on more easily acquired sparse data, such as that provided by the image of the surgical field through the operating microscope or intraoperative ultrasound to shift and deform preoperative images, proves more cost-effective and less cumbersome remains an open question. ^{15 , 16}