Key point

Introduction: intraoperative imaging modalities

Soon after Röntgen discovered x-rays in 1895 their use was introduced in medical imaging. Conventional fluoroscopy and angiography were the first imaging methods applied intraoperatively. Ultrasonography and computed tomography (CT) followed later, although their initial image quality was less than satisfactory for neurosurgical procedures. The first attempts in the late 1970s and 1980s applying ultrasonography or CT to determine the extent of a resection were disappointing. At that time, image quality was limited, so neither modality gained wide acceptance for intraoperative application. Since then, MRI has become the primary imaging modality for preoperative diagnosis of brain and spinal diseases, providing critical information about anatomy, metabolism, structure, and function. However, the first magnetic resonance (MR) scanners could not be used in an operating environment because of the closed-bore design and the strong magnetic fields. Then, in the mid-1990s, the concept of intraoperative imaging experienced a distinct renaissance with the development of open MRI systems. For the first time, these systems made it possible to apply MRI for intraoperative imaging.

Methods like intraoperative x-ray fluoroscopy are applied as quick, easy to use, and reliable navigation tools showing where a surgical instrument is located in the patient in relation to bony landmarks, and this approach is still used; for example, in transsphenoidal and spine surgery. Ultrasonography is also a method for direct localization and navigation, because of its real-time imaging capability. Tomographic methods like CT and MRI need additional tools to provide localization and navigation, unless the procedure is performed directly in the scanner. These tools are classic frame-based stereotaxy or frameless stereotaxy, clinically known as navigation. All tomographic modalities allow delineating the placement of an implant or device in three-dimensional (3D) space, as well as evaluating the extent of a resection.

Another technique that has emerged in the last few years as alternative or adjunct to the use of classic imaging possibilities is the application of fluorescence methods; that is, application of a fluorescent dye/fluorophore that visualizes vasculature or tumor extent in the surgical field.

Because the role of resection in surgery for gliomas has become increasingly accepted in the recent years, intraoperative methods to optimize the extent of resection are getting more attention from the neurosurgical community. The combination of preserving function while optimizing the extent of resection seems to be the optimal intraoperative treatment strategy.

Introduction: intraoperative imaging modalities

Soon after Röntgen discovered x-rays in 1895 their use was introduced in medical imaging. Conventional fluoroscopy and angiography were the first imaging methods applied intraoperatively. Ultrasonography and computed tomography (CT) followed later, although their initial image quality was less than satisfactory for neurosurgical procedures. The first attempts in the late 1970s and 1980s applying ultrasonography or CT to determine the extent of a resection were disappointing. At that time, image quality was limited, so neither modality gained wide acceptance for intraoperative application. Since then, MRI has become the primary imaging modality for preoperative diagnosis of brain and spinal diseases, providing critical information about anatomy, metabolism, structure, and function. However, the first magnetic resonance (MR) scanners could not be used in an operating environment because of the closed-bore design and the strong magnetic fields. Then, in the mid-1990s, the concept of intraoperative imaging experienced a distinct renaissance with the development of open MRI systems. For the first time, these systems made it possible to apply MRI for intraoperative imaging.

Methods like intraoperative x-ray fluoroscopy are applied as quick, easy to use, and reliable navigation tools showing where a surgical instrument is located in the patient in relation to bony landmarks, and this approach is still used; for example, in transsphenoidal and spine surgery. Ultrasonography is also a method for direct localization and navigation, because of its real-time imaging capability. Tomographic methods like CT and MRI need additional tools to provide localization and navigation, unless the procedure is performed directly in the scanner. These tools are classic frame-based stereotaxy or frameless stereotaxy, clinically known as navigation. All tomographic modalities allow delineating the placement of an implant or device in three-dimensional (3D) space, as well as evaluating the extent of a resection.

Another technique that has emerged in the last few years as alternative or adjunct to the use of classic imaging possibilities is the application of fluorescence methods; that is, application of a fluorescent dye/fluorophore that visualizes vasculature or tumor extent in the surgical field.

Because the role of resection in surgery for gliomas has become increasingly accepted in the recent years, intraoperative methods to optimize the extent of resection are getting more attention from the neurosurgical community. The combination of preserving function while optimizing the extent of resection seems to be the optimal intraoperative treatment strategy.

Navigation and imaging

Typical real-time intraoperative imaging modalities like fluoroscopy and ultrasonography are tools for immediate localizing. However, CT and MRI as standard means of diagnostics in radiological units generally provide preoperative image data that historically were just displayed in the operating room. The integration of the stereotactic principle allowed using CT and MRI also in the sense of localization, leading to modern image-guided surgery in which navigation technology allows the visualization of the essentials of preoperative imaging in the surgical field.

In standard navigation the physical space of the surgical field is registered to the 3D image space, which is based on anatomic data from CT, MRI, or even ultrasonography. Microscope-based navigation provides an intuitive data visualization directly in the surgical field. Navigation accuracy is influenced by a variety of factors. Major factors degrading navigation accuracy are related to an unwanted movement of the registration coordinate system (positional shift), as well as intraoperative events like altering the intraoperative geometry by tumor resection or brain deformation (brain shift). Positional shift and brain shift can both be compensated for by updating the navigation with intraoperative image data, increasing navigation accuracy as well as the safety for the patient.

The overall application accuracy of navigation systems is additionally influenced by the quality of imaging, by the technical accuracy of the system, and by the quality of patient registration, which defines the process of registering image space and real/surgical space.

Standard anatomic navigation is based on anatomic image information only. Integration of additional data obtained from other imaging submodalities results in multimodal navigation. An initial step in establishing multimodal navigation was the development of functional navigation in which preoperative data from magnetoencephalography (MEG) and functional MRI (fMRI), which both define localizations of cortical eloquent brain areas, such as the motor and speech areas, were coregistered with the standard anatomic data and thus could be visualized in the surgical field. This method of functional navigation allowed more thorough resections of tumors in risk zones with low morbidity. Integration of fiber tracking data derived from diffusion tensor imaging (DTI) delineating the course of major white matter tracts extended this concept to subcortical areas, whereas the coregistration of PET data and information from MR spectroscopy added metabolic information leading to true multimodal navigation.

Placing CT and MRI scanners into an operating room for intraoperative imaging provides immediate intraoperative feedback. The most important aspect is to prevent increased neurologic deficits despite increased resections that might result from the attempt to remove initially overlooked tumor remnants that are detected by intraoperative imaging. Therefore, intraoperative imaging should be accompanied by the integration of navigation.

Intraoperative x-ray fluoroscopy and intraoperative angiography

Radiographs were the first imaging modality introduced in routine operating procedures. In 1980, Rey concluded that, “Intraoperative control in neurosurgery requires mainly fluoroscopy, with the possibility of a single exposure for checking purposes.” Fluoroscopy is still used routinely as a reliable and quick means for intraoperative orientation; for example, in transsphenoidal procedures to identify the trajectory to the sella turcica, in stereotactic surgeries, as well as in spine procedures for level identification and assisting pedicle screw trajectory adjustment. Catheter and electrode placements are easily controlled during surgery; for example, in placement of atrial shunt catheters, road mapping in deep brain stimulation procedures, as well as in other kinds of electrode placements procedures such as spinal stimulation or epilepsy surgery.

Further developments of fluoroscopy imaging led to volumetric imaging by acquiring multiple fluoroscopic images about an isocentric point in space, providing axial plane tomographic images that may be reconstructed into an accurate 3D volume (Iso-C arm technology). The usefulness of the Iso-C lies in its convenience and diversity. It provides quick, seamless, and accurate data acquisition for intraoperative imaging with or without navigation. It can be used during traditional open spinal and cranial base approaches or in conjunction with minimally invasive approaches such as thoracoscopy, vertebroplasty, biopsy, and minimally invasive pedicle screw placement. This technology is readily adaptable to any operating room.

X-ray fluoroscopy is also the basis of intraoperative angiography, which is applied in vascular procedures as well. Technological progress like in the standard C-arm systems has led to the development of biplanar systems, with flat panel detectors, where 1 imaging arm can function as a tool for 3D imaging, which leads, for example, to the DynaCT. High-contrast structures, such as bone, calcified lesions, and metallic materials, can be visualized by DynaCT, as well as by multidetector CT. Although contrast differentiation of soft tissue, such as cerebral cortex, muscle, and hematoma, on DynaCT is inferior to multidetector CT scans, the images are sufficiently used as intraoperative CT-like images. However, DynaCT images acquired from flat panel detector–based systems were found to be superior to those images acquired from image intensifier–based systems.

Recently, modern robotic mounting devices further improved these systems and have found broad application in neuroradiological units but also in hybrid operating room concepts in which angiography units are placed in an operating room environment.

Intraoperative fluorescence techniques and others

There are 2 major applications of fluorescence techniques in neurosurgical procedures. One is the intravenous application of a fluorescent biomarker to visualize vasculature, which is beneficial in aneurysm, angioma, and bypass surgery. Indocyanine green (ICG) video angiography provides high-resolution, real-time images of the arterial, capillary, and venous flow of cerebral vasculature. Although it gives adequate information about the clipped neck, parent/branching artery, and perforator involvement, it has some limitations, like viewing the neck residuals located behind the aneurysm, thick-walled atherosclerotic vessels, and thrombosed aneurysms. In arteriovenous malformation (AVM) surgery, it is useful in detecting the residual nidus in diffuse-type AVM, but cannot be relied on in deep-seated AVMs, and it gives exact information about the anastomosis site in extracranial to intracranial bypass, thus avoiding early bypass graft failure. ICG video angiography is a simple, reliable, and quick method to pick up subtle findings in cerebrovascular procedures. However, in selected cases of aneurysms, endoscopy and intraoperative digital subtraction angiography may be helpful, whereas, in deep-seated AVMs, navigation may be required as an adjunct to confirm intraoperative findings. Application of ICG can also be combined with the use of endoscopes. ICG video angiography is a quick and safe method of intraoperatively mapping the vascular architecture of superficial AVMs, but is less helpful for deep-seated lesions. This modality alone does not improve identification of residual disease or clinical outcomes. Surgeon experience with extensive study of preoperative vascular imaging is paramount to achieving acceptable clinical outcomes. Formal angiography remains the gold standard for evaluation of AVM obliteration.

In contrast, there is fluorescence guidance, which has potential in maximizing the extent of high-grade glioma resection. Different fluorophores (ie, fluorescent biomarkers), including 5-aminolevulinic acid (5-ALA), have been examined with the use of several imaging techniques. 5-ALA is the fluorophore mostly commonly examined and applied in glioma surgery; however, there are potential alternatives, like hypericin, that might have advantages with respect to photodynamic therapy. It is still unclear whether fluorescent molecules have specific affinity for neoplastic cells or whether they just reflect a leaky local vasculature and inflammatory change. These techniques seem to be safe and effective for increasing gross total resection; however, the development of more tumor-specific fluorophores is needed to resolve problems with subjective interpretation of fluorescent signal at tumor margins. Techniques such as quantum dots and polymer-based or iron oxide–based nanoparticles have shown promise as potential future tools.

Confocal imaging techniques, as well as other local imaging techniques, such as optical coherence tomography, that might provide a level of imaging that goes beyond the classic operating microscope to a more histologic level might offer interesting alternatives in the future.

Intraoperative ultrasonography

In 1978, Reid reported on the application of intraoperative ultrasonography during surgery of a cystic cervical astrocytoma. Intraoperative ultrasonography is a real-time imaging modality. This ability to depict real-time anatomic data during a surgical procedure is a valuable surgical adjunct and immediately affects surgical decision making. Intraoperative ultrasonography is a rapid and effective way to localize and characterize diseases of the brain, osseous spine, and adjacent spinal cord, providing accurate localization of anatomy and reducing the risk of injury to the spinal cord or surrounding brain parenchyma.

Despite being used for nearly 40 years in neurosurgical operating theaters, particularly by neurosurgeons, it has yet to gain wide acceptance, possibly because much expertise is needed for interpretation of ultrasonography images. The limitations of intraoperative ultrasonography for determining the extent of resection include difficult image interpretation needing many years of experience, as well as limitations in tissue differentiation, especially after having resected a tumor, with loss of cerebrospinal fluid and other surgeon-induced alterations at the resection border.

Even the combination of intraoperative ultrasonography with navigation systems, which allows an easier ultrasonography image interpretation by the parallel display of corresponding MRI or CT slices for easier orientation and identification of structures, as well as techniques like native 3D ultrasonography, have not solved these issues completely. Establishing grading systems, as well as emerging techniques to apply ultrasonography contrast media, might offer further possibilities for intraoperative ultrasonography to determine the extent of resection more reliably. Also recent technological advancements in intraoperative ultrasonography technology, including higher frequencies and reduced probe dimensions, might extend the application of ultrasonography technology in neurosurgery.

There are several case reports and case series on the application of intraoperative ultrasonography in glioma and pituitary tumor surgery. There are no large prospective studies on the effect of intraoperative ultrasonography on glioma resection. In a larger retrospective study on 192 patients with glioblastoma Saether and colleagues showed that survival improved in this patient group within the same period that intraoperative ultrasonography and navigation was introduced and established in their neurosurgical department.

As for intraoperative MRI there are also attempts to combine intraoperative ultrasonography with functional navigation, including fMRI and DTI. However, only a few groups have managed to establish such a working setup. Few studies comparing different intraoperative imaging modalities have shown that ultrasonography is a helpful adjunct as a real-time imaging modality, but the sensitivity of intraoperative MRI was higher.

Intraoperative ultrasonography, as well as duplex sonography, has a major role in vascular neurosurgery. Here the advantages of intraoperative ultrasonography become clear: it is easily available, no excessive preparation is necessary when it is needed during surgery, and it is a real-time modality.

Intraoperative computed tomography

The first intraoperative application of CT technology was published by Shalit and colleagues in 1979. The entire surgical procedure for a recurrent glioma was performed on the table of a CT scanner. In his overview on radiograph control in the operating theater, Rey states that, “As a conclusion, let us have a glance into the future. Shalit from Tel Aviv has started using CT scan control during tumor surgery, and has assessed the quality of removal in two cases of gliomas.” Lunsford and colleagues and Okudera and colleagues published similar early attempts in the following years and adapted CT imaging technology to the operating room environment. However, at that time no larger case series were published.

The next step in technical development was a fully mobile CT scanner that was used in the operating room, as well as in the intensive care unit by several groups. The main investigations were stereotactic procedures, tumor resections, and also the first spinal applications. This setting, as in the first intraoperative MRI systems that were investigated in the same time period, combined intraoperative imaging with navigation.

With the upcoming low-field and high-field MRI systems for intraoperative use, the focus shifted away from intraoperative CT, which had its strengths in depicting bony structures, although it provided only a restricted soft tissue contrast, as well as only allowing fixed slice positions.

About 10 years later, fixed CT systems on rails were integrated in operating room settings, often in close combination and integration with navigation systems. Modern multislice CT scanning offers a reasonable technology for intraoperative imaging, especially in conjunction with a navigation system for intraoperative image updating. The system can be installed in a preexisting operating environment without the need for special surgical instruments. It increases the safety of the patient, the surgeon, and the anesthesiologist without the need to change existing routine protocols and workflows. Imaging and updating of the navigation system can be performed at any time during surgery. Multidisciplinary use increases use of the system, thus improving the cost-efficiency relationship. Hosodoa and colleagues showed prolonged survival in a series of low-grade gliomas caused by usage of intraoperative CT (iCT), whereas an increased extent of resection correlated with longer survival.

The extent of tumor resection in transsphenoidal surgery was reliably assessed by iCT in a study by Lee and colleagues. Compared with postoperative MR imaging, iCT provided good guidance for residual tumor resection, and it not only showed significant promise for maximizing the extent of tumor resection but also increased the percentage of complete removals without any major complications. Similar benefits were found in a recent study investigating the application of DynaCT in skull base surgery.

Intraoperative CT also has a classic application in stereotactic procedures, like the application of the mobile CT-like O-arm for deep brain stimulation procedures.

In vascular neurosurgery, iCT systems are used to visualize the vascular architecture and the extent of vascular anomaly resection or clipping, to give information about the perfusion, as well as to delineate the extent of hematoma evacuation. iCT images can be obtained and merged with those obtained by other modalities, such as MRI, if desired. Depending on the iCT system, limitations may be caused by lower resolution compared with standard multislice systems.

Spine surgery with its increasing technically complex procedures is ideal for iCT technology because of its ability to visualize bony structures in great detail, as well as the capability to image implants. Comparing intraoperative fluoroscopy in spine surgery for pedicle screw placement iCT provided better results. In a study by Costa and colleagues the overall median accuracy of screw placement using the assistance of a navigation system was 93.5%: when preoperative CT data set images were used, the rate of accuracy was 91.8%, whereas with the intraoperative CT acquisition the accuracy was 95.2%. The intraoperative CT acquisition and subsequent automerging of the surgical anatomy with the 3D model allowed a significant reduction in surgical time. With the use of a system allowing the intraoperative CT scan acquisition, such as an O-arm, the need to revise was almost completely avoided. Furthermore, the spinal navigation system allows a significant reduction in exposure to ionizing radiation for the surgeons and nurses. Bandela and colleagues confirmed this by comparing the use of intraoperative CT-based navigation systems with fluoroscopic-based methods.

Recently, mobile iCT solutions integrating navigation have triggered an increased interest for iCT, especially in complex spine procedures in which navigation and intraoperative imaging are almost indispensable.

Intraoperative MRI

Since the mid-1990s intraoperative MRI has gained increasing interest. Intraoperative imaging can identify an incomplete resection; residual tumor that was initially missed can be removed during the same operation without much guesswork. Updated navigation facilitates localizing these tumor remnants. In contrast with the subjective surgeon’s estimate, intraoperative imaging allows an objective evaluation of the intraoperative situation, thus acting as quality control during surgery.

Many brain tumors, particularly low-grade gliomas, have the look and feel of normal brain. Without iMRI, achieving gross total resection without being unduly aggressive is virtually impossible. However, with iMRI neurosurgeons can evaluate the brain at any time during resection, and thus can both avoid eloquent structures and achieve a more complete resection.

In the mid-1990s, with the development of open configured magnets, intraoperative imaging based on MR technology became feasible. The first designs were based on low-field magnets with magnetic field strengths up to 0.5 T. The use of MR scanners in the operating environment proved to be safe and reliable, as well as applicable to neurosurgical procedures, even if these procedures had to be adapted to a certain extent for the MR environment. Several reports on intraoperative low-field MRI for large numbers of patients were published.

In contrast with the development of an MR scanner dedicated for the operating room as pioneered by Black and colleagues, F. Jolesz, and General Electric Medical Systems at the Brigham and Women’s hospital in Boston, in the neurosurgical departments in Erlangen and Heidelberg together with Siemens Medical a low-field MR scanner (0.2-T MR Magnetom Open) was adapted to surgical use. In addition to intraoperative imaging, an integral part of the Erlangen concept was the simultaneous application of navigation. Standard navigation based on anatomic information only was soon enhanced by preoperative functional data from MEG or fMRI defining eloquent brain areas, resulting in so-called functional neuronavigation. Between March 1996 and July 2001 intraoperative low-field MRI was performed in the Erlangen setting in 330 patients. The simultaneous use of intraoperative MRI and functional navigation allowed preservation of neurologic function despite extended resections.

Soon after implementation of the intraoperative MRI technology it became obvious that the most important indications for intraoperative imaging included gliomas, hormonally inactive pituitary tumors, and pharmacoresistant epilepsy. Intraoperative MRI also enabled compensation for brain shift by an update of the navigation system with intraoperative image data.

However, the diagnostic quality of the intraoperative low-field MR systems could not compete with the image quality of routine neuroradiological diagnostics based on high-field magnets. Advances in scanner design, including those caused by active magnetic shielding, have made it possible to adapt modern high-field scanners to the surgical environment. Two major different high-field concepts were initially realized. As for the intraoperative low-field magnet settings, there are 2 basic principles: either taking a standard diagnostic scanner and adapting it to the operating environment, as was done in Minneapolis with a Philips scanner, or designing a high-field MRI scanner specifically dedicated for the requirements of an operating room, as implemented in Calgary with a ceiling-mounted magnet, which is moved into the appropriate imaging position during surgery. Our approach to realize intraoperative high-field MRI combined with microscope-based navigation resembled the Minneapolis setup, necessitating some kind of intraoperative patient transport for intraoperative imaging. The active magnetic shielding of the high-field magnets resulted in a 5G zone close to the scanner, so that the adaptation of a rotating operating table enabled combining intraoperative high-field MRI with microscope-based navigation. The main operating position in which navigation can be applied is located in the 5G zone, where the use of all standard neurosurgical equipment is possible. Since the implementation in April 2002 more than 2500 patients have been operated on with this setting.

Meanwhile several other variants of how to implement intraoperative MRI have emerged at the extremes of magnetic field strengths. There is the development of a mobile iMRI system with magnetic field strengths between 0.12 T and 0.15 T. In contrast, there are many centers moving from intraoperative 1.5 T to 3 T because the magnetic field strength of 3 T has become standard in routine neuroradiological diagnostics. The latest developments in technology relate to the integration of robotics in 3 T intraoperative imaging.

A major cost-saving approach is the dual use of intraoperative systems, which is often the case for the high-field strength systems, allowing parallel intraoperative usage, as well as diagnostic imaging, which clearly enhances the economic effectiveness. Alternatively, some sites try to find low-cost solutions, which interrupt the surgical workflow, so that these approaches do not provide real intraoperative imaging but a kind of inter-resection imaging.

Intraoperative MRI not only provides anatomic imaging for the evaluation of the extent of resection but also structural, functional, and metabolic submodalities (diffusion imaging, fMRI, MR spectroscopy) leading to fiber tracking and fMRI, which can all also be applied and updated intraoperatively, and which are addressed in the accompanying articles.