Learning Objectives:After participating in this CME activity, the neurosurgeon should be better able to:
- Select ideal candidates and assess operability criteria for surgical resection of insular tumors.
- Differentiate the various intraoperative neuromonitoring techniques and explain their usefulness.
- Interpret recent intraoperative modalities.
This article is the third of 4 parts.
Since the first insulectomies reported by Penfield and Faulk, there has been controversy about the insula as a clearly defined functional entity. Surgery in this region may induce postoperative hemiparesis. Moreover, various degrees of speech impairment after dominant insula infarction, and mutism and apraxias after nondominant insula damage, have been reported. However, it should be noted that in most of these cases the posterior pathologic examination revealed damage on the surrounding opercular cortex and corona radiata, probably due to opercular and subcortical involvement of the lesion, or direct surgical manipulation of these neural structures and ischemic injury due to middle cerebral artery (MCA) branch manipulation. The aforementioned pioneer surgical explorations of the insular lobe were reported by Penfield in the 1950s, who performed direct stimulations on the insular cortex in awake patients, followed by various degrees of insular resection in certain epilepsy cases. The surgical technique he used consisted of anterior temporal lobectomy to safely expose the insular surface. Following the same philosophy, Guillaume et al. promoted transopercular approaches (frontal or temporal depending on the electrocorticographic activity) to reach the insula.
Yaşargil's first long series of limbic and paralimbic tumors included 57 insular and parainsular tumors, and 93 more cases were added. The 93% of the benign reported cases (91% among the malignant tumors) showed good postoperative conditions, and acceptable rates of seizure-free patients. The clinical results were substantially improved after 3 to 6 months postoperatively. After the advent of microsurgical techniques, these first encouraging results reported by Yaşargil et al. convinced many neurosurgeons to take a renewed interest in insular tumor surgery. However, there is still controversy about optimal treatment for patients with these lesions and the various options, including observation, stereotactic biopsy, radiosurgery, and direct surgery that have been proposed.
Thus, the management of patients with insular tumors has been dramatically changing during the last 2 decades, mainly due to the technological developments in neuroimaging and surgical armamentarium. The benefits provided by these technical upgrades have generally improved the decision-making process (diagnosis, treatment, and prognosis) of intrinsic brain tumors. However, this evolution must be clearly understood and managed, because this leap in quality may also carry more complex decisions.
The decision to operate should be individualized to each patient, attending to the symptoms, tumor topography, and individual patients' features. Insular tumors generally appear as producing intermittent but recurrent symptoms, or sometimes asymptomatically. This “indolent” clinical presentation, added to their deep location in the floor of the Sylvian fissure, covered by highly eloquent neurovascular structures, have encouraged many surgeons to recommend biopsies for large dominant-hemisphere insular tumors to avoid postoperative deficits. Nevertheless, it has been demonstrated that this form of treatment does not provide any benefit when compared with large resections of these lesions.
Another option is represented by the wait-and-see strategy. It is fair to highlight that most of these low-grade gliomas experience a prolonged and slowly progressive growing course, which is not as aggressive as that of tumors in other locations. This fact allows the surgeon, but specially the patient, to consider a future decision. These patients can be followed-up closely for a period, and the decision to operate can be taken calmly in the event of imaging progression or bad response to the medical treatment for epilepsy. However, the generally unpredictable natural history of intrinsic insular tumors must also be taken into consideration when deciding on optimal treatment, and surgery seems to represent the safest therapeutic option. Some exceptions to this strategy are the presence of big and/or hemorrhagic tumors and lesions causing an important mass effect and/or presenting as contrast-enhancing tumors. Generally, these features correspond well with high-grade tumors, and patients are more likely to undergo surgery quickly, before their condition can worsen.
With regard to seizures, low-grade insular gliomas are associated with 60% rates of refractory epilepsy. In these cases, microsurgical excision has been promoted as the only solution, achieving results in recent series of 73% seizure-free, 12% occasional, 11% reduced, and 3.6% unchanged seizures. Thus, when pharmacologically refractory seizures are present, the decision to operate seems clear. Attending to other features, Simon et al. demonstrated the importance of patient selection as a key factor in determining functional outcomes. In this sense, young patients who are free of neurologic deficit and World Health Organization (WHO) grade I-III gliomas represent ideal candidates for surgical resection, achieving 3-month postoperative Karnofsky Performance Scale (KPS) scores of 80 to 100. This good prognostic group excludes patients with glioblastoma, which, despite its general poor survival outcome, would benefit from decompressive surgery. Poor overall survival and high complication rates have been reported in patients older than 60 years with preoperative neurologic impairment and a KPS score less than 80. Recent groups have analyzed the impact of extent of resection and other important factors on overall survival and progression-free survival, showing better results in patients with insular low-grade gliomas compared with gliomas with the same histologic features in various locations.
Thus, it seems that young patients in a good clinical condition and low-grade neuroimaging features, and patients in good general condition with important mass effect and intracranial hypertension symptoms, will benefit from careful and gross total removal, which provides strong benefits in survival and symptom alleviation. Consequently, microsurgical removal seems to be the most adequate option when a symptomatic insular glioma is diagnosed. To achieve this goal, the surgeon's experience, anatomic knowledge, and training, aided by recently improved pre- and intraoperative tools, represent key points to complete a successful surgery.
Various intraoperative mapping techniques have improved results of resection of brain tumors near eloquent areas. Some of these techniques require the patient's collaboration to detect, for example a speech arrest during Broca's area stimulation, or visual impairment when stimulating the optic radiations and/or the primary visual area.
Awake craniotomy for brain tumors surgery is a useful technique in selected cases. When the tumor involves the frontoparietal and/or temporal opercula near the language areas, the surgical procedure may be performed under conscious sedation, in which the patient actively participates (object naming, counting, and reading numbers) during much of the procedure. Therefore, the patient must exactly understand the technique, objectives, limitations, and complications of the method.
In these cases, the aim consists of achieving conscious sedation, which means that the patient is snugly sedated but responds to verbal commands. In these cases, the anesthesiologist will take care to provide comfort to the patient by providing sufficient sedation and analgesia, mainly with drugs such as propofol and remifentanil, to achieve good cooperation. The most common sequence is asleep-awake-asleep. During the first stage, the patient will be completely anesthetized to proceed with the scalp block and craniotomy. Before the mapping step, administration of anesthetic drugs will be stopped, and the patient will wake up to collaborate. This will be objectified through an anesthetic depth monitor (bispectral index, which values should be between 70 and 80). Once the cortical mapping is completed, the patient will be anesthetized again until the end of surgery, or at any point of surgery if the surgeon needs to test any other subcortical region mainly related to language or memory.
If the patient's cooperation is not needed, general anesthesia is used, preferably with appropriate depth level (bispectral index 40–50), by use of propofol with remifentanil both in continuous IV infusion. The use of drugs that reduce or alter the potential, such as neuromuscular relaxants, volatile anesthetics, and benzodiazepines, should be avoided. However, neuromuscular relaxants can be used to obtain a partial blockage, in such a way as to reduce movement and facilitate the surgical procedure. Care must be taken to obtain adequate monitoring (with an appropriate lock), except during mapping. One of the possible complications of this method is the appearance of postdischarges associated with stimulation, such as to trigger a seizure. If a postdischarge appears over 30 seconds, it suggests the use of midazolam 2 mg, lorazepam 2 mg, or low-IV dose of barbiturates. It is also recommended to use Ringer's solution at a lower temperature than ambient directly over the exposed cortex. (See Supplemental Digital Content 1, published online, http://links.lww.com/CNS/A10.)
In clinical practice, neuromonitoring techniques are generally used to check nervous system integrity before, during, and once the surgery is over. Measuring the nervous system preoperative status is mandatory to have objective control to easily and quickly detect any change during surgery.
Cortical Mapping: Awake and Nonrelaxed
This neurophysiological technique allows the surgeon to localize functional areas with awake and nonrelaxed patients through direct brain stimulation.
The technique of mapping by direct cortical electrical stimulation allows creation of a functional map of the exposed cerebral cortex, both to identify areas that are functionally significant and areas that are not. The neurosurgeon places the electrode at a small region of cortical area of the brain, and the stimulator from the computer applies a train of stimuli, which can result in neurologic changes as patient movement or numbness or inhibit neurologic function as speech arrest. When stimulating a focal region of the brain produces any of the aforementioned symptoms not accompanied by a crisis or postdischarge, it is confirmed that the stimulated cortex region owns functions. The criterion for considering an area as eloquent involves that a functional response occurs for at least 3 separate stimuli into a single cortical region (Figure 1).
The most common technique for direct cortical stimulation is repetitive bipolar cortical stimulation or Penfield technique. Bipolar stimulation technique of Penfield is based on the activation of the cortical circuit applying electrical pulses repeatedly. A bipolar electrode with 5 mm carbon tips, with a cable connected to an external stimulator that acts as a generator of continuous electrical power trains of biphasic pulse, is commonly used for the stimulation. To assess motor or sensory function, applying a train of 2 to 3 seconds long is enough, whereas assessing language function requires longer durations of 5 to 7 seconds. Stimuli should not be encouraged 2 consecutive times at the same area to prevent postdischarges and seizures. Occasionally, it is necessary to extend the duration of the train to 7 to 10 seconds to map language, due to the complexity of questions and answers to the questions.
It is recommended to start at 1–2 mA amplitude increasing to the minimum intensity that produces a sensorimotor response (not higher than 8–10 mA). The procedure finishes by marking with sterile labels the cortex, with positive response using the legend chosen by each surgeon, to avoid damaging these eloquent areas.
Motor Areas. These tests are mainly performed to identify the precentral, subcentral, and ventral premotor areas above the Sylvian fissure. Registration can be done with the patient awake, determining the evoked movement by direct visualization or with the patient asleep, using continuous electromyographic recording as a control.
The motor mapping can be done through the Penfield method using a bipolar probe, biphasic pulses, 1-ms duration stimuli, 50 to 60 Hz applied for 3 to 4 seconds. It is recommended to start at 1–2 mA amplitude and not higher than 8 mA. The other procedure involves the so-called high-frequency monopolar stimulation technique, in which a monopolar probe delivers a train of 5 pulses, 500-ms duration each, at a 4-ms interstimulus interval (250 Hz).
After the central sulcus and precentral gyrus are identified, we recommend the following steps:
- The stimulation intensity must be increased in 1-mA ranges. Registration can be done with the patient awake, determining the evoked movement by direct visualization and/or with electrodes on target muscles. The stimulus causes muscle jerking or tonic contractions that can begin immediately to stimulation or after several seconds.
- Stimulation intervals of 1 to 2 ms are enough to induce movements when the primary motor cortex is being stimulated. Secondary motor areas require longer stimuli (2–4 ms). In other regions, the stimulation can cause inhibition of movement (when an awake patient is told to move his fingers, stimulation can slow or halt the movement), as supplementary motor cortex stimulation and other regions that can assume an integrative role in motor function.
- Once a movement is induced, we highly recommend exploring all the other needed functions under this motor stimulation threshold. The exposed cortical region will be tested every 5 mm systematically.
- Sterile labels are placed over the cortex with numbers or letters pointing out the areas with positive response, with the aim of easing the fast identification of this region (Figure 1).
Sensitive Areas. The lower postcentral gyrus is mainly checked to look for sensitive responses during insular tumor surgery in awake patients. At this location, the Penfield technique is based on the activation of the cortical/subcortical circuit, using a bipolar handheld probe with 2 tips separated 5 mm (most often used) and connected to an external stimulator that delivers biphasic pulses in short trains, 1 ms in duration, at 50-Hz (Europe) or 60-Hz (North America) frequency. Recommended intensity is 1 to 6 mA, by intensity increments of 0.5 mA and application time of stimulus 2 to 4 seconds.
The presence of sensory impairment is assessed by the patient, as paresthesias in the contralateral regions of the body and occasionally on both sides of tongue or face, or both sides of the neck.
Language Areas. This is performed with patients awake and requires their collaboration. Administration of drugs must be stopped at least 15 minutes before cortical stimulation begin. It is desirable to involve a neuropsychologist for interpretation of various language mistakes caused by the stimulation. There should be a preoperative training about the tasks to be performed intraoperatively. Monitoring is stopped if the patient in a basal situation fails more than 25% of the presented tasks. In a first step, the sensorimotor response should be mapped to confirm a positive response.
After confirming the response, next steps are the following:
- Mapping of cortical areas.
- Language sites, which answers after stimuli are known, are mapped and produce the same inhibition (speech arrest, dysarthria, or anomia). A good beginning may be naming and counting tasks (expressive language): asking the patient to count (1–10, again and again) or name objects presented visually (various tests can be used, selected according to variables such as frequency, familiarity, age of acquisition, and education). It helps to use pictures of items that the patient can name quickly and easily in the test.
- Phonetic association: the patient is asked to say as many names as possible beginning with the letters F, P, and L (1 minute for each letter).
- Semantic association: the patient is asked to read the largest number of cars, fruits, or animals. Nomination of famous faces: the patient is shown 50 pictures of famous and 50 nonfamous people.
- Object naming.
- Simple calculations: multiplication or subtraction.
Failures are coded and responsible areas are registered again with sterile labels, as follows:
- Aphasia (loss of the ability to produce and/or comprehend language).
- Phonemic paraphasias (changes in the articulation of one or more phonemes).
- Semantic paraphasias (changes in the phonology of the word, weakness with the onset of the alterations).
- Anomia (difficulty in retrieving words when speaking).
- Perseverance (repeat previous items as the following items were submitted).
- Speech arrest.
Stimulation techniques for language areas include the following:
- Penfield technique: bipolar probe, biphasic pulses, 1-ms duration stimuli, 50 to 60 Hz. Longer stimulus durations are required to induce stimulation effects on language and cognition compared with motor mapping because of complexity of questions and answers.
- High-frequency monopolar stimulation technique: Monopolar probe delivers a train of 3 to 5 monophasic pulses, 500-ms duration each, 4-ms interstimulus interval (250 Hz), repetition rate of 2 Hz.
Intensity is determined by positive mapping in the motor and/or sensory cortex. Later, language mapping is performed beginning with a current of 1 mA until 3 to 6 mA, by a step of 0.5 mA, and not higher than 8 mA.
Phase Reversal of Somatosensory Evoked Potentials Technique: N20 Wave Technique
This method is performed with patients under general anesthesia. The N20 wave technique has been developed to identify the central sulcus and, consequently, the pre- and postcentral gyri to perform somatosensory evoked potentials (SEPs).
The electrode subdural grid should be placed across the alleged central sulcus, covering the area of the hand or foot at the sensorimotor gyri, with an angle of 15 degrees to the sagittal direction. For SEP of the median nerve, the electrode should be placed in a cortical area between 3 and 8 cm from the midline. An electrical stimulator of constant current in the wrist is used for this purpose. Intensity should be gradually increased until a motor response appears in the first–third fingers (median nerve). Once the wave is obtained, it is easy to locate the central sulcus between the electrodes in which the phase reversal is given . (See Supplemental Digital Content 1, published online, http://links.lww.com/CNS/A10.)
The SEP reverse phase is based on polarity changes in generated dipole field by the cortical afferent pathway. The applied stimuli to peripheral nerve generate an electric dipole on the postcentral gyrus. The dipole polarity changes on adjacent precentral gyrus. Thus, a cortical SEP (N20 for the median nerve) can be recorded from the postcentral gyrus and a reverse image from the turn precentral (P'20/N'25).
This technique provides great help to recognize motor area and to put up a subdural electrode grid appropriately over the precentral region for corticospinal tract monitoring during surgery.
Somatosensory Evoked Potentials
SEPs consist of distal stimulation of afferent pathways, with the aim of identifying short-latency responses evoked from specific cortical areas. When stimulating the sciatic nerve distally, the response is detected on its specific area of the contralateral postcentral gyrus devoted to the representation of the lower extremity, whereas stimulation of the upper extremity will elicit a similar response from a different portion of the contralateral postcentral gyrus. (See Supplemental Digital Content 1, published online, http://links.lww.com/CNS/A10.)
The evoked potential method has been developed as a practical noninvasive clinical method to study conduction in the visual, auditory, and somatosensory systems. In clinical practice, this technique is used to check the somatosensory pathway integrity before, during, and after surgery. Its preoperative status is of great importance to have a control to quickly detect any changes during surgery.
Motor Evoked Potentials
Once craniotomy has been performed and the dura is opened, a silicone strip with a variable electrode number (4 or 8) of platinum (each 4 mm in diameter and with 1-cm interelectrode distance) is used to localize the central sulcus, record cortical SEPs, and elicit motor evoked potentials by direct cortical stimulation during the entire procedure. This technique will allow the surgeon to have continuous control of the corticospinal tract integrity. (See Supplemental Digital Content 1, published online, http://links.lww.com/CNS/A10.)
In patients with insular tumors, these are good methods to detect possible damage of the thalamocortical projections and corticospinal fiber tracts, mainly located in the genu and posterior limb of the internal capsule when dealing with the most posterosuperior aspect of these tumors around the posterior insular point. We combine this technique with direct subcortical stimulation of the corticospinal fiber tract at this stage of the surgery.
Direct Subcortical Stimulation: Classic Bipolar and Monopolar Electrodes and Continuous Ultrasonic Aspirator Stimulation Technique
Subcortical stimulation has become as important as cortical mapping in brain tumor surgery. Its main is tumors with a subcortical extension in the surroundings or inside important white matter fiber bundles. The technical details and the tools used are the same as explained in the cortical mapping section.
Nowadays, insular tumor surgery cannot be understood without direct subcortical stimulation. Once the main tumor volume is removed, the oncologic surgical concept forces the surgeon to resect as much surrounding tumor tissue as possible. The aim is to complete total macroscopic resection respecting important white matter fiber tracts closely related with the deeper tumor components, including the inferior fronto-occipital fascicle (IFOF), the superior longitudinal fascicle (SLF), and the corticospinal fiber tract. When tumor resection approaches these highly eloquent white matter fiber tracts, the price to pay for total resection may be too high. In this sense, the direct subcortical stimulation appears as a great tool (Figure 2). The most accepted approach consists of stimulating each wall of the resection cavity 1 or 2 mm, using a bipolar stimulator during 2 to 4 seconds, with intensities between 2 and 10 mA for the motor and sensitive pathways. In patients with language-related fascicles as the IFOF and SLF, during tumor resection, anesthetic agents are stopped and patients are awakened to carry out the functional pathway mapping, which will follow outlining the eloquent cortical sites at depth. The patient should be nominating and counting in different phases until the end of tumor resection, especially when performing this deep resection.
Recently, we have incorporated the suction pipe stimulator as well as the novel ultrasonic surgical aspirator stimulation device, both of which allow subpial and ultrasonic suction, respectively, while continuously stimulating every single thin layer of tumor. (See Supplemental Digital Content 1, published online, http://links.lww.com/CNS/A10.) This method allows extension of resection in a safer, faster, and more precise way as stimulation is performed during suction, letting the surgeon discriminate as to how far the alleged functional fascicle remains from the resection edge.
Intraoperative Imaging Modalities
Class II data show that the extent of resection for malignant gliomas improves survival and, in patients with low-grade gliomas, improves survival and time to tumor recurrence. This is the main reason why, until new treatments for glial tumors are developed, the cytoreductive treatment provided by surgery will maintain its value. In this sense, various intraoperative imaging modalities have been incorporated to clinical practice, with the ultimate objective being to optimize resection limits, extending them to the maximum while minimizing the eventual associated morbidity.
Image-assisted surgery is continuously being developed to help neurosurgeons operated on patients with brain tumors more safely and effectively. Neuronavigation allows surgeons to locate intra-axial brain tumors more accurately, choosing the best path to the lesion. All intraoperative navigation systems are fed with imaging studies, either CT or MRI. The choice of appropriate images is important, as, for example, low-grade gliomas are best defined by T2-weighted sequences, whereas high-grade primary tumors are best seen on T1-weighted images with contrast. However, the main limitation of neuronavigation is its reliance on preoperative images. Upon opening the skull and dura mater, movements inevitably occur, making data derived from these images unreliable and then loosing accuracy. Only intraoperative imaging can offer the updated information needed to maintain accurate navigation during the surgical procedure. These images are useful to confirm that tumor resection has been completed, a fact often not verifiable under the surgical microscope vision.
Intraoperative MRI (iMRI) is not often used due to its high cost and special requirements in the operating room (OR). An OR must be previously designed to host an iMRI. The complexity of the OR setup is greater for the use of iMRI, and safety and equipment details increase proportionally to the field magnetic strength. How iMRI may influence quality of life and survival remains to be studied, but it seems that selected patients with low- and high-grade gliomas will clearly benefit from the use of intraoperative imaging techniques such as iMRI, due to their ability to offer direct control of the extent of resection in real time, along with the location of the cortical and subcortical eloquent regions relative to the tumor.
Technical pitfalls associated with iMRI have led to standardization of intraoperative ultrasonography at many centers. This is our preferred intraoperative imaging modality when combined with other intraoperative vision techniques and anatomic knowledge. The utility of this device in patients with insular tumors is focused on the first stages of the procedure to confirm the surgeon's view of the relationship between the tumor with the opercula and the Sylvian fissure compartments. Its viability to check the tumor remnant compared with iMRI is diminished, mainly due to the artifacts produced by bleeding, and ultrasonography loses its usefulness in patients with previously irradiated tumors. However, comparison of the snapshot taken before the resection with the postresection picture can be useful enough for experienced surgeons.
Ideally, all brain tumors are at least partially hyperechogenic. Diffuse calcifications inside lesions produce stronger echo patterns, whereas cysts or areas of necrosis are hypoechogenic. Local invasion of gliomas tends to appear as intermediate echogenicity. Edema can be distinguished from the surrounding healthy parenchyma. The main limitation of ultrasound is image resolution and the correlation of preoperative MRI scans and intraoperative ultrasonography images. This problem has been partially solved with the aim of modern neuronavigation systems, which allow overlapping of real-time ultrasound images with preoperative MRI data.
Indocyanine Green Technique
Intraoperative indocyanine green (ICG) fluorescence videoangiography has been a major practical advance in the practice of cerebrovascular microsurgery. Numerous applications have been defined as an adjunct to surgical procedures as diverse as aneurysm clipping, bypass, arteriovenous malformation resection, and dural arteriovenous fistulae obliteration, among others. IGC fluorescence videoangiography is safe and available in most of neurosurgical centers, and numerous applications in tumor surgery have been reported. In our experience, ICG has been useful to understand venous flow at the opercular and superficial Sylvian vein system and to identify the safest way to perform Sylvian fissure splitting. In some cases, this technology has been of great help to identify MCA branch vasospasm during insular component resection (Figure 3).
Ultrasonic aspiration was introduced in neurosurgical procedures 25 years ago and has become an indispensable tool in the neurosurgical armamentarium for resection of intracranial tumors. The ultrasonic aspirator presents 2 effects over the tissue interface. The first effect provides suction that brings the surrounding tissue to the tip of the aspirator and forces it to vibrate, accelerate, and decelerate with the tip, fragmenting it away from harder tissues as vessels. The second important effect consists of a rapidly oscillating tip that produces localized pressure waves, which cause vapor pockets around cells in tissues with high water contents. The collapse of these pockets causes the tissue cells to rupture. The speed of fragmentation depends on the amplitude setting of the system.
The use of this technology remains of great interest when dealing with intrinsic brain tumors, especially when these tumors are close to eloquent areas. Classical resection of intrinsic brain tumors is performed with an electric coagulating bipolar system to destroy tissue and a conventional suction system to aspirate it. The result can be as effective as ultrasonic aspiration, but there is a great difference. Bipolar coagulation systems increase the local temperature of the surrounding tissue and have been demonstrated to be responsible of damaging neighboring areas to the resected tissue. In contrast, the ultrasonic aspirator is at least as effective removing infiltrated tissue, but it has the advantage of not damaging the healthy surrounding and sometimes functional tissue. The use of ultrasonic aspirators can increase the extent of resection, respecting vessels and healthy surrounding tissue, and improving the functional result.
Fluorescence-Guided Resection With 5-Aminolevulinic Acid
This technique allows the visualization of malignant tissue during surgery for malignant glioma (grades III and IV WHO). Tumor resection guided by fluorescence involves giving the patient a natural precursor, 5-aminolevulinic acid (5-ALA HCl), which is taken up by cells of malignant gliomas and, when summed, becomes a fluorescent substance. Thus, by applying a special light during surgery, malignant cells are stained red, offering the surgeon a clear distinction between those that are healthy and those that are not, increasing the extent of tumor resection and minimizing brain damage. (See Supplemental Digital Content 1, published online, http://links.lww.com/CNS/A10.) This technique still presents many pitfalls, and our recommendation is not to strictly follow the information it provides. However, resection with 5-ALA seems a promising tool, and, in this sense, we must support and study it in real conditions with the aim of improving it.