Studies looking at resection in high-grade gliomas have had mixed results. The authors briefly review the literature regarding the value of the extent of resection. They proceed to the preoperative and intraoperative tools available to the neurosurgeon to distinguish eloquent from noneloquent language cortex and fibers, including the emerging roles of functional magnetic resonance imaging diffusion tensor imaging tractography and direct cortical/subcortical stimulation in the surgical management of tumors in eloquent areas. Finally, the authors evaluate the postoperative course of these patients and the effect of language deficits on their quality of life.
- •
Significant retrospective data exist to support the hypothesis that maximal safe resection benefits patients with glioma in terms of survival, accuracy of diagnosis, and response to chemotherapy.
- •
There are multiple modalities for preoperatively localizing eloquent language cortex and fibers, including anatomic landmarks, functional magnetic resonance imaging, and diffusion tensor imaging tractography.
- •
Considerable interpatient variation exists in the location of critical language areas. Thus, intraoperative cortical and subcortical electrostimulation mapping remains the gold standard.
- •
As our knowledge of human language function advances, our view of the brain will likely evolve from the identification of isolated areas of the cortex to a better understanding of integrated functioning circuits.
Introduction
It is thought that the cytoreduction of gliomas is a worthy goal and that adjuvant therapies (ie, radiation, chemotherapy, immunotherapy, and so forth) would be more effective with a smaller cell volume leading to delayed recurrence. However, studies looking at resection in high-grade gliomas have had mixed results. Although there is a consensus that obtaining a histologic diagnosis and relieving compression and mass effect are worthwhile goals, the value of further microsurgical resection still remains controversial. This question becomes even more salient when considering a glioma located in eloquent areas, such as the language cortex. The risk of a postoperative language deficit in these surgeries has been reported to be as high as 26%. Thus, the benefit of a gross total or near gross total resection needs to outweigh these risks. Although there is some inconsistency in the literature regarding the impact of the extent of resection on outcomes, an increasing number of reports of both low-grade and high-grade gliomas suggest that extensive resection is beneficial. As such, it is imperative to use all available strategies to obtain safe and extensive resections to ensure that any benefits of further resection outweigh the risks. The authors briefly review the literature regarding the value of the extent of resection. They proceed to the preoperative and intraoperative tools available to the neurosurgeon to distinguish eloquent from noneloquent language cortex and fibers, including the emerging roles of functional magnetic resonance imaging (fMRI) diffusion tensor imaging (DTI) tractography and direct cortical/subcortical stimulation in the surgical management of tumors in eloquent areas. Finally, the authors evaluate the postoperative course of these patients and the effect of language deficits on their quality of life.
Rationale for extensive resection
High-Grade Gliomas
One of the earliest and most influential retrospective studies looking at the survival benefits of a gross total microsurgical resection for patients with glioblastomas was performed by Lacroix and colleagues more than a decade ago. In this study, the investigators combined the results of 416 patients with newly diagnosed and recurrent glioblastomas and concluded that a 98% resection was associated with significantly improved survival. This finding has led to the all-or-none mentality that has existed over the last decade. This study, however, was designed to test whether complete or near complete resections had a survival advantage over biopsy and was not designed or powered to discover the threshold value whereby debulking had a survival advantage over biopsy. There were insufficient numbers of subtotal resections to perform this analysis.
Since that time, there was an avalanche of case series attempting to quantify the benefit, if any, of subtotal resection. A recent review identified 28 studies between 1990 and 2007 that compared the outcome of patients with subtotal versus gross total resections. Out of these studies, 16 demonstrated evidence that gross total resection was a significant predictor of overall survival or progression-free survival or both. Twelve studies, however, demonstrated no significant benefit based on extent of resection (EOR). The most quantitative study to date used compiled case series and Kaplan-Meier survival curve analyses, which suggested that a cutoff of 78% tumor resection provides a survival advantage. However, all of these studies were nonrandomized and suffer from the same statistical confounder of selection bias. Despite some studies attempting to control for various tumor characteristics and baseline Karnofsky Performance Status (KPS), the fact remains that larger, more invasive, and difficult tumors in older patients with poor preoperative KPS scores are more likely to be subtotally resected, whereas younger patients with smaller tumors are more likely to get gross total resections.
Low-Grade Gliomas
The evidence for extensive resection in low-grade gliomas (LGGs) is more persuasive than that for high-grade gliomas. LGGs differ significantly from their higher-grade counterparts in many important respects. A meta-analysis identified 10 studies investigating the benefit of resection in LGGs. Seven of the 10 studies found EOR to be a statistically significant predictor of survival. The survival benefit from gross total resection was approximately 30 months more than subtotal resection, with the average life expectancy increased from 61.1 to 90.5 months. In regard to LGGs, although the question of when to observe versus intervene is still controversial, there exists consensus that once the tumor begins to show progression, the extent of resection does correlate with survival and all efforts should be made to obtain extensive resection. Besides the potential impact on survival, other compelling reasons exist for surgical resection, including the treatment of mass effect; potentially increasing the efficacy of adjuvant therapy; and, perhaps most importantly, increasing diagnostic accuracy.
Effect of Resection on Adjuvant Therapy
There is some indication in the literature that patients with extensive resection respond better to adjuvant therapy. There have been 2 prospective, randomized, phase 3 studies that have shown the efficacy of chemotherapy in patients with glioblastomas: one using carmustine (BCNU) wafers (Gliadel) and one using temozolamide in glioblastomas. In the case of BCNU wafers, the treatment was only significantly better than control in those patients who had a greater than 90% tumor resection. The increase in life expectancy was modest (14.8 vs 12. 6 months; P = .01) but significant. A similar trend was seen in the trial investigating the effectiveness of concurrent radiation and temozolomide following surgical resection. Although this trial was not designed to examine the extent of resection and postoperative imaging was not mandated, patients were stratified into gross total resection (39%), partial resection (44%), and biopsy (16%). The survival advantage of radiation with concurrent temozolamide was greater in the gross total resection group (+4.1 months) than the partial resection group (+1.8 months) and was nonsignificant in the biopsy-only group (+1.5 months).
Diagnostic Accuracy
Finally, and perhaps most importantly, there is indisputable evidence that resection provides significantly superior diagnostic accuracy over stereotactic biopsy alone. Accurate diagnosis can be evasive when the histologic characteristics are heterogeneous. The grade of a glioma is defined by its most aggressive area, yet the tumor may still contain areas with less malignant features, which, if biopsied, may result in significant sampling and diagnostic error. Stereotactic biopsy series report a diagnostic yield of around 90%; however, a diagnosis made from such a biopsy cannot be confirmed unless the biopsy is followed by an extensive resection. In a series of 64 patients who had undergone stereotactic biopsy followed by a more extensive resection, Sawaya found that the final diagnosis from resection was significantly different, leading to a change in therapy in 34 patients (53%). Further, as our ability to characterize gliomas on a genetic and molecular level increases, having more stored tumor may be vital to perform subsequent analyses and further personalized therapy.
Preventing Symptomatic Mass Effect
In patients who present with symptomatic mass effect, surgical resection is unequivocally indicated, even if the tumor involves eloquent areas. Prior studies have shown that gross total resections are associated with better patient neurologic performance scores than those observed after more limited resections. Further, it is unusual for a large high-grade glioma with contrast enhancement to show a significant reduction in size after either radiation or chemotherapy, which, in some cases, requires the surgeon to perform a second surgery for symptomatic debulking.
Rationale for extensive resection
High-Grade Gliomas
One of the earliest and most influential retrospective studies looking at the survival benefits of a gross total microsurgical resection for patients with glioblastomas was performed by Lacroix and colleagues more than a decade ago. In this study, the investigators combined the results of 416 patients with newly diagnosed and recurrent glioblastomas and concluded that a 98% resection was associated with significantly improved survival. This finding has led to the all-or-none mentality that has existed over the last decade. This study, however, was designed to test whether complete or near complete resections had a survival advantage over biopsy and was not designed or powered to discover the threshold value whereby debulking had a survival advantage over biopsy. There were insufficient numbers of subtotal resections to perform this analysis.
Since that time, there was an avalanche of case series attempting to quantify the benefit, if any, of subtotal resection. A recent review identified 28 studies between 1990 and 2007 that compared the outcome of patients with subtotal versus gross total resections. Out of these studies, 16 demonstrated evidence that gross total resection was a significant predictor of overall survival or progression-free survival or both. Twelve studies, however, demonstrated no significant benefit based on extent of resection (EOR). The most quantitative study to date used compiled case series and Kaplan-Meier survival curve analyses, which suggested that a cutoff of 78% tumor resection provides a survival advantage. However, all of these studies were nonrandomized and suffer from the same statistical confounder of selection bias. Despite some studies attempting to control for various tumor characteristics and baseline Karnofsky Performance Status (KPS), the fact remains that larger, more invasive, and difficult tumors in older patients with poor preoperative KPS scores are more likely to be subtotally resected, whereas younger patients with smaller tumors are more likely to get gross total resections.
Low-Grade Gliomas
The evidence for extensive resection in low-grade gliomas (LGGs) is more persuasive than that for high-grade gliomas. LGGs differ significantly from their higher-grade counterparts in many important respects. A meta-analysis identified 10 studies investigating the benefit of resection in LGGs. Seven of the 10 studies found EOR to be a statistically significant predictor of survival. The survival benefit from gross total resection was approximately 30 months more than subtotal resection, with the average life expectancy increased from 61.1 to 90.5 months. In regard to LGGs, although the question of when to observe versus intervene is still controversial, there exists consensus that once the tumor begins to show progression, the extent of resection does correlate with survival and all efforts should be made to obtain extensive resection. Besides the potential impact on survival, other compelling reasons exist for surgical resection, including the treatment of mass effect; potentially increasing the efficacy of adjuvant therapy; and, perhaps most importantly, increasing diagnostic accuracy.
Effect of Resection on Adjuvant Therapy
There is some indication in the literature that patients with extensive resection respond better to adjuvant therapy. There have been 2 prospective, randomized, phase 3 studies that have shown the efficacy of chemotherapy in patients with glioblastomas: one using carmustine (BCNU) wafers (Gliadel) and one using temozolamide in glioblastomas. In the case of BCNU wafers, the treatment was only significantly better than control in those patients who had a greater than 90% tumor resection. The increase in life expectancy was modest (14.8 vs 12. 6 months; P = .01) but significant. A similar trend was seen in the trial investigating the effectiveness of concurrent radiation and temozolomide following surgical resection. Although this trial was not designed to examine the extent of resection and postoperative imaging was not mandated, patients were stratified into gross total resection (39%), partial resection (44%), and biopsy (16%). The survival advantage of radiation with concurrent temozolamide was greater in the gross total resection group (+4.1 months) than the partial resection group (+1.8 months) and was nonsignificant in the biopsy-only group (+1.5 months).
Diagnostic Accuracy
Finally, and perhaps most importantly, there is indisputable evidence that resection provides significantly superior diagnostic accuracy over stereotactic biopsy alone. Accurate diagnosis can be evasive when the histologic characteristics are heterogeneous. The grade of a glioma is defined by its most aggressive area, yet the tumor may still contain areas with less malignant features, which, if biopsied, may result in significant sampling and diagnostic error. Stereotactic biopsy series report a diagnostic yield of around 90%; however, a diagnosis made from such a biopsy cannot be confirmed unless the biopsy is followed by an extensive resection. In a series of 64 patients who had undergone stereotactic biopsy followed by a more extensive resection, Sawaya found that the final diagnosis from resection was significantly different, leading to a change in therapy in 34 patients (53%). Further, as our ability to characterize gliomas on a genetic and molecular level increases, having more stored tumor may be vital to perform subsequent analyses and further personalized therapy.
Preventing Symptomatic Mass Effect
In patients who present with symptomatic mass effect, surgical resection is unequivocally indicated, even if the tumor involves eloquent areas. Prior studies have shown that gross total resections are associated with better patient neurologic performance scores than those observed after more limited resections. Further, it is unusual for a large high-grade glioma with contrast enhancement to show a significant reduction in size after either radiation or chemotherapy, which, in some cases, requires the surgeon to perform a second surgery for symptomatic debulking.
Preoperative imaging for language mapping
Given the preponderance of evidence suggesting the importance of extensive resection, it behooves us to identify comprehensive strategies to safely resect gliomas without impairing eloquent function and quality of life. This point is particularly true when resecting tumors near or within canonical language areas. When planning surgery near perisylvian cortices in the dominant hemisphere, the localization of language cortices is of paramount importance in preventing postoperative deficits. The gold standard for locating essential language cortices has been electrical stimulation mapping (ESM). However, this technique is not without obstacles. Unlike intraoperative mapping of motor regions, patients must be awake and able to respond. This requirement leads to longer operative times, a higher chance of intraoperative seizures, and the potential for considerable patient distress. Moreover, awake intraoperative mapping is limited to those patients who have sufficient language ability and behavioral control to participate. Understandably, there is considerable interest in additional noninvasive modalities to identify patients in whom intraoperative awake mapping may be of low yield or not needed at all as well as means to make intraoperative mapping safer and more efficient.
Anatomic Considerations
It has long been recognized that the human brain has a stereotypical pattern of gyri and sulci. As early as 1980, Kido and colleagues described the relationship between the posterior end of the superior frontal sulcus and the precentral sulcus. Similarly in 1997, Yousry and colleagues described the omega sign as a method to identify the hand portion of the precentral gyrus.
It was initially hoped that this link between structure and function would provide surgeons with much-needed guidance to distinguish an eloquent from a noneloquent cortex. However, with advances in neuro-functional imaging, we are finding more interpatient neuroanatomical variability even among typical patients. For example, the aforementioned omega sign can either represent a primary motor or premotor cortex. Other groups have also reported variability in the functional organization of the primary sensorimotor cortices. For example, within the precentral gyrus, the stimulation of individual cortical sites has been shown to recruit both sensory and motor phenomenon; in other cases, stimulation has been shown to recruit motor movements in more than 1 motor group.
The language cortices are even more variable. While performing cortical stimulation mapping on patients with gliomas undergoing resection, Quinones and colleagues found more than 4 cm of variability in the localization of speech arrest when using classical anatomic landmarks. This finding may be because the cortical representation of speech is more complex than the motor cortex, with multiple essential and nonessential speech areas throughout the frontal, temporal, and parietal lobes. Fortunately, although the location of essential speech areas is variable among individuals, once it is found it is typically small and discrete.
However, the difficulty still remains in predicting where the essential language areas will be in any individual patient. Although it is difficult to positively predict where vital areas will be, it is easier to define where they will not be. Ojemann and colleagues reported that the posterior inferior frontal region is essential in 79% of patients, whereas the anterior middle temporal gyrus is essential in only 5% of patients. In perhaps the most extensive study of intraoperative mapping, Sanai and colleagues tested 3281 cortical sites in 250 patients. In the 151 patients in whom the frontal lobe was tested, only 92 (60.9%) had essential areas of language that were identified on ESM. Further subdividing the frontal lobe squares revealed that even the most prevalent areas only yielded speech arrest in less than 25% of the stimulations.
The presence of an intracranial neoplasm seems to further compound this variability. Intracranial lesions can affect functional localization in 3 ways. First, developmental and vascular lesions may affect how the overlying cortex develops and which functions it assumes. Two studies have noted a greater preponderance of right-sided language lateralization in patients with cerebrovascular malformations. Further, in patients with left temporal lobe epilepsy, earlier age of onset has been associated with a greater likelihood of right-sided or bilateral language lateralization.
Second, intracranial pathologic conditions can lead to functional reassignment. Developmental lesions, destructive injuries, and malignancies acquired in adulthood can all lead the brain to compensate by reassigning neurologic functions to other areas of the cortex. Lucas and colleagues compared language maps in patients with acquired pathologic conditions (gliomas, subarachnoid hemorrhage, and traumatic brain injury) with age-matched controls and found significant migration of language function to the nondamaged cortex in the pathologic group. The best evidence for this phenomenon is that, in stark contrast to ischemic stroke, LGGs rarely present with acute neurologic deficits. In fact, language mapping of patients with LGGs demonstrate multiple patterns of reorganization and compensation. Robles and colleagues reported on 2 patients in whom maps of eloquent language cortices changed between surgeries spaced by several years, allowing a multistage surgical approach for the resection of LGGs in eloquent cortices.
Third, the effect of intracranial disease on the accuracy of the imaging modality is unclear, possibly leading to disease-related imaging artifacts. It has long been suspected that fMRI cannot be used to map eloquent cortices adjacent to arteriovenous malformations (AVMs) because AVMs may alter the perfusion-dependent response that fMRI relies on or because AVMs cause susceptibility artifacts that can interfere with the detection of the blood oxygen level–dependent fMRI response. To investigate this claim, the authors’ group specifically tested the accuracy and reliability of blood oxygen level–dependent fMRI mapping in patients with vascular malformations and found that fMRI is highly sensitive and specific for determining language localization in patients with vascular malformations, even directly adjacent to these lesions. In the authors’ practice at the University of California, Los Angeles, it was found that relying on anatomic localization alone fails to identify up to 25% of the cases in which preoperative mapping (described later) suggested that awake intraoperative ESM mapping was necessary to achieve extensive resection.
fMRI
Recently, the use of fMRI has increased in prevalence. The use of the technology has expanded from simple language lateralization to specific language localization. fMRI works by detecting localized changes in blood flow and metabolism that is coupled to neuronal activity, such as during word language exercises. In contrast to ESM in which only essential cortices are identified, fMRI detects changes in all cortices (essential or not) that are activated during language tasks, resulting in an overly sensitive but nonspecific language map. A recent meta-analysis by Giussani and colleagues identified 9 reports in the literature of case series in which patients with surgical lesions in the eloquent language cortex underwent preoperative fMRI followed by intraoperative electrocortical stimulation (ECS). Of the 9 studies cited, 5 of them computed a sensitivity and specificity of fMRI in comparison with ECS as a gold standard. The sensitivities ranged from 59% to 100% and the specificities ranged from 0% to 97%. The investigators stated that the varied methods and results used in these studies precluded any definitive conclusions about the utility of fMRI in preoperative planning. At this point, fMRI is not universally reliable and depends largely on the quality of the equipment and expert analysis and interpretation. Besides variability across institutions, the variability within a subject across cortices (frontal vs temporal vs parietal) is also not fully understood.
Despite its potential limitations, fMRI has been demonstrated at several institutions to be of value in identifying patients who require awake intraoperative language mapping and in identifying cortical regions that must be specifically interrogated intraoperatively for eloquence, thereby facilitating intraoperative awake mapping and making it more time efficient.
DTI Tractography
Gliomas often grow along white matter tracts in an infiltrative fashion. The method of DTI is a modification of diffusion-weighted imaging that is sensitive to the preferential diffusion of water along white matter fibers and can detect subtle changes in white matter structure and integrity. Over the past 5 years, the authors have routinely integrated DTI into the preoperative evaluation of patients harboring brain tumors ( Fig. 1 ). This imaging modality can be used in a variety of capacities. For example, DTI can be used to differentiate normal white matter from edematous brain and nonenhancing tumor margins. More commonly, however, DTI has been used to evaluate the effect of intraparenchymal tumors on adjacent white matter tracts, including displacement, infiltration, and possible disruption by the tumor. Likewise, combined with functional imaging data (eg, fMRI), DTI has been used to identify the subcortical connections between essential eloquent cortices. This identification provides the surgeon with invaluable 3-dimensional information about spatial relationships of eloquent structures and their connectivity intraoperatively. It should be noted that DTI provides only anatomic and not functional information. Despite this limitation, the use of this technology can be useful in aiding the resection of tumors in the eloquent brain. For instance, DTI, when combined with 3-dimensional intraoperative guidance, may be used to locate the pyramidal tracts in patients with insular gliomas or the arcuate fasciculus between the Broca area and Wernicke area.
Related posts:
Modern Management of High Grade Glioma, Part II
Immunotherapy for Glioma
Molecular Characteristics and Pathways of Avastin for the Treatment of Glioblastoma Multiforme
Clinical Trials of Small Molecule Inhibitors in High-Grade Glioma
Passive Immunotherapeutic Strategies for the Treatment of Malignant Gliomas
Quality of Life and Outcomes in Glioblastoma Management
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
