Epidemiology

Glial tumors constitute about half of newly diagnosed primary brain tumors, with low-grade gliomas (LGGs) accounting for about 15% of all brain tumors in adults.¹ The subset of tumors classified as LGG is a heterogeneous group of tumors with astrocytic, oligodendroglial, ependymal, or mixed cellular histology. In adults, the term LGG typically refers to the diffuse, infiltrating variety of tumors classified as World Health Organization (WHO) grade II lesions—specifically, low-grade astrocytomas, oligodendrogliomas, or mixed oligoastrocytomas.² Among low-grade astrocytomas, the most common histologic subtypes are the fibrillary, protoplasmic, and gemistocytic variants. There is no indication in the literature that LGGs are more prevalent in a specific ethnic or national group.

About 1500 new cases of LGG are diagnosed in North America each year.³ Age-specific data show that low-grade astrocytomas constitute 15% of brain tumors in adults and 25% of brain tumors in children.¹ Pediatric LGGs, which include cerebellar astrocytomas, optic pathway and hypothalamic gliomas, brainstem gliomas, and hemispheric LGGs, are discussed elsewhere in this book. These tumors demonstrate a slight male predominance and a biphasic age distribution, with the first peak occurring during childhood (age 6 to 12 years) and a second peak in adulthood (between the third and fifth decades). The median age of presentation in adults is 35 years.

Clinical Presentation

LGGs typically arise in the frontal lobes, followed by temporal and parietal lobe lesions, in order of decreasing incidence. Fifty percent to 80% of patients present with seizures as their initial symptom, with most remaining otherwise neurologically intact.⁴ Patients may present with or develop other signs and symptoms, which are largely dictated by the tumor’s size and location. These include signs and symptoms of raised intracranial pressure (headache, nausea, vomiting, lethargy, papilledema), focal neurological deficits (weakness, sensory disturbance or neglect, visual neglect, agnosia, aphasia), and impaired executive function (altered personality, disinhibition, apathy).

In some studies, seizures have accompanied presentation of LGG in up to 81% of patients.⁵ Of the patients who presented with seizures, about 50% have uncontrolled seizures at the time of resection despite antiepileptic treatment. Partial seizure type, temporal location, and longer seizure duration also appear to predispose patients to poorer preoperative seizure control.⁵ Careful consideration of a patient’s seizure status is of paramount importance for LGG patients because seizures significantly affect quality of life. Beyond antiepileptic agents, surgical resection is an effective means of reducing seizure burden on patients with LGGs. Postoperatively, the factors associated with freedom from seizures are gross total tumor resection, preoperative seizure history of less than 1 year, and non–simple partial seizure type. However, continued use of antiepileptic drugs can be necessary, and in some patients, an additional operation may be required for persistent seizure activity. In our experience, optimal control of intractable epilepsy without the use of postoperative anticonvulsants is possible when perioperative (i.e., extraoperative or intraoperative) electrocorticographic mapping of separate seizure foci accompanies the tumor resection. In most cases of epilepsy, as in patients with occasional breakthrough seizures, such mapping is not needed, but complete tumor resection is. When mapping is not used and radical tumor resection (including adjacent brain) is carried out, seizures occur less frequently, but most patients continue to take antiepileptic drugs.⁶

Conventional Neuroimaging Studies

The typical computed tomography (CT) appearance is one of an either discrete or diffuse hypodense to isodense mass lesion, showing minimal or no enhancement with intravenous contrast administration. In about 15% to 30% of patients, however, tumor enhancement can be discerned.^7,8 Calcifications may also occur, particularly among oligodendrogliomas or mixed oligoastrocytomas. In addition, cystic changes may be seen with any histologic subtype.

Magnetic resonance imaging (MRI) is the diagnostic procedure of choice for LGG, delineating the lesion as hypointense to isointense on T1-weighted images and as hyperintense on T2-weighted images (Fig. 121-1). Similar to their appearance on CT scans, most LGGs do not show gadolinium enhancement on MRI. LGGs are intra-axial lesions but do not typically exert significant mass effect on surrounding structures. They do, however, display a tendency to reside within and extend along white matter tracts (e.g., corpus callosum, subcortical white matter). Neuroimaging is not diagnostic but may suggest a particular pathologic subtype of LGG by virtue of the tumor’s location and imaging characteristics. Oligodendrogliomas, for example, are frequently located within the frontal lobes, involve the cortex, and display calcifications, in contrast to other LGGs. Importantly, T1-weighted MRI with gadolinium contrast may underestimate the extent of an LGG. The true extent is shown on the T2-weighted sequences, although on these sequences, tumor extent and surrounding edema are indistinguishable. More recently, diffusion tensor MRI has been used as a surrogate marker of glioma infiltration.^9,10

FIGURE 121-1 T1-weighted, axial (A) and T2-weighted, axial (B) magnetic resonance imaging (with gadolinium contrast) of a non–contrast-enhancing, low-grade astrocytoma in a 41-year-old female patient.

Emerging Neuroimaging Technologies

Magnetic Resonance Imaging

Continued improvement in the resolution of anatomic imaging and innovations in functional and physiologic imaging modalities have the potential to improve our ability to diagnose, treat, follow up, and predict outcome in LGG patients. The increasing use of 3-tesla (3-T) MRI (instead of the standard 1.5-T magnet) will provide more anatomic detail and exquisite cytoarchitectural data on intracranial lesions.^11,12 Proton magnetic resonance spectroscopy (MRS) allows for the noninvasive assessment of metabolite levels within intracranial lesions (Fig. 121-2). Of particular interest are the metabolites N-acetyl aspartate (NAA), choline (Cho), creatine (Cr), and lipids. In contrast to normal brain, gliomas typically demonstrate a decrease in NAA and Cr levels and a rise in Cho levels, indicative of their proliferative potential, cellular heterogeneity, and high cell turnover. In general, higher grade lesions display higher Cho/NAA and Cho/Cr ratios than do lower-grade tumors. The utility and reliability of MRS in predicting tumor grade noninvasively is being evaluated,^13–15 but it does not supplant the need for tissue diagnosis. Nevertheless, MRS may facilitate the identification of targets for surgical biopsy, focusing our attention on regions with elevated Cho peaks, suggestive of increased cellular proliferation, and thereby on regions of maximal tumor activity. In addition, MRS has proved useful in monitoring LGG patients after radiotherapy because it can often distinguish between tumor recurrence and radiation necrosis.

FIGURE 121-2 A 35-year-old male patient with low-grade insular glioma. Three-dimensional magnetic resonance spectroscopy demonstrates multiple voxels with elevated choline, reduced N-acetyl aspartate, and low creatine, compatible with tumor metabolism. Lipid peaks seen within voxels at the more inferior levels may represent either lipid contamination from the skull base or tumor necrosis products.

Magnetic resonance techniques have also been developed for assessment of cerebral blood volume (CBV). A 2- to 3-minute dynamic acquisition of T2-weighted images during intravenous injection of a bolus of gadolinium–diethylenetriamine pentaacetic acid (DTPA) allows estimations of CBV. A voxel-by-voxel CBV map can be created by integrating the area under the dynamic contrast uptake curve and provides a relative measure of CBV with a spatial resolution of about 1 × 2 × 5 mm³ or better. Magnetic resonance perfusion has already demonstrated utility in predicting histopathologic diagnosis and tumor grade noninvasively^16,17 (Fig. 121-3) and will probably play a role in selecting biopsy locations, evaluating treatment response, and differentiating effects of treatment from those produced by recurrent tumor effects.

FIGURE 121-3 Awake mapping results for 1237 cortical sites stimulated in 151 glioma patients. The red squares denote the total number of sites that were stimulated, and the blue squares denote the total number of stimulations that induced speech dysfunction. A lateral view of the dominant-hemisphere cortex, indicating the total number of stimulations per square centimeter of the frontal cortex, is shown in A. The number of stimulations (upper value) and the percentage of total stimulations (lower value) that induced speech arrest (B), anomia (C), and alexia (D) are shown in each square centimeter of the frontal cortex.

(Adapted from Sanai N, Mirzadeh Z, Berger MS. Functional outcome after language mapping for glioma resection. N Engl J Med. 2008;358:18-27.)

Patient Outcome and Survival

The median overall survival time for LGG patients is about 6.5 to 8 years.^20,21 Published survival estimates for patients diagnosed with LGG range from 3 to more than 20 years.^7,20,22–28 Overall, 5- and 10-year survival rates of about 70% and 50%, respectively, have been reported.²⁹ Interestingly, the clinical course of individual LGGs can demonstrate substantial heterogeneity, with certain lesions tending to behave more aggressively, whereas others follow a more indolent course. This diversity of clinical behavior is matched by the anatomic and histopathologic diversity inherent in LGGs. Not surprisingly, this contributes to the controversy among experts regarding the most appropriate strategy for treating patients with LGG. However, Smith and colleagues³⁰ recently demonstrated that after adjusting for the effects of age, Karnofsky Performance Scale (KPS) score, tumor location, and tumor subtype, extent of resection was a significant predictor of overall survival and showed a trend toward predicting progression-free survival. This volumetric extent of resection analysis revealed that patients with greater than or equal to 90% resection had an 8-year overall survival rate of 91% and a progression-free survival rate of 43%, whereas patients with less than 90% resection had an 8-year overall survival rate of 60% and a progression-free survival rate of 21%.

Prognostic Factors

In light of this heterogeneity in clinical outcome, it has been particularly important to identify reliable prognostic factors and assign LGG patients to low- and high-risk subgroups, allowing for the implementation of upfront treatment of lesions predicted to behave more aggressively. Additionally, reliable prognostic factors allow for the rational stratification of patients enrolled in clinical trials.

Clinical characteristics associated with improved LGG patients’ survival outcomes include age younger than 40 years at diagnosis, seizures present at diagnosis, the absence of additional neurological deficits at diagnosis, KPS score of no less than 70, and Folstein Mini-Mental Status Examination (MMSE) score greater than 26/30.^{20,26,27,29,31–34} Imaging factors predictive of poor survival include maximum tumor diameter greater than 5 to 6 cm and the presence of contrast enhancement.^20,27 Increasingly, the extent of surgical resection has been found to be a significant predictor of outcome and progression-free survival among LGG patients (see “Surgical Resection” section under “Treatment Options”). Furthermore, histopathologic factors associated with better prognosis include an MIB-1 labeling index of less than 8% and a histologic diagnosis of either low-grade oligodendroglioma or oligoastrocytoma (especially if the tumor is harboring chromosome 1p deletions, an indication of chemosensitivity).^26,27,35

Dedifferentiation or malignant transformation is a well-described phenomenon observed in LGGs. In the literature, 13% to 86% of tumors initially diagnosed as low grade were observed to recur at a higher histologic grade.^{4,24,33,36–40} Similar to its broad range of incidence, the time to malignant differentiation is also variable, ranging from 28 to 60 months.^4,37,40–42 However, the factors resulting in the transformation to a malignant phenotype are unclear, and the effect of treatment on this malignant transformation remains controversial. In one series, 58% of patients who did not initially undergo biopsy and treatment of a tumor suspected to be an LGG after diagnostic imaging studies eventually required surgery at a median interval of 29 months, and 50% of the tumors then showed anaplastic features.⁴¹ Although these patients had a higher incidence of malignant transformation at the time of operation and shorter time to tumor progression relative to patients who were operated on initially, the study concluded that no difference was observed in overall survival. Nevertheless, it is likely that the timing of malignant transformation affects patient outcome, and this phenomenon may be detected by future studies that are more robust. Furthermore, extent of resection studies suggest that the natural history of malignant transformation can be altered by greater extent of resection.³⁰

Genetic Expression Profile

The cause of LGG is unknown, and with the exception of patients with one of the phakomatoses, there is no defined genetic predisposition that leads to the development of these tumors. The only genetic alteration consistently observed in patients with low-grade astrocytomas is mutation of the p53 gene.⁴³ The p53 gene is located on chromosome 17 at 17p13.1, and this site is often deleted in astrocytomas of all grades. The remaining copy of p53 is usually inactivated through a subtle mutation. Because this gene is essential in the regulation of apoptosis and cell cycle progression, loss of normal p53 function promotes the accelerated growth and malignant differentiation of astrocytes.^44,45 Astrocytomas are the only brain tumors displaying significant p53 mutation rates. Fifty percent to 60% of grade 2 and grade 3 astrocytomas exhibit p53 mutations, suggesting that inactivation of this tumor suppressor gene is an early lesion among gene alterations associated with the development of malignant gliomas.⁴⁶ Although some glioblastomas exhibit p53 mutations, a significant subset of them do not and instead have amplification of the epidermal growth factor receptor (EGFR) gene, suggesting that this subset arises from a different genetic pathway. Other common alterations observed in adult low-grade astrocytomas are gain of chromosome 7 and structural abnormalities, including possessing double-minute chromosomes. Losses of chromosomes 10, 13, 15, 20, and 22 and structural rearrangements involving chromosomes 4, 11, 12, 13, 16, 18, and 21 have also been reported in patients.⁴⁷

Although p53 is only rarely mutated in oligodendrogliomas, more than half of these tumors show a characteristic loss of the long arm of chromosome 1 and the short arm of chromosome 19. Because loss of 1p and 19q is not seen in astrocytic tumors (with rare exception), analysis of the presence or absence of p53, 1p, and 19q can be used to distinguish an astrocytic from an oligodendroglial genotype in cases that are difficult to distinguish histologically. Similarly, most mixed oligoastrocytomas appear to segregate genetically into astrocytic or oligodendroglial genotypes, suggesting that such mixed tumors may not be a distinct biologic entity.⁴⁸

Treatment Options