Key points

Introduction

Low-grade gliomas (LGG; World Health Organization grade II) are a heterogeneous population of intrinsic brain tumors whose natural history is to evolve to higher grade tumors. Over the last few decades, significant advances in the management of LGGs have been made, particularly in the areas of neuroimaging, treatment paradigms, and the genetic make-up of these tumors. However, important questions remain to be answered. One aspect that has been explored more recently is the management of LGG that present incidentally without any obvious clinical signs when they are investigated with an MRI study for unrelated symptoms, such as headache or dizziness. These incidental gliomas are relatively rare and we are only beginning to understand their clinical and biological features. In this review, we discuss the characteristics of incidental gliomas, and their management, with a particular emphasis on their surgical treatment.

LGG constitute 15% of all adult brain tumors, and they most commonly present with seizures (in 80% of cases). A model for the natural history of gliomas posits 4 phases: (1) the occult stage, in which tumor-initiating cells proliferate but there is no detectable tumor on MRI; (2) the clinically silent stage, in which tumor mass becomes apparent on MRI but the patient does not have any symptoms (incidental glioma); (3) the symptomatic stage, in which the tumor elicits symptoms such as seizures or weakness; and (4) malignant transformation, in which the LGG switches to a more biologically aggressive high-grade glioma.

Brain imaging studies of healthy subjects have estimated the incidence of incidental LGGs to be between 0.05% and 0.2% in the general population, and a study of 4309 gliomas from the French Brain Tumor Study Bank revealed that 3% of LGG patients were asymptomatic at the time of diagnosis. A study on the natural history of incidental gliomas demonstrated that if, not treated, these tumors become symptomatic at a median time of 48 months after diagnosis, typically with onset of seizures, or other neurologic signs such as hemiparesis; the same study showed these incidental gliomas grow at a rate of 3.5 mm per year as seen on radiological imaging. Moreover, incidental LGGs are proliferating at similar rates as symptomatic LGGs, with both tumor sets having a median Ki67 proliferative index of 5.0% in one study. Collectively, these findings suggest that incidental LGGs are an earlier phase in the natural history of LGGs.

Introduction

Low-grade gliomas (LGG; World Health Organization grade II) are a heterogeneous population of intrinsic brain tumors whose natural history is to evolve to higher grade tumors. Over the last few decades, significant advances in the management of LGGs have been made, particularly in the areas of neuroimaging, treatment paradigms, and the genetic make-up of these tumors. However, important questions remain to be answered. One aspect that has been explored more recently is the management of LGG that present incidentally without any obvious clinical signs when they are investigated with an MRI study for unrelated symptoms, such as headache or dizziness. These incidental gliomas are relatively rare and we are only beginning to understand their clinical and biological features. In this review, we discuss the characteristics of incidental gliomas, and their management, with a particular emphasis on their surgical treatment.

LGG constitute 15% of all adult brain tumors, and they most commonly present with seizures (in 80% of cases). A model for the natural history of gliomas posits 4 phases: (1) the occult stage, in which tumor-initiating cells proliferate but there is no detectable tumor on MRI; (2) the clinically silent stage, in which tumor mass becomes apparent on MRI but the patient does not have any symptoms (incidental glioma); (3) the symptomatic stage, in which the tumor elicits symptoms such as seizures or weakness; and (4) malignant transformation, in which the LGG switches to a more biologically aggressive high-grade glioma.

Brain imaging studies of healthy subjects have estimated the incidence of incidental LGGs to be between 0.05% and 0.2% in the general population, and a study of 4309 gliomas from the French Brain Tumor Study Bank revealed that 3% of LGG patients were asymptomatic at the time of diagnosis. A study on the natural history of incidental gliomas demonstrated that if, not treated, these tumors become symptomatic at a median time of 48 months after diagnosis, typically with onset of seizures, or other neurologic signs such as hemiparesis; the same study showed these incidental gliomas grow at a rate of 3.5 mm per year as seen on radiological imaging. Moreover, incidental LGGs are proliferating at similar rates as symptomatic LGGs, with both tumor sets having a median Ki67 proliferative index of 5.0% in one study. Collectively, these findings suggest that incidental LGGs are an earlier phase in the natural history of LGGs.

Glioma genetics

The genetic basis of gliomas has been investigated intensively. The cell of origin of gliomas is currently unclear, but may be a neural stem cell or oligodendrocyte precursors. Large-scale sequencing studies have been enormously helpful in elucidating a distinction in the genetic background of LGG compared with high-grade gliomas. Mutations in isocitrate dehydrogenase 1 (IDH1) are present in 80% of grade II and III gliomas and secondary glioblastomas, of which the IDH1 R132H mutation is most frequent. In contrast, primary high-grade gliomas typically lack IDH1 mutations, and are more likely to have other genetic changes, such as mutations or amplifications in epidermal growth factor receptor.

An important genetic alteration in LGG is that of p53, mutations in which are found in two-thirds of diffuse LGGs that later transform to more aggressive tumors. The p53 gene is normally activated after DNA damage to cells, inducing transcription of genes whose ultimate effects include apoptosis. Mutations in p53 are thought to have effects such as inhibition of apoptosis, stimulation of cell proliferation, and neovascularization, which are hallmarks of cancer.

It has been proposed that mutations in IDH1 are an early genetic event in LGGs that is followed by TERT promoter mutations and/or 1p/19q codeletions. Recent work has led to the classification of gliomas into 5 groups based on their status of TERT mutations, IDH mutations, and 1p/19q codeletion, with the best prognosis for TERT- and IDH-mutant tumors and triple-positive tumors, and the worst prognosis for gliomas that only have a TERT mutation. To determine whether incidental LGGs harbor a similar mutational profile as LGGs and therefore are a part of a common natural history, a recent study analyzed for the presence of IDH1 mutations, TERT promoter mutations and 1p/19q codeletions in 23 incidental LGGs, and found these all occur with high frequency in this group of tumors. This suggests that incidental LGGs and symptomatic LGGs share a common genetic basis.

Diagnostic imaging

The gold standard imaging modality for diagnosing LGGs is currently 1.5 T MRI (although 3-T MRI improves image resolution ), in which LGGs typically are isointense or hypointense on T1-weighted imaging and hyperintense on T2-weighted imaging, and usually do not enhance with contrast (although oligodendrogliomas do so in 25%–50% of cases). Because LGGs are slow growing, vasogenic edema and mass effect are less commonly seen on MRI. More recently, diffusion tensor imaging has emerged as a complementary imaging modality to structural MRI, because it can delineate functional tract deflection by tumors, which can certainly aid in preoperative planning to guide the optimal surgical approach and the resection margins. Intraoperative MRI scanning is being increasingly adopted for continuously assessing the progress of tumor resection during the operation and studies suggest that using MRI intraoperatively may enable increased extent of LGG resection with improved outcomes.

Recent physiologic and metabolic imaging modalities can aid the diagnosis and targeting of LGGs. In particular, proton MR spectroscopy is helpful in this regard by quantifying the distribution of cellular metabolite levels: a dominant choline peak (owing to higher membrane synthesis), low N -acetylaspartate (reflecting a reduced neuronal signature), and no lactate or lipid (owing to a lack of hypoxia or necrosis that is typical of high-grade gliomas), are characteristic of LGGs. This imaging modality can further guide biopsy target selection because the level of cellular proliferation in the tumor is correlated with the choline peak. The normalized creatinine/phosphocreatine levels of these tumors are also associated with both progression-free survival and malignant progression-free survival, thereby providing useful prognostic information. At this point, however, it remains unproven whether MR spectroscopy is adequate for monitoring the progression of presumed LGGs.

Another complementary imaging technique is PET scanning, which allows for quantification of tumor metabolism and therefore can help to guide biopsy location and identify histologic upgrading. LGGs are hypometabolic on PET imaging with 18F-fluorodeoxyglucose, in contrast with high-grade gliomas. Radiolabeled amino acids are increased in two-thirds of LGGs, and an example that is used by some in clinical practice is O-(2–18F-fluoroethyl)- l -tyrosine (FET). In a retrospective study of 174 patients with new cerebral lesions, it was found that PET imaging with 18F-FET helped to differentiate gliomas from other lesions: a maximum tumor-to-brain ratio (TBR _max ) of 18F-FET uptake above a threshold of 2.5 had 65% positive predictive value and 84% negative predictive value for detection of a tumor, supporting a further procedure such as biopsy or resection. The TBR _max seems to increase with malignant progression of LGGs, and using TBR _max in addition to other parameters for PET imaging with 18F-FET (time–activity curve pattern of 18F uptake) was found to have a higher diagnostic accuracy for malignant progression than contrast enhancement on MRI, particularly with serial PET imaging with 18F-FET for dynamic information.

Stereotactic biopsy

Incidental gliomas are generally low grade, but histologic diagnosis is essential to confirm the diagnosis, for which tumor tissue can be obtained either via biopsy or from resection. The indications for biopsy of a suspected LGG are if the lesion is diffuse, such as gliomatosis, or the patient is medically unfit for a major operation of surgical tumor resection. A major disadvantage of histologic diagnosis from tissue obtained via biopsy is diagnostic inaccuracy owing to sampling error, which is particularly problematic in mixed gliomas and gliomas with low proliferative rates. Indeed, 28% of grade III gliomas are undergraded and 11% of grade II gliomas are overgraded from biopsy tissue alone. As suggested, however, the diagnostic accuracy of biopsy can be improved significantly if complementary imaging approaches are used, such as PET imaging with 18F-fluorodeoxyglucose and proton MR spectroscopy, the latter of which can be performed at the time of routine brain MRI scanning and improves the diagnostic yield of biopsy for LGGs to close to 100%. We anticipate these methodologies will become more widely used in this context as the technologies become more easily accessible.

Microsurgical resection

An important consideration for incidental gliomas is that these tumors tend to be significantly smaller at the time of diagnosis than LGGs. Moreover, they are also less likely to be situated near eloquent regions of the brain than their symptomatic counterparts. These points offer a considerable advantage to the surgical management of incidental gliomas, in that they are more amenable to gross total resection. Comparing incidental and symptomatic LGG resections, a study reported gross total resection in 60% of incidental gliomas, much higher than the 31.5% rate for symptomatic LGGs. Even in eloquent brain regions, it seems that incidental gliomas are more amenable to fuller resection: supratotal and total resections were achieved in 27% and 36%, respectively, in 1 report. Accordingly, studies have recently demonstrated improved overall survival in patients with resections for incidental glioma compared with those for LGGs.

It is now well-established that the greater the extent of resection of LGGs, the better the long-term outcome for the patient. Several studies over the last few decades, using either volumetric or nonvolumetric tumor assessment, have evaluated the impact of extent of LGG resection on outcome, with the majority of studies demonstrating that the greater the extent of resection the better the overall survival and progression-free survival. These results seem to apply both to hemispheric LGGs as well as to LGGs in limited to certain regions, for example, insular gliomas. A review of the studies assessing the effect of extent of resection on outcomes has been published recently ; to summarize, 3 studies using volumetric assessment to determine the amount of resection for LGGs (462 patients in total) all showed increased 5-year survival with greater resection on univariate and/or multivariate analysis ( Table 1 ), with one of these studies also showing improved malignant progression-free survival.

Recent analyses demonstrate survival benefits from more complete LGG resections are maintained in the very long term, even up to 10 years after the surgery. When complete resections are achieved, the 10-year survival is close to 100%, but this progressively declines as the extent of resection decreases to 40% ( Fig. 1 ). It has also been demonstrated that leaving radiographically evident residual tumor tissue behind negatively impacts outcome, even if the original tumor is large. As imaging techniques become more advanced, the effects of resecting the last few percentile of LGG on patient outcome will become clearer, and this will be important in guiding neurosurgeons as to how aggressive they should be when resecting these tumors, particularly if they are adjacent to eloquent regions or tracts. To date, a volumetric “threshold” for LGG extent of resection remains unknown. A correlation between extent of resection of LGGs and survival may be related to biased treatment allocation, however. To exclude this, a retrospective analysis of 148 LGG patients showed eloquent location of the tumor was the strongest predictor of extent of resection, and the presence of a neurodeficit and extent of resection were the strongest predictors of overall survival, tumor recurrence, and malignant progression; importantly, after stratification by eloquent tumor location to correct for treatment bias, it was shown that extent of resection was still associated with improved overall survival.

Kaplan-Meier curves of overall patient survival with different levels of resection of low-grade glioma. The graph demonstrates stepwise improvement in survival with increasing extent of resection.

( Data from Smith JS, Chang EF, Lamborn KR, et al. Role of extent of resection in the long-term outcome of low-grade hemispheric gliomas. J Clin Oncol 2008;26(8):1338–45; and Adapted from Sanai N, Chang S, and Berger MS. Low-grade gliomas in adults. J Neurosurg 2011;115(5):956, with permission.)

There is growing interest in the use of intraoperative MRI studies for improving the extent of resection for LGGs. In a retrospective study of 102 patients undergoing LGG resection with the aid of intraoperative MRI, intraoperative MRI showed residual tumor present in 79 of these patients, of which 54 patients had tumor amenable for and underwent resection of this residual tumor volume. Their data suggested therefore that intraoperative MRI improved the extent of resection, in particular for nonenhancing gliomas. Similar results were confirmed in a multicenter, retrospective study of 288 patients for whom intraoperative MRI was used to guide LGG resection: gross total resection was found to be an independent prognostic factor for progression-free survival, and patients with residual tumor left inadvertently the prognosis was similar to those with partially resectable tumors.

It is well-known that glioma cells infiltrate beyond the observed margins from T2-weighted MRI and/or fluid-attenuated inversion recovery imaging, in 1 study this infiltration was up to 26 mm beyond the observed margins on imaging. Therefore, some have proposed a supratotal resection of LGGs in cases where the tumor is away from noneloquent regions, and this can alter the natural history by delaying malignant transformation in a subset of LGG. Nevertheless, the optimal distance for resection beyond observable tumor margin remains debatable. In such cases where supratotal resection is preferred, intraoperative functional mapping is especially important for maximizing resection while minimizing risk of inducing neurologic deficits, particularly for incidental gliomas where operative morbidity must be kept to an absolute minimum. A recent analysis of long-term outcomes from supratotal resection for LGG, with 16 patients all with intraoperative mapping for supratotal resection (including a margin beyond fluid-attenuated inversion recovery-weighted MRI abnormalities) and a minimum of 8 years follow-up duration, showed that one-half of the patients had a recurrence but there were no cases of malignant progression or death even after such a long follow-up. These data suggest the benefits of supratotal resection in cases without involvement of eloquent structures may be maintained in the long term, although more extensive and ideally prospective studies are required.

As discussed, molecular markers are known to have prognostic relevance for gliomas and these tumors can be stratified according to such markers as TERT mutations, IDH mutations, and 1p/19q codeletions. This potentially creates a confounding factor for studies, suggesting that the extent of LGG resection improves outcome. To distinguish between these, a study looked at IDH1, p53, and 1p/19q status in 200 grade II gliomas that underwent resection, and demonstrated that a greater extent of resection was not attributable to tumors having favorable molecular markers and, therefore, that maximizing surgical resection independently predicts a better prognosis.