Clinical manifestations

Presentation and course

Dysembryoplastic neuroepithelial tumors are classified in the WHO category of glioneuronal and neuronal tumors. They occur almost exclusively in the cerebral cortex. In 10 large series that have been published to date, comprising 346 patients (09; 08; 10; 36; 41; 43; 34; 18; 32; 42), the tumor involved the temporal lobe in about three fourths of patients and the frontal lobe about another 20%. Parietal and occipital tumors made up the remainder. One series described 10 patients with tumors in the region of the septum pellucidum (02). There are also scattered case reports of solitary or multifocal occurrence in the basal ganglia posterior fossa and within the ventricles (25; 05; 12; 26). About 60% of patients are male. The age at surgical diagnosis in these 346 patients ranged from 1 week to 55 years, with symptoms present for 1 month to 39 years prior to diagnosis. This tumor is extremely indolent with stability over years to decades and may be diagnosed at any age; in fact, there are reports of postmortem diagnosis at the ages of 50 and 75 years (13; 25). The onset of symptoms occurs almost exclusively in childhood or early adulthood, usually between the ages of 1 month and 17 years, with mean age of symptom onset of 9 years. Onset of symptoms in adulthood is less common, though in a series of patients who had undergone epilepsy surgery, more than 40% had adult onset of seizures (03).

Almost every patient with dysembryoplastic neuroepithelial tumor presents with epilepsy, which, as mentioned above, may be of substantial duration. The seizures are complex partial in type, though simple partial seizures and secondary generalization of partial seizures may occur. Most patients are medically refractory. Rare presentations have been recorded with symptoms such as headache, vomiting, and increased intracranial pressure or hydrocephalus, particularly when the tumor occurs in a periventricular location and obstructs the foramen of Monro (36; 05; 02). Nonepileptic focal neurologic deficits are much less common. Occasionally there is protrusion of the skull overlying the lesion. One report describes occurrence of this lesion in 2 (and possibly 3) members of a family (16).

Prognosis and complications

The overall prognosis for patients with dysembryoplastic neuroepithelial tumor is excellent. Follow-up studies in patients whose tumor was not resected have shown no growth for periods of 2 to 12 years with CT or MR imaging (08; 05; 10). After resection, recurrent tumor is rare, with most studies reporting no recurrence in any patient (36; 41; 10; 18). Two patients who died from other causes 6 to 15 years after resection had no evidence of recurrent tumor at autopsy (09; 18). A minority of patients do experience recurrence, however (34; 32). Malignant transformation of a dysembryoplastic neuroepithelial tumor following subtotal resection is quite rare (15).

The prognosis for patients' epilepsy is also excellent overall. The literature regarding intermediate-term outcome (more than 6 months of follow-up) of glioneuronal tumors, including dysembryoplastic neuroepithelial tumors and gangliogliomas, demonstrates most (approximately 80%) of patients are free of disabling seizures after surgery, with outcome predicted by less than 1-year duration of epilepsy, absence of generalized tonic-clonic seizures, and gross total (as opposed to subtotal) resection (11). However, this control rate may be artificially high due to the limited follow up in some studies. For example, one study found that although 85% of children were free from disabling seizures at 1 year after resection, this proportion dropped considerably over the ensuing 2 years, with only 62% being seizure-free at last follow-up (32). The only factor found to be associated with long-term seizure freedom was complete tumor resection. For tumors in the temporal lobe, temporal lobectomy may be superior to mere lesionectomy. This might be due to the coexisting cortical dysplasia typically found adjacent to these tumors (06).

Clinical vignette

A 20-year-old man presented to the emergency department after a seizure. The seizure was described as focal in onset with left arm stiffening and left gaze deviation which progressed to clonic movement of the left arm and face and ended in a generalized convulsion. He reported a prior 3-year history of episodic deja vu and visual hallucinations for which he never sought medical attention. MRI of the brain showed a multicystic-appearing mass in the right temporal lobe extending into the amygdala, hippocampus, and thalamus without enhancement.

Stereotactic biopsy revealed a low-grade glioma. He opted for observation with serial MRI scans and management of his seizures with anticonvulsants. Initially he was treated with levetiracetam but over the next 12 months his seizures increased in frequency and required addition of anticonvulsants including levetiracetam, lacosamide, and oxcarbazepine. Repeat MRI shows a largely unchanged right temporal lobe mass. EEG showed occasional right anterior temporal discharges. He underwent a right anterior temporal lobectomy with an uncomplicated postoperative course. Pathologic examination revealed dysembryoplastic neuroepithelial tumor. He was able to be slowly weaned off several anticonvulsant medications over the following year. At the time of last follow-up he was seizure free on levetiracetam monotherapy and his imaging showed no sign of tumor recurrence.

Biological basis

Etiology and pathogenesis

Dysembryoplastic neuroepithelial tumor appears in situ as enlargement of a cortical gyrus, which exceeds the thickness of the surrounding normal cortex. Macroscopically, the tumor is often contained within the cortical ribbon, with involvement of the underlying white matter typically only in mesial temporal locations (24), though a different series suggested much more frequent white matter involvement than had previously been thought (18). The lesion has a mucinous or semi-liquid consistency with macroscopic nodules visible within. The soft consistency of this tumor renders it prone to damage during removal, and great care must be taken in handling the pathologic specimen or diagnosis may be hindered (09). Meningeal involvement occurs occasionally.

Classically, the dysembryoplastic neuroepithelial tumor displays 3 distinct pathologic features. The one that is most specific for this tumor type is the oddly named "specific glioneuronal element," consisting of columns of monotonous, small, round cells with little cytoplasm that resemble oligodendrocytes.

These columns are oriented perpendicular to the cortical surface and typically form an alveolar pattern, with the interstitial spaces filled with pale, mucinous fluid. Thin capillaries run through these columns, and cytologically normal neurons give the appearance of floating within the fluid matrix. Mitotic figures are rare. Ultrastructural studies demonstrate that the columns are made up of bundles of axons to which are attached the cell processes of the presumptive oligodendrocytes (08; 41). These small round cells stain positively for S-100 (an oligodendroglial marker), and the majority of tumors showed staining with myelin oligodendrocyte glycoprotein, a marker for mature oligodendrocytes (14). These cells stain only rarely with GFAP (an astrocytic marker); however, multiple reports also demonstrate that in 7% to 40% of cases, small fractions of these cells stain positively for neurofilament protein, synaptophysin, neuron-specific enolase, or class III beta-tubulin, suggesting early neuronal differentiation. In addition, electron microscopy has demonstrated dense-core granules (a characteristic of neuronal differentiation) in a few of these cells and occasional synapse formation involving them (17; 25; 41; 34). This evidence suggests origin from an early glioneuronal precursor rather than from an oligodendroglial cell.

The second, more variable, histologic characteristic of this tumor is that of glial nodules. These sharply demarcated regions are most often situated at the edge of the specific glioneuronal element, bordering the white matter. They are composed variously of astrocytes (usually pilocytic, but sometimes fibrillary or even anaplastic), small round oligodendrocyte-like cells, and mature neurons, with astrocytic nodules being most common. Occasionally, nodules may be composed of cells of glioneuronal lineage that cannot be further characterized (09).

The third pathologic element of this tumor type is that of cortical dysplasia, which is present in a graded fashion at the edges of the specific glioneuronal element, shading off into normal cortex at either end. Neurons do not demonstrate any cytologic atypia.

Tumors containing all 3 of the aforementioned elements were subsequently labeled the "complex form" of dysembryoplastic neuroepithelial tumor, as further studies by Daumas-Duport unearthed many cases in which only the specific glioneuronal element was present, with none of the other features. These patients shared the clinical profile of the previously reported patients and were, therefore, considered to have the "simple form" of dysembryoplastic neuroepithelial tumor (08). The existence of cases featuring only nodules and dysplasia without the specific element was also documented in 40% of cases in the original report and has been found by others as well (20). In light of these findings, Daumas-Duport considered either the specific glioneuronal element alone or the glial nodules with surrounding cortical dysplasia to be sufficient to make a diagnosis of dysembryoplastic neuroepithelial tumor (08).

The most likely cellular origin of this tumor is the subpial granular layer of the cortex, 1 of the secondary germinal layers of the nervous system. These cells are believed to migrate into the cerebral cortex in the middle to late stages of gestation and then to mature into cortical glial cells (08). Thinning of the inner table of the skull, a frequent radiologic finding in this tumor, and the surrounding cortical dysplasia seen histologically also attest to its development in the earliest stages of life. This hypothesis does not explain the remarkable stability of this lesion; it has been proposed that early control of the abnormal growth by embryonic tissues may be responsible (09). Others have suggested that this lesion is actually a complex hamartoma, with overgrowth of what may represent normal embryonic nervous system tissue, rather than true neoplastic cells (33; 17); however, given the clinical and cytologic characteristics of the lesion, most authors consider this growth to be a benign tumor (09; 41), and the occasional recurrence or growth of the lesion support this. The reason for the male predominance of this lesion is not known.

Proliferative indices seen in dysembryoplastic neuroepithelial tumor range from no labeling (about 25% to 40% of cases) to slight or intermediate labeling of less than 1% of nuclei (50% to 65% of cases) to a high labeling index that may reach the level of cell division seen in high-grade gliomas (10% to 20% of cases) (08; 41; 43; 34; 10; 18). The oligodendrocytic cells of the specific glioneuronal element are rarely labeled. The labeling index of the nodules does not correspond to the aggressiveness of their histologic appearance, nor does the clinical course of highly labeled tumors deviate from the usual indolence that is a constant feature of this entity, suggesting that cell loss keeps up with cell proliferation in these cases. Molecular features of dysembryoplastic neuroepithelial tumors include mutations in fibroblast growth factor genes, mostly FGFR1 (28).

As is the case for other causes of refractory epilepsy, it is not known why the seizures produced by this entity often appear to be drug resistant. Some evidence suggests that the same multidrug transporter proteins that may produce resistance to chemotherapy in tumors (such as p-glycoprotein, multidrug resistance-associated protein 1, and major vault protein) and that are present in other causes of refractory epilepsy (such as hippocampal sclerosis) are likewise present in dysembryoplastic neuroepithelial tumors and might contribute to their antiepileptic drug resistance (38; 39).

Epidemiology

Because this tumor has been described almost exclusively in patients undergoing epilepsy surgery, the incidence or prevalence of dysembryoplastic neuroepithelial tumor in the general population has been difficult to establish. Data from the Surveillance, Epidemiology, and End Results (SEER) database from 2004 to 2013 demonstrated the incidence of dysembryoplastic neuroepithelial tumor to be 0.033 per 100,000 person years (31). Although these tumors may rarely be encountered incidentally (M Blaivas, written communication, December 1998), a study of 500 unselected brains at autopsy found no cases of dysembryoplastic neuroepithelial tumor (nor of cortical dysplasia) (43). Within the population of patients with chronic epilepsy seen at tertiary referral centers, the incidence of brain tumors is 15%; 6% of these tumors are dysembryoplastic neuroepithelial tumors, yielding an overall incidence of 0.9% in this patient group (30). Two more case series, however, using expanded diagnostic criteria for the tumor, found that 18% to 19% of patients undergoing temporal lobe resection for refractory epilepsy had a dysembryoplastic neuroepithelial tumor (10; 18).

Prevention

Differential diagnosis

Confusing conditions

The differential diagnosis for this lesion consists of:

Pleomorphic xanthoastrocytoma and ganglioglioma usually appear as cystic lesions with enhancing mural nodules, and oligodendroglial tumors are often calcified. Hamartoma must also be considered in the differential diagnosis.

Histologically, the specific glioneuronal element may be mistaken for oligodendroglioma or due to the appearance of the predominant cell type. The nodules may be mistaken for astrocytoma. Biopsy may yield only 1 of the components, setting up ready misdiagnosis.

Ganglioglioma is often considered within the pathologic differential but is distinguished by extracortical location or extension, lack of nodules, lack of neighboring dysplasia, and the presence of atypical neuronal cells and abundant connective tissue matrix (09). Proper diagnosis of this entity is crucial for patient care, as the treatment is different than that for low-grade astrocytomas.

Diagnostic workup

As with all neoplasms, the workup for this tumor includes cerebral imaging to identify and pinpoint the lesion followed by surgery to obtain tissue for pathologic analysis. On CT scanning, dysembryoplastic neuroepithelial tumor appears as a hypodense lesion, with calcification in about 25% of patients and contrast enhancement in 15% (09; 08; 36; 22; 41). Perhaps 5% of cases are CT occult. Thinning of the overlying skull is apparent on CT in approximately one third of cases, particularly when the tumor has a superficial location. In early studies, cystic appearance was reported in 28% to 50% of cases with CT or MRI (09; 10; 36; 41). However, a series using contemporary MRI techniques noted small cysts, sometimes separated from the main tumor mass, in 100% of 37 cases and found that these “pseudocysts” actually corresponded pathologically to the specific glioneuronal element (04). MRI reveals the lesion to have hyperintense signal on T2-weighted images and hypointense signal on T1-weighted images in the large majority of cases. The intracortical location of the tumor is seen much more clearly on MRI than on CT. Neither mass effect nor edema is seen. Contrast enhancement, generally focal or punctate in appearance, was seen in about 40% of cases in early series (08; 36; 22; 41) but in only 22% of cases in a more recent series (04). Thinning of the skull was described in 60% of cases in a report. The same report found that thick gyriform or nodular configuration, often with well-demarcated tumor lobules, was seen in all patients and appeared to be highly specific for dysembryoplastic neuroepithelial tumor (22). Angiography is often normal but may show neovascularization or an avascular mass with shift of normal structures.

Two studies have shown that patients with these tumors frequently have discordant or bilateral interictal epileptiform activity on EEG. However, this does not correlate with seizure outcome after surgery and, thus, need not be a source of concern (36; 23).

Molecular characterization reveals that some dysembryoplastic neuroepithelial tumors harbor FGFR fusions or duplications (37; 40). The incidence of BRAF mutations appears to be much lower than in other low-grade glial tumors, seen in less than 5% of tumors (07; 37; 01).

Surgical biopsy or resection is recommended for confirmatory pathologic diagnosis and to rule out other conditions. Tumor resection may be curative for associated epilepsy.

Management

Because patients with dysembryoplastic neuroepithelial tumor neither die nor become disabled as a consequence of neoplastic growth, active management is indicated only to treat tumor symptoms, particularly epilepsy. A substantial portion of patients with this tumor have epilepsy that is refractory to anticonvulsant medication (though there may be some ascertainment bias, with resection and diagnosis occurring more often in patients with intractable seizures). No studies examine the use of particular anticonvulsant agents in this population.

The mainstay of management of this tumor is surgical resection. Surgery is usually indicated to obtain tissue diagnosis. Surgical resection tends to be safe and a review of 185 pediatric patients with dysembryoplastic neuroepithelial tumor who underwent surgical resection showed a 12% rate of postoperative complications (35). Radiation therapy and systemic therapies have limited utility in these tumors. Patients who have received radiation for dysembryoplastic neuroepithelial tumor have been shown to have worse outcomes and decreased overall survival (31). This may be driven, in part, by more aggressive tumors receiving radiation.

Outcomes

Total or subtotal resection resulted in elimination of seizures at follow-up of 1 year or more in about three quarters of patients in 5 studies (09; 10; 36; 34; 03). However, a subsequent study indicated that such seizure-free rates decline measurably over years 2 and 3 to around 60% (32). The latter investigation also pointed to the crucial importance of complete tumor resection for long-term seizure freedom, with results far better in the patients having complete resection than in those with residual tumor (40% seizure-free vs. 88% seizure-free, respectively). Longer duration of epilepsy has been associated with worse seizure outcomes in dysembryoplastic neuroepithelial tumor patients (19; 27). It is possible that this is due to the cortical dysplasia that frequently abuts this tumor and may suggest that tumor resection should be accomplished with an ample margin to remove such dysplastic regions.

Positive prognostic factors for improved seizure control following tumor resection include younger age at diagnosis, earlier diagnosis, focal onset seizures at presentation, and larger extent of tumor resection (28).

In very rare cases malignant transformation of dysembryoplastic neuroepithelial tumor can be seen. Risk factors for malignant transformation include complex-type histologic features, location outside the temporal lobe, and subtotal resection (29).

Special considerations

In light of the presence of FGFR or BRAF abnormalities in these tumors, targeted therapies may be considered; however, at this time clinical data are limited (21).

Pregnancy

No information is available regarding the effects of pregnancy on this tumor, or vice versa.

Anesthesia

No information is available regarding the use of specific anesthetic agents in patients with dysembryoplastic neuroepithelial tumor.