Address correspondence and reprint requests to Dr. T. Maehara at Neurosurgery, Department of Brain Medical Science, 1-5-45, Yushima, Bunkyo-ku, Tokyo 113-8519, Japan. E-mail: email@example.com
Summary: Purpose: We assessed the diagnostic value of [11C]methionine (MET) positron emission tomography (PET) in the differential diagnosis of dysembryoplastic neuroepithelial tumors (DNETs) among benign tumors associated with temporal lobe epilepsy (TLE).
Methods: This series consisted of seven TLE patients with benign tumors in the temporal lobe. After MET-PET study, all seven patients underwent tumor resection along with focus excision. The uptake of tracers was evaluated by the lesion-to-contralateral ratio (L/C ratio) and the standardized uptake value (SUV). We also assessed the relation between MET uptake and proliferation capacity observed in the surgical specimens.
Results: Whereas four patients with DNETs did not show high MET uptake visually, the ganglioglioma and gliomas of the remaining three patients were identified as high–MET-uptake lesions. In the DNETs, the SUV ranged from 1.03 to 1.41, and the L/C ratio ranged from 0.99 to 1.14. MET uptake was significantly lower in the patients with DNETs than in the patients with ganglioglioma and brain gliomas (SUV, p = 0.045; L/C ratio, p = 0.0079). The Ki-67 labeling index was 4% in one patient with DNET and 5% in one patient with pleomorphic xanthoastrocytoma (higher labeling index). The higher labeling index was not related to high MET uptake based on the SUV (p = 0.91) and L/C ratio (p = 0.38).
Conclusions: Negative MET uptake in benign temporal lobe tumors with TLE is consistent with a preoperative diagnosis of DNET.
Benign tumors are one of the most important pathologies in patients with temporal lobe epilepsy (TLE) (1,2). Among them, dysembryoplastic neuroepithelial tumors (DNETs) are well recognized as surgically curable tumors with strong epileptogenic potential (3,4). Thus the differential diagnosis of DNETs from other low-grade tumors is mandatory for planning surgical strategy. The DNET on magnetic resonance imaging (MRI) is typically demonstrated as a well-demarcated lobulating tumor without perifocal edema, often showing a multicystic appearance (5,6). In cases without a multinodular and multicystic appearance, it is not always easy to differentiate DNETs from gangliogliomas and low-grade gliomas on MRI examination.
In addition to MRI, positron emission tomography (PET) studies have been recently introduced as important examinations for epilepsy patients with DNETs (6–9). On [18F]fluorodeoxyglucose (FDG) PET, DNETs are observed as hypometabolic lesions (6–8). On [11C]flumazenil (FMZ) PET, a procedure that discloses focal abnormalities of type A benzodiazepine (BZD) receptor, DNETs are demonstrated as low-uptake lesions (9). Given the associations of these findings with epileptogenicity, it can be difficult to make the differential diagnosis for benign tumors associated with TLE.
PET with [11C]methionine (MET) is considered a useful measure for identifying brain gliomas (10). Recent reports indicated that MET uptake of DNETs is the same as that of surrounding cerebral cortex (8,11). Fortunately, a negative finding on MET-PET seems to be helpful for the differential diagnosis of DNET, because gliomas are usually manifested as increased-uptake lesions (10,12).
In this study, we assessed the diagnostic value of MET-PET in the differential diagnosis of DNET among benign tumors associated with TLE. To accomplish this, we analyzed the MET uptake of the benign tumors visually and quantitatively. In addition to differential diagnosis, we assessed the relation between MET uptake and the proliferation capacity observed in the surgical specimens.
PATIENTS AND METHODS
This series consisted of seven patients with a type of complex partial seizure (CPS) associated with TLE who were found to have benign tumors in the temporal lobe. They included three male and four female patients, ranging in age from 13 to 48 years. The age at seizure onset ranged from 6 to 26 years, and the time to the operation ranged from 2 to 28 years (Table 1). High-resolution MRI with a 1.5-T magnet detected organic lesions in the temporal lobes. MET-PET was performed in all seven patients, and fluorodeoxyglucose (FDG)-PET was performed in four patients.
After presurgical examinations including repeated EEGs, video-EEG monitoring, and neuropsychological testing, three patients underwent invasive monitoring for focus detection and functional mapping. All seven patients underwent tumor resection along with focus excision under intraoperative electrocorticographic (ECoG) guidance (Table 1).
The PET study was performed with 50 axial slices of tomographic images with a center-to-center interval of 3.125 mm, scanned by a Headtome-V scanner (Shimadzu Corp., Kyoto, Japan). The in-plane spatial resolution was 6.0 mm full width, half maximum (FWHM). For the imaging of possible cerebral tumors, the subjects were intravenously injected with 11C-labeled l-methionine (300 MBq). Twenty minutes after the injection, equilibrated brain radioactivity was measured for 3 min. For the imaging of cerebral glucose metabolism, the subjects were intravenously injected with 18F-labeled fluorodeoxyglucose ([18F]FDG; 100 MBq). Forty-five minutes after the injection, equilibrated brain radioactivity was measured for 3 min.
Medical staff affirmed an absence of clinical seizures throughout the 3 h leading up to the PET scan and during the PET scan itself. The patients themselves also indicated that they had remained seizure free during the PET scan. With this information, we confirmed that all of the PET measurements were performed in an interictal state. Using Dr. View image-analysis software (Asahi Kasei Joho, Tokyo, Japan) on a personal computer, the regions of interest (ROIs) were placed on the tumor and contralateral brain over the MRI and then transferred to the PET images co-registered with MRI (12). The uptake of tracers was evaluated by the lesion-to-contralateral ratio (L/C ratio) and the standardized uptake value (SUV), a parameter calculated by dividing the brain activity of tracer per volume of brain by the injected activity of tracer per body weight.
Surgical specimens were examined to investigate the tumor histology and Ki-67 labeling index. The microdysgenesis (13,14) of the resected foci also was examined.
The differences in MET uptake based on the SUV and L/C ratio between patients with DNETs and patients with ganglioglioma or gliomas were analyzed for statistical significance by using Student's t test. The difference in MET uptake based on the SUV and L/C ratio between patients with higher and lower labeling index also was analyzed for statistical significance.
All seven of the patients showed interictal spikes in the anterior temporal region ipsilateral to the tumor location. Four patients also showed interictal spikes in the lateral temporal region. Among three patients with invasive monitoring, seizure onset was recorded in the lateral temporal lobe in two patients and in the medial temporal lobe in one patient. The sensory speech area of the three patients detected by electrical stimulation was located outside the tumor or focus area. After tumor resection, a focus excision was performed until the intraoperative ECoG showed no spikes surrounding the focus cortex.
After surgery, five of the patients remained seizure free, and two had a rare-seizure outcome for ≥1 year (Table 1). Tumor recurrence has not been observed in any of the seven patients.
Figure 1 shows the MET-PET images for each patient. The tumors were identified as high–MET-uptake lesions in three cases and were undetected in four patients. Among the four patients who underwent FDG-PET, only two showed high MET uptake in the tumors, although all four showed low glucose uptake in the tumor and surrounding cerebral cortex (Table 1).
Pathological examination demonstrated DNETs in four cases, ganglioglioma in one case, low-grade astrocytoma in one case, and pleomorphic xanthoastrocytoma (PXA) in one case. Patient 2 (Table 1) showed a number of ectopic neurons in the white matter, a finding compatible with microdysgenesis. However, none demonstrated typical findings of focal cortical dysplasia (fCD), including balloon cells and dysplastic neurons [15)].
Comparison between MET-PET and histopathologic data
Although four patients with DNETs did not show high MET uptake visually, the ganglioglioma and gliomas of the remaining three patients were identified as high–MET-uptake lesions. In the DNETs, SUV ranged from 1.03 to 1.41, and the L/C ratio ranged from 0.99 to 1.14 (Table 1). MET uptake was significantly lower in patients with DNETs than in patients with ganglioglioma and brain gliomas (SUV, p = 0.045; L/C ratio, p = 0.0079).
In the assessment of the proliferation capacity, the Ki-67 labeling index was 4% in one patient with DNET and 5% in one patient with PXA (higher labeling index). The labeling index of the remaining five patients was <1% (Table 1). The high Ki-67 labeling index was not related to high MET uptake based on the SUV (p = 0.91) and L/C ratio (p = 0.38).
In the present study, we tried to elucidate the clinical role of MET-PET for patients with DNETs. Although the patient population is as small as seven patients, our findings demonstrated that negative MET uptake was useful for the differential diagnosis of DNET. A ganglioglioma and several brain gliomas were clearly found as high–MET-uptake lesions, whereas DNETs could not be visualized. With quantitative evaluation, we found that the SUV and L/C ratio were significantly lower in patients with DNETs than in patients with ganglioglioma and brain gliomas (SUV, p = 0.045; L/C ratio, p = 0.0079).
Kaplan et al. (8) were the first to report the clinical application of MET-PET in patients with DNETs and complex partial seizures. In five children from their series, they found that DNETs were associated with negative or minimal MET uptake. In a series of 30 intracranial small tumors studied by Braun et al. (11), one DNET case was revealed as a MET-negative lesion. These findings are compatible with the difficulty we experienced in visualizing DNETs by MET-PET. We further introduced quantitative analysis and demonstrated that the L/C ratios of the four DNETs ranged from 0.99 to 1.14. We confirmed that DNETs show negative MET uptake by both visual and quantitative analysis.
By contrast, gangliogliomas and brain gliomas are usually identifiable as MET-positive lesions (10). Several case reports demonstrated high accumulations of MET in patients with gangliogliomas and PXA (11,16). Previous articles also affirmed that oligodendrogliomas and astrocytomas were sensitive to MET tracers (10,17). Kaplan et al. (8) reported that 20 pediatric cases with low-grade gliomas, either astrocytoma or oligodendroglioma, were consistently associated with increased MET uptake. In the present series of patients with ganglioglioma and brain gliomas, tumors were clearly visualized by MET-PET, and quantitative analysis demonstrated that MET uptake was significantly higher in the lesions than in the DNETs. Among benign tumors with TLE, negative MET uptake is consistent with a preoperative diagnosis of DNET.
Although DNETs show a negative MET uptake, we must consider the possibility of other benign tumors being found as low–MET-uptake lesions. For example, several reports have indicated that negative MET uptake was observed in 3.4% of 89 cases with low-grade gliomas (18) and 8.3% of 12 cases with low-grade astrocytomas (19). However, benign tumors with TLE have unique characteristics in their histology and incidence. In a recent large series of temporal lobectomies, Pasquier et al. (20) reported 94 benign tumors including 61 cases of DNET, 29 cases of ganglioglioma, and four cases of PXA, whereas Clusmann et al. (21) reported 113 benign tumors including 55 cases of ganglioglioma, 16 cases of DNT, 27 cases of astrocytomas, nine cases of oligoastrocytomas, two cases of oligodendrogliomas, and four cases of PXA. All of these tumors, with the exception of DNETs and a few cases of low-grade gliomas, are demonstrated as high–MET-uptake lesions (10,11,16,17). Therefore negative MET uptake has a significant value in the preoperative diagnosis of DNET.
We also investigated the relations between MET uptake and proliferation activity. Because high accumulation of MET in brain gliomas is usually associated with malignancy grade (10,17), we examined whether negative MET uptake in DNET results from a low proliferation capacity. In the present study, the high Ki-67 labeling index was not related to high MET uptake. All four of the patients with DNET showed negative MET uptake, regardless of whether the Ki-labeling index was low or high. In a series of 27 pediatric tumors, Utriainen et al. (17) reported that MET-PET had a high value for tumor detection, but less value for assessing the malignant potential of the tumors. They pointed out that MET uptake is associated with tumor malignancy when analyzed in single tumor histologies, but not always when analyzed in the histologies of plural tumor types. In benign tumors, MET uptake seems to be related to tumor histology but may not be applicable to compare the proliferation capacity.
We also noticed those reports stating that epileptogenic areas based on fCD were clearly visualized by MET-PET (22,23). None of our cases showed MRI features of fCD, but we carefully examined the presence of various candidates of microdysgenesis, such as neuronal clustering, the neuronal columnar structure, superficial hypermyelination, white matter neurons, and perivascular glial satellitosis (13), in our samples. This was done because, in a series of 40 DNET cases by Daumas-Duport et al. (3,4), 19 (47%) cases revealed CD, either resembling the typical form described by Taylor et al. (15), or rather resembling microdysgenesis. In the present study, one patient showed a number of ectopic neurons in the white matter, which has recently been defined as a definite component of microdysgenesis (14), but none demonstrated typical findings of fCD including balloon cells and dysplastic neurons (15). Sasaki et al. (22) reported that four patients who had typical fCD features on both MRI and histologic examinations showed a moderately increased MET uptake, which ranged from 1.10 to 1.37 in L/C value. With respect to the neuroradiologic diagnosis of fCD, Yagishita et al. (24) reported that 12 of 13 cases pathologically proven as typical fCD were demonstrated by MRI examination. These data suggest that positive MET uptake reported in fCD had little effect on the differential diagnosis of DNET, when MRI examinations do not show typical features of fCD.
In the seven patients with TLE and benign tumors in this study, we observed a high MET uptake in ganglioglioma and brain gliomas, but not in DNETs. Negative MET uptake in the benign tumors with TLE is consistent with a preoperative diagnosis of DNET. However, the accumulation of more cases will be needed before we can reach a firm conclusion.