Neural damage due to temporal lobe epilepsy: Dual-nuclei (proton and phosphorus) magnetic resonance spectroscopy study

Authors


Dr Takayuki Obata, Department of Medical Imaging, National Institute of Radiological Sciences, 4-9-1 Anagawa, Inage-ku, Chiba 263-8555, Japan. Email: t_obata@nirs.go.jp

Abstract

Abstract  The aim of this study was to evaluate the usefulness of proton and phosphorus (1H and 31P) magnetic resonance spectroscopy (MRS) for temporal lobe epilepsy (TLE) patients, and to evaluate neural damage and metabolite dysfunction in the TLE patient brain. We performed 1H and 31P MRS of medial temporal lobes (MTL) in the same TLE patients (n = 14) with a relatively wide range of severity from almost seizure-free to intractable, and calculated the ratio of N-acetylasparate to choline-containing compounds and creatine + phosphocreatine (NAA/Cho + Cr) in 1H MRS and inorganic phosphate to all main peaks (%Pi) in 31P MRS. There was no significant correlation between NAA/(Cho + Cr) and %Pi  in  each  side  (ipsilateral,  r = −0.20;  contralateral,  r =−0.19). The values of NAA/(Cho + Cr) showed a significant difference between ipsilateral and contralateral MTLs to the focus of TLE patients (P < 0.01, paired t-test). Although %Pi also had a tendency to show the laterality of TLE, there was no significance. Ipsilateral (r = −0.90, P < 0.0001) and contralateral (r = −0.70, P < 0.005) NAA/(Cho + Cr) decreases and contralateral %Pi increase (r = 0.81, P < 0.001) had significant correlation with seizure frequency. 1H MRS provides more important information concerning neuronal dysfunction in MTL of TLE patients than 31P MRS.

INTRODUCTION

Surgical treatment for medically intractable temporal lobe epilepsy (TLE) is an effective procedure in selected patients in whom the epileptogenic focus is localized in a unilateral anterior and medial temporal lobe (MTL).1 Accurate localization of the epileptogenic focus has largely and traditionally been dependent on a scalp-sphenoidal electroencephalogram (EEG). Recently, magnetic resonance imaging (MRI), positron emission tomography (PET) and single photon emission computed tomography (SPECT) have provided important additional information about the epileptogenic focus.2,3 However, subdural EEG electrode implantation is still required when imaging and scalp EEG are inconclusive. Thus, a more sensitive and non-invasive neurological measurement modality is needed that interictally lateralizes the epileptogenic focus and provides useful information about the pathological changes of TLE.

Magnetic resonance spectroscopy (MRS) is widely used in the field of psychiatry and clinical neuroscience.4–17 Proton and phosphorus magnetic resonance spectroscopies (1H and 31P MRS) were non-invasive functional neurological measurement modalities. Proton MRS of the brain includes signals from N-acetylasparate (NAA), choline-containing compounds (Cho) and creatine + phosphocreatine (Cr). There have been several studies that observed a reduction in the NAA signal or its ratio to other metabolic signals in the ipsilateral temporal lobe to the epileptogenic focus in TLE patients, suggesting that NAA concentration reflects neural damage.4–13 The values of inorganic phosphate (Pi)  and  phosphocreatine  (PCr)  that  can  be  detected in 31P MRS change according to the condition of oxidative phosphorylation, and they were reported to be good indexes for detecting brain metabolite dysfunction.14–16

Magnetic resonance nuclei have been used in TLE studies, but studies using both nuclei MRS of the same TLE patients have hardly been reported. In the present study, we aimed to evaluate the 1H and 31P MRS signal changes in bilateral temporal lobes of TLE patients with a relatively wide range of severity from almost seizure-free to intractable, and to clarify the relation between neural damage and metabolite dysfunction in the TLE patient brain. Some 1H MRS data in this study was from that already reported by Someya et al.11

METHODS

Subjects

This study was approved by the Ethics and Radiation Safety Committee of the National Institute of Radiological Sciences, Chiba, Japan. We studied 14 patients, four males and 10 females, aged 31.5 ± 8.4 years (mean ± SD), who were diagnosed with the amygdala-hippocampus type of TLE on the basis of clinical symptoms and EEG findings.18 They were asked to participate in this study, and written informed consent was obtained. The patients showed unilateral temporal spikes or sharp waves in their repetitive (three or more) scalp EEG and/or stereotactic implanted depth electrode EEG record. The data of the patients are summarized in Table 1. Seven healthy volunteers, five males and two females, aged 27.0 ± 5.4 year (mean ± SD), were also studied as controls. No significant difference was observed in gender (P = 0.06; Pearson's χ2 test) or age (t = 1.37, P = 0.19; t-test) between controls and patients.

Table 1.  Patient demographic information
SubjectSexAgeDrug (mg)Onset age (y)Epilepsy duration (y)EEG focusSeizure type
per month
Seizure frequencyMRI
  • CBZ, carbamazepine; CZP, clonazepam; DPH, diphenylhydantoin; PB, phenobarbital; VPA, valproate sodium; ZNS, zonisamide; SPS, simple partial seizure; CPS, complex partial seizure; SGS, secondarily generalized seizure; MTL, medial temporal lobe.

  • Results by stereotactic implanted depth electrode EEG.

 1F20DPH 225, CBZ 800, ZNS 20015 5LeftSPS + CPS<1Atrophy of Lt-MTL
 2F22CBZ 300, DPH 10020 2LeftSPS + CPS<1(−)
 3F27CZP 4, ZNS 400, VPA 1000 225LeftCPS<1Atrophy of Lt-MTL
 4F30CBZ 600, ZNS 5001020LeftSPS + CPS<1(−)
 5M22DPH 20014 8LeftSPS + CPS<1(−)
 6M30CBZ 40026 4RightCPS<1(−)
 7F20ZNS 200, CBZ 800, DPH 10015 5LeftSPS + CPS18.3(−)
 8F34CBZ 800, ZNS 4001024LeftCPS+SGS 8.5(−)
 9F37PB 50, CBZ 1200, DPH 3501819RightSPS + CPS+SGS33.3(−)
10F37CBZ 800, ZNS 400 928LeftSPS + CPS+SGS 3.0(−)
11F40CBZ 800, ZNS 2001525LeftCPS+SGS 2.6(−)
12F47CBZ 1000, VPA 6002027LeftSPS + CPS+SGS 3.6Atrophy of Lt-MTL
13M36VPA 600 630RightCPS 9.5Atrophy of Rt-MTL
14M39DPH 450, CBZ 800, VPA 1600 336RightCPS+SGS 8.8(−)

Magnetic resonance spectroscopy

The MR system used in this study was a Gyroscan ACS2 operated at 1.5T (Philips Medical Systems Ltd, Best, The Netherlands). 1H and 31P MRS were performed with a birdcage-type coil. The volumes of interest (VOIs) were set at the bilateral MTL under the guidance of proton scout images, with their size being 6 × 3 × 3 cm so as to include all parts of the hippocampus: the head, body, tail, gray and white matter, and amygdala. For 1H MRS, volume selection was done by spin echo sequence for each side. Repetition time (TR) was 1500 msec. The echo time (TE) used was 136 msec because lactate peaks are visible at this TE. There were 1024 data points, and the bandwidth was 2000 Hz. Scan averages were 256 times and scan time was 6.4 min for each side. For 31P MRS, the volume selection was done by two-VOI image-selected in vivo spectroscopy (2-VOI ISIS) sequence. TR was 3000 msec. Data points were 1024 and the bandwidth was 2000 Hz. The scan time was 20 min (384 scans).

To evaluate the structural brain changes, approximately 90 coronal MR images, covering the entire brain, were taken for each subject. An inversion recovery pulse sequence (TR, 2500 msec; TE, 20 msec; inversion time, 300 msec) was used. Matrix size was 205 × 256, and the FOV was 230 mm. Slice thickness was 2 mm without slice gaps. Total scan time was 21 min and 25 s. The coronal slice images were parallel to the baseline of the fourth ventricle. Atrophy of the MTL was judged by two experienced specialists, a radiolologist and a neuropsychiatrist, while were blind to the EEG lateralization.

Data processing and analysis

The spectra in 31P MRS were processed with a 15-Hz exponential filter in the time domain. No filter was used in 1H MRS. To measure the area under each peak, a semi-automatic line-fitting routine assuming a purely Lorentzian line shape was used. NAA intensities were evaluated as relative ratio to the total ratio of Cho and Cr. The Pi value was shown as relative value to the total value of all the peaks (phosphoesters, Pi, phosphocreatine, and adenosine triphosphate).

The paired t-test was used for comparisons of ipsilateral and contralateral MTL to the EEG focus, and the unpaired t-test was used for comparisons of control and patient MTL. Logarithmic transformation was used to normalize the data of seizure frequency. Pearson product-moment correlation analysis was used to evaluate the correlation between each set of MRS data and seizure frequency. Differences were considered to be significant at P < 0.05.

RESULTS

There was no significant correlation between NAA/(Cho + Cr) and %Pi on each side (ipsilateral r = −0.20; contralateral r = −0.19). Atrophy of MTL was detected by MRI in four patients (Table 1). Their lateralities were consistent with the EEG and MRS lateralities except for one case that had lower %Pi in the ipsilateral MTL.

Proton MRS

Proton MRS lateralization based on a reduced NAA/(Cho + Cr) ratio was in agreement with the clinical EEG lateralization in 12 of 14 patients, and a significant difference between the ipsilateral and contralateral MTL (P < 0.01, paired t-test) (Table 2) was obtained. The NAA/(Cho + Cr) ratio in the ipsilateral or contralateral MTL was not significantly different from that in MTL of control subjects. Two cases with disagreement between the results of 1H MRS and clinical data were in the low seizure-frequency group (< 1/month).

Table 2.  Relative signal intensity of proton and phosphorus MRS
 ControlIpsilateralContralateral
  • Values are mean ± SD.

  • NAA, N-acetylasparate; Cho, choline-containing com-pounds; Cr, creatine + phosphocreatine; Pi, inorganic phosphate.

  • *

    P < 0.01, paired t-test.

%NAA/(Cho + Cr) 82 ± 772 ± 9*79 ± 7*
%Pi5.0 ± 1.45.8 ± 1.44.9 ± 1.3

NAA/(Cho + Cr) in the ipsilateral and contralateral MTL decreased as the seizure frequency increased. Both ipsilateral and contralateral NAA/(Cho + Cr) were significantly correlated with seizure frequency (r = −0.90, P < 0.0001; r = − 0.70, P < 0.005, respectively; Fig. 1a).

Figure 1.

(a) Correlation of seizure frequency vs NAA/(Cho + Cr) ratio and (b) seizure frequency vs%Pi ratio. Logarithmic transformation was used to normalize seizure frequencies. Horizontal lines and gray areas show mean values and standard deviations of the controls. The lines represent the expected values in contralateral (dotted line) and ipsilateral (solid line) medial temporal lobes to the focus obtained by linear regression. Both ipsilateral and contralateral NAA/(Cho + Cr) were significantly correlated with the seizure frequency (r = −0.90, P < 0.0001; r= −0.70, P < 0.005, respectively; Fig. 1a). The contralateral %Pi values showed significant correlation with seizure frequency (r = 0.81; P < 0.001) but, as the ipsilateral values were very scattered, no significant result was obtained. f, seizure frequency; Ipsi-NAA, ipsilateral NAA/(Cho + Cr); contra-NAA, contralateral NAA/(Cho + Cr); Contra-Pi, contralateral %Pi.

Phosphorus MRS

Phosphorus MRS lateralization based on an increase in %Pi ratio was consistent with the clinical EEG lateralization in 10 of 14 patients, but there was no significant difference (Table 2). The %Pi value in the ipsilateral or contralateral MTL was not significantly different from that in MTL of control subjects. In 10 of 14 patients, the %Pi values in the ipsilateral MTL were higher than the mean value of normal control. They were very scattered, and had no significant correlation with the seizure frequency (r = 0.09, Fig. 1b). Four cases with disagreement between the results from 31P MRS and clinical data were in the high seizure-frequency group (> 1/month). The %Pi values in the contralateral MTL of all the low seizure-frequency group patients were lower than the mean value of the normal control, while those in six of eight patients in the high seizure-frequency group were higher than the mean value of normal control, resulting in significant correlation (r = 0.81; P < 0.001, Fig. 1b).

DISCUSSION

Significant changes in the NAA/(Cho + Cr) ratio of 1H MRS and the %Pi ratio of 31P MRS in the medial temporal lobes of TLE patients were observed. Both seem useful for the detection of TLE lateralization, but they have different tendencies.

The values of NAA/(Cho + Cr) in 1H MRS showed significant differences between ipsilateral and contralateral MTL of TLE patients. Although %Pi also had a tendency to show the laterality of TLE, there was no significance. This indicates that 1H MRS may be more useful for generally assessing the lateralization of the TLE focus. Most previous studies have shown that TLE patients have lower NAA values in their ipsilateral MTL in 1H MRS. As N-acetylasparate is mainly located in neural cells, a reduction in NAA value is believed to reflect some neural damage.3–13 In contrast, the concentrations of Cho and Cr are much higher in astrocytes and oligodendrocytes than in neurons,19 meaning that changes in their values may reflect reactive astrocytosis. Thus, the decrease in NAA/(Cho + Cr) ratios in ipsilateral MTL may reflect neuronal loss and reactive gliosis as observed in medial temporal sclerosis in TLE patients. We also found a significant correlation between NAA/(Cho + Cr) and the seizure frequency of TLE patients, suggesting that 1H MRS is useful not only for assessment of TLE lateralization but also for evaluation of TLE severity. Significant correlation between contralateral NAA/(Cho + Cr) and the seizure frequency was also detected, suggesting that neural damage due to TLE extended to the contralateral MTL in the high-seizure frequency group.

Although the mean value of %Pi in ipsilateral MTL was higher than in normal controls, the values were widely scattered, and no significant correlation with the seizure frequency was detected. The mean value of %Pi in the contralateral MTL in the low seizure-frequency group (< 1/month) was lower than that of normal controls. The %Pi increase is believed to reflect dysfunction of energy metabolism.16 These results may indicate a hyper-energy status in the contralateral MTL as compensation for dysfunction in the ipsilateral MTL.15 In contrast, the mean value of %Pi in the contralateral MTL in the high seizure-frequency group (> 1/month) was higher than that of normal controls. This seems to indicate a spreading of neural dysfunction to the contralateral MTL. These quite different conditions under different seizure frequencies in contralateral %Pi produced significant correlation with seizure frequency.

The NAA/(Cho + Cr) ratio in the ipsilateral MTL has the largest correlation coefficient with seizure frequency among the four spectroscopy data sets, suggesting that the NAA/(Cho + Cr) ratio in the ipsilateral MTL is the best index for showing TLE severity. The lateralization of the TLE focus is of major importance, especially for presurgical decisions for intractable TLE.  In  the  present  study,  the  lateralization  of  TLE in the low seizure-frequency group (< 1/month) was detected more clearly using 31P MRS, while 1H MRS showed highly significant lateralization detectability in the high seizure-frequency (> 1/month) group. Two cases with disagreement between the 1H MRS results and clinical data were in the low seizure-frequency group, but four cases with disagreement between the 31P MRS results and clinical data were in the high seizure-frequency group. In total, 1H MRS seems more useful for clinical application.

Although only four patients showed atrophies, the MTL atrophies detected by MRI were located on the ipsilateral side to the focus, and all of the four had smaller NAA/(Cho + Cr) ratios in ipsilateral MTL than in contralateral MTL. Some investigations have reported that morphological MRI study of TLE patients is very useful for evaluating the severity and laterality of TLE.20,21 The reason for the low atrophy detectability by MRI in the present study may lie in the fact that we did not use a pulse sequence with specificity for evaluating MTL atrophy.

In conclusion, we performed 1H and 31P MRS for detecting the lateralization of the TLE focus. There was no significant correlation between NAA/(Cho + Cr) and %Pi. The values of NAA/(Cho + Cr) in 1H MRS showed significant differences between ipsilateral and contralateral MTL of TLE patients. The decrease in ipsilateral NAA/(Cho + Cr) had the largest correlation coefficient with seizure frequency. These results suggest that 1H MRS provides more important information than 31P MRS concerning neuronal dysfunction in MTL of TLE patients.

ACKNOWLEDGMENTS

This work was supported in part by a Grant-in-Aid from the Ministry of Education, Science, Sports and Culture of Japan, by funds for the Research Project with Heavy Ions at the National Institute of Radiological Sciences – Heavy-ion Medical Accelerator in Chiba (NIRS-HIMAC), by the fund for the Neuroscience Project at the National Institute of Radiological Sciences, and by a Postdoctoral Fellowship from the Japan Science and Technology Corporation.

Ancillary