To quantitatively compare the diagnostic capability of double inversion-recovery (DIR) with F-18 fluorodeoxyglucose positron emission tomography (FDG-PET) for detection of seizure focus laterality in temporal lobe epilepsy (TLE).
To quantitatively compare the diagnostic capability of double inversion-recovery (DIR) with F-18 fluorodeoxyglucose positron emission tomography (FDG-PET) for detection of seizure focus laterality in temporal lobe epilepsy (TLE).
This study was approved by the institutional review board, and written informed consent was obtained. Fifteen patients with TLE and 38 healthy volunteers were enrolled. All magnetic resonance (MR) images were acquired using a 3T-MRI system. Voxel-based analysis (VBA) was conducted for FDG-PET images and white matter segments of DIR images (DIR-WM) focused on the whole temporal lobe (TL) and the anterior part of the temporal lobe (ATL). Distribution of hypometabolic areas on FDG-PET and increased signal intensity areas on DIR-WM were evaluated, and their laterality was compared with clinically determined seizure focus laterality. Correct diagnostic rates of laterality were evaluated, and agreement between DIR-WM and FDG-PET was assessed using κ statistics.
Increased signal intensity areas on DIR-WM were located at the vicinity of the hypometabolic areas on FDG-PET, especially in the ATL. Correct diagnostic rates of seizure focus laterality for DIR-WM (0.80 and 0.67 for the TL and the ATL, respectively) were slightly higher than those for FDG-PET (0.67 and 0.60 for the TL and the ATL, respectively). Agreement of laterality between DIR-WM and FDG-PET was substantial for the TL and almost perfect for the ATL (κ = 0.67 and 0.86, respectively).
High agreement in localization between DIR-WM and FDG-PET and nearly equivalent detectability of them show us an additional role of MRI in TLE.
Temporal lobe epilepsy (TLE) is the most common form of medically refractory epilepsy, and surgical treatment is adopted in such cases (Semah et al., 1998; Choi et al., 2008). Therefore, presurgical examinations are important in locating a seizure focus and its laterality. In TLE, fluorine-18 fluorodeoxyglucose positron emission tomography (FDG-PET) has been shown to achieve high detectability regarding the lateralization of a seizure focus, as decreased glucose uptake in the epileptogenic temporal lobe (Hammers, 2012) and better surgical prognosis has been reported in cases with FDG-PET hypometabolism (LoPinto-Khoury et al., 2012). Relative to FDG-PET, magnetic resonance imaging (MRI) has been less reliable in detecting abnormalities related to seizure foci (King et al., 1998; Engel et al., 2008; Takaya et al., 2012), although it can provide detailed anatomic information that is indispensable for clinical practice.
Double inversion-recovery (DIR) is a relatively new MR sequence that nullifies signals from both the cerebrospinal fluid and white matter (WM) and improves lesion detection (Turetschek et al., 1998); a recent development is three-dimensional (3D) acquisition (Busse et al., 2006). DIR is considered helpful for increased detectability of hippocampal sclerosis (HS; Zhang et al., 2011) and other abnormal signal changes in epilepsy (Rugg-Gunn et al., 2006; Salmenpera et al., 2007; Morimoto et al., 2013). Recently, by visually evaluating the presence or lateralization of the anterior temporal lobe white matter abnormal signal (ATLAS), better diagnostic capability of DIR at 3T was reported in determining seizure focus laterality of TLE in higher agreement with clinically diagnosed seizure focus laterality than that of T2-weighted imaging (T2WI) or fluid-attenuated inversion recovery (FLAIR). (Morimoto et al., 2013). In patients with TLE, ATLAS is observed as an increased signal area or a loss of gray–white matter demarcation in the anterior temporal lobe (ATL) WM ipsilateral to the seizure focus on T2WI (Coste et al., 2002). This finding has been considered as an indicator of focus laterality (Kuzniecky et al., 1987; Choi et al., 1999; Meiners et al., 1999; Mitchell et al., 1999; Coste et al., 2002; Adachi et al., 2006; Morimoto et al., 2013), and it is not observed in healthy controls (Briellmann et al., 2004; Morimoto et al., 2013). High detectability of seizure focus laterality on DIR gives us an expectation that MRI would play a more important role, comparable to PET for TLE patients without radiation exposure. However, objective and quantitative comparison of these modalities has not yet been carried out.
The purposes of this study were the following: (1) to evaluate diagnostic capability of DIR in quantitatively determining seizure focus laterality of patients with TLE, and (2) to compare the capability of DIR with that of FDG-PET.
This study was approved by the institutional review board. All subjects gave written informed consent prior to enrollment in this study.
All of the patients (25 women, 26 men) with a mean age of 33 years (range, 2 months–79 years) suspected of epilepsy at our institution from September 2009 to October 2010 were prospectively enrolled at the initial stage. Seizure foci were diagnosed by consensus of board-certified epileptologists (A.I. and R.M., with 26 and 17 years of clinical experience in epileptology, respectively) as located in the temporal, frontal, parietal, and occipital lobes in 28, 6, 5, and one case, respectively. Foci were considered generalized or uncategorized in three and eight cases, respectively. For focus localization, all available clinical examinations were used, namely: routine electroencephalography (EEG) studies based on the International 10-20 System with additional anterior temporal electrodes (T1 and T2) (n = 51); long-term video-EEG monitoring with scalp electrodes (n = 16); single photon emission computed tomography (SPECT) (n = 6); magnetoencephalography (MEG) (n = 4); and routine MRI (n = 51). MR images were used only to find a lesion suggestive of a seizure focus, without evaluating ATLAS. Detailed neurologic history and physical examinations were also referenced. In addition, four patients were examined using intracranial EEG with subdural electrodes.
Among 28 TLE patients, 13 patients were excluded due to presence of apparent abnormal lesions other than HS in the temporal lobes (TLs); postinflammatory changes (n = 1), postoperative changes (n = 2), ectopic gray matter (n = 1), pilocytic astrocytoma (n = 1), mass lesion suspected of low grade glioma (n = 1), inadequate MR image quality (n = 1), no PET study (n = 3), and use of a different PET system (n = 3). The remaining 15 TLE patients (8 women, 7 men) with a mean age of 33 years (range, 17–63 years) were included for further analysis. Four, six, and five of these patients were clinically diagnosed as having right, left, or bilateral seizure, respectively. The seizure foci of all five patients with bilateral TLE were predominantly one-sided (having bilateral epileptiform discharges, but observed predominantly on one side ranging from 60–80%: right in four patients and left in one patient). Therefore, among these 15 patients, 8 and 7 patients were sorted to the right TLE group (RTLE) and left TLE group (LTLE), respectively. Seven patients (47%) in 15 TLE patients had HS evaluated by two neuroradiologists (T.O. and E.M., with 22 and 8 years of experience in neuroradiology, respectively) on routine T2WI and FLAIR (Hanamiya et al., 2009). The diagnosis was determined in consensus of the evaluators in the case of discordant results. The details of clinical information of the 15 patients with TLE are summarized in Table 1.
|Case||Seizure laterality||Focus area||Aura||Seizure type||Duration of epilepsy||Hippocampal sclerosis|
|3||B(R)||Mesial||+||CPS, GTCS||2 years||−|
|4||L||Mesial||+||CPS, GTCS||28 years||+|
|8||B(L)||Mesial||−||CPS, GTCS||2 years||+|
|9||L||Mesial||+||CPS||2 years 3 months||+|
|11||B(R)||Mesial||+||SPS, GTCS||5 months||−|
|12||L||Lateral||+||CPS, GTCS||7 years||−|
|14||L||Mesial||+||CPS, SPS||20 years||+|
|15||B(R)||Mesial||+||CPS||4 years 7 months||−|
As normal controls, 21 healthy subjects were enrolled for MRI and an additional 21 healthy subjects for FDG-PET imaging. None of these subjects had a history of epileptic attack, head trauma, or other disorders of the central nervous system. In both MR and PET imaging, two subjects from each group were excluded because of inadequate image quality owing to motion artifacts. The remaining 19 healthy volunteers for MR (7 women, 12 men) with a mean age of 32 years (range, 24–63 years) and 19 healthy subjects for PET (12 women, 7 men) with a mean age of 36 years (range, 18–54 years) were used as the controls in this study.
All MR examinations of both normal volunteers and patients with epilepsy were conducted using a 3T whole-body MR system (MAGNETOM Trio, A Tim System; Siemens Healthcare, Erlangen, Germany) with a 32-channel head coil. The single-slab 3D DIR was implemented using a turbo spin-echo sequence with variable flip angles (work in progress provided by Siemens; Mugler et al., 2000). The acquisition parameters are summarized in Table S1, including those of T2WI and FLAIR, as well as those of 3D magnetization-prepared rapid acquisition gradient-echo (MPRAGE) images acquired for anatomical reference for all subjects.
All PET images were acquired using a PET scanner (Advance; General Electric Medical Systems, Milwaukee, WI, U.S.A.). After fasting for at least 4 h, FDG was administered to the subjects intravenously, and images were acquired approximately 40 min later. The administered dose of FDG was approximately 370 MBq. Emission images were reconstructed into 128 × 128 matrix images with a pixel size of 1.95 × 1.95 mm and a slice thickness of 4.25 mm. The mean interval period between the FDG-PET study and MRI study was 30 days (1 day–6 months).
Two neuroradiologists (T.O. and E.M.), blinded to clinical information, independently evaluated 15 patients with TLE for ATLAS using DIR and conventional MR images: T2WI and FLAIR, randomly (Morimoto et al., 2013). The diagnosis was determined in consensus of the evaluators in the case of discordant results.
The acquired images were processed using statistical parametric mapping software (SPM8; Wellcome Trust Centre for Neuroimaging, London, United Kingdom) and its extension VBM8 on the MATLAB 7.14 platform (Mathworks, Natick, MA, U.S.A.). For each patient, PET images and DIR images were coregistered onto 3D-MPRAGE images, which were spatially normalized to the Montreal Neurological Institute (MNI) space and segmented. The same spatial transformation was applied to FDG-PET and DIR images. The WM segment of MPRAGE was used to extract the WM segment of DIR images (DIR-WM). Normalized FDG-PET and DIR-WM images were smoothed with an isotropic Gaussian kernel of 8-mm full width at half maximum.
The normal volunteer groups of PET and DIR-WM and TLE patients were compared using one-way analysis of variance (ANOVA) and the chi-square test for age and sex, respectively. RTLE and LTLE were compared using the chi-square test for presence of aura or having a history of generalized tonic–clonic seizure (GTCS) and an independent t-test for duration of epilepsy.
Comparisons were conducted for hypometabolic areas in FDG-PET images and for increased intensity areas in DIR-WM images, separately at both group and individual levels for the TL and the ATL, using a two-sample t-test for each voxel by SPM8 with a threshold at p < 0.001 and the spatial extent threshold calculated according to the theory of Gaussian random fields (Luders et al., 2004). Using the same method, each and every normal subject was compared with the other normal subjects for FDG-PET and DIR-WM. In group analysis, the patients were sorted into LTLE or RTLE, and these two groups were separately compared with the control groups. Significant clusters in the TL or ATL were extracted with the masks provided using the Pickatlas (WFU Pickatlas, version 3.0.4; Maldjian et al., 2003).
Location of significant clusters was determined by referencing an atlas of the International Consortium for Brain Mapping (ICBM)-152 coordinates (Oishi et al., 2008, 2011). Statistically defined laterality was compared with the clinically defined seizure focus laterality, and the correct diagnostic rate was evaluated (Morimoto et al., 2013). When no laterality was found, the case was sorted into the false group. Agreement of seizure focus laterality between DIR-WM and FDG-PET images was assessed using Cohen κ statistics. The detectability of seizure focus laterality by visual analysis of ATLAS on DIR was calculated and compared with that by voxel-based analysis (VBA) using Cohen κ statistics on DIR-WM at ATL. Statistical analysis was conducted with MedCalc, version 12.3 (MedCalc software, Mariakerke, Belgium), except two-sample t-tests using SPM8.
No statistically significant difference in age (p = 0.67) or sex (p = 0.26) was found among the TLE patients and the two normal volunteer groups. There was no significant hypometabolic or increased signal intensity area when each normal subject was compared with the other normal subjects. Between RTLE and LTLE patients, no significant difference was observed for presence of aura (p = 0.51), having a history of GTCS (p = 0.85), and duration of epilepsy (p = 0.83). See Table 1.
Fourteen patients had mesial TLE, and only one patient had lateral TLE. ATLAS was detected visually in 4 (27%) and 10 (67%) of 15 TLE patients on conventional sequences and DIR, respectively (Table 2).
|FDG-PET||DIR-WM||T2WI FLAIR||DIR||FDG-PET||DIR-WM||T2WI FLAIR||DIR|
Significant hypometabolism was detected widely in the ipsilateral TL for LTLE (Fig. 1A). Hypometabolic areas were found in the whole areas of the ATL, except for a part of the superior temporal gyrus (STG). In the dorsal part of the TL, hypometabolic areas were observed at the middle temporal gyrus (MTG), inferior temporal gyrus (ITG), fusiform gyrus (FUG), and the hippocampus (HI). Also in the contralateral ATL, hypometabolic areas were found limited almost to the FUG.
In RTLE, significantly hypometabolic areas were primarily limited to the ipsilateral ATL: planum polare (PPo: mesial part of the STG), the entorhinal cortex, MTG, FUG, and lateral area of the HI, all of which were smaller than those in LTLE (Fig. 1B). There was also a small hypometabolic area in the contralateral MTG.
In LTLE, significant signal increase was observed in the ipsilateral TL (Fig. 1C). Increased intensity areas were found in the whole white matter of the ATL, other than a part of the STG white matter (STG-WM). MTG-WM, ITG-WM, and inferior frontooccipital fasciculus (IFOF) were also involved in the dorsal part of the ipsilateral TL. Also at the FUG-WM and ITG-WM of the contralateral ATL, small areas of increased intensity were detected.
In RTLE, significantly increased signal intensity was found limited to the ipsilateral ATL (Fig. 1D) and smaller than that in LTLE.
Significantly hypometabolic areas were found in the TL and the ATL ipsilateral to the clinically defined seizure focus laterality in 11 (73%) patients and 10 patients (67%), respectively. Hypometabolic areas were also observed contralaterally in three subjects. One of them had bilateral hypometabolism with opposite laterality, but the other two were not. Hence, in 10 (91%) of 11 and 9 (90%) of 10 patients, hypometabolic areas were larger ipsilaterally in the TL and the ATL, respectively. In 9 (90%) of 10 patients with ipsilateral hypometabolic areas, they were predominantly observed in the ipsilateral ATL. These hypometabolic areas were predominantly observed at the MTG, FUG, PPo, and ITG in seven, seven, six, and five of nine patients, respectively.
The lateral part of the STG and the entorhinal cortex were less involved. Hypometabolic areas at the lateral STG had a tendency to be observed in such patients that had large hypometabolic areas in the ipsilateral ATL. The ipsilateral HI was involved in five patients. These results are summarized in the Table 3, and overlaps of significantly hypometabolic areas are presented in Fig. S1. As for the dorsal part of the TL, hypometabolic areas were detected at the ipsilateral MTG in four patients, which were rarely observed at the ipsilateral ITG and FUG.
The one patient of lateral TLE had hypometabolic areas on FDG-PET, not in the ATL but in the dorsal lateral areas in the ipsilateral TL.
Significant signal increase in the WM was observed in 12 patients (80%) and 10 patients (67%) in the ipsilateral TL and ATL, respectively. One patient, who had contralateral laterality of hypometabolism on FDG-PET, also had significant abnormality only on the contralateral side of the ATL (case 11 in Table 2). Contralateral areas of increased intensity were also observed in the TL and ATL in eight and five patients, respectively, but they were smaller than ipsilateral areas. Therefore, significantly increased signal intensity areas were larger ipsilaterally for TL in 12 (92%) of 13 patients and ATL in 10 (91%) of 11 patients.
Ipsilateral ATL was predominantly involved at the MTG-WM, ITG-WM, PPo-WM, FUG-WM, and lateral STG-WM in 10, 10, 9, 9, and 4 patients, respectively (Table 3). Increased signal intensity areas were often observed at the dorsal part of the ipsilateral MTG-WM, but less so at the ITG-WM, FUG-WM, and IFOF. Similar to the distribution of the hypometabolic area, the one patient with lateral TLE had increased intensity areas not in the ATL but in the dorsal lateral areas in the ipsilateral TL. Distributions of significantly abnormal areas in the ATL are illustrated as overlaps of individual results for both DIR-WM and FDG-PET (Fig. S1).
When FDG-PET revealed large hypometabolic areas, increased signal intensity areas on DIR-WM were frequently large on the same side. In the ATL, increased intensity areas were often observed at the WM, corresponding to hypometabolic gray matter (GM) areas (Table 3; Fig. 2), but less so in the dorsal part of the TL (Fig. 3).
Correct diagnostic rates for DIR-WM (0.80 and 0.67 for the TL and the ATL, respectively) were slightly higher than those for FDG-PET (0.67 and 0.60 for the TL and the ATL, respectively). Agreement of seizure focus laterality between FDG-PET and DIR-WM was substantial for the TL and almost perfect for the ATL (κ = 0.67 and 0.86 for the TL and the ATL, respectively; Table S2). There were seven patients (47%) with HS on conventional MR images in this study. In the other eight patients, correct diagnostic rates of DIR-WM (0.63 and 0.38 for the TL and the ATL, respectively) were slightly higher than or equal to those of FDG-PET (0.50 and 0.38). DIR-WM could find laterality in MR-negative subjects with detectability comparable to FDG-PET (Table 2).
Seizure focus laterality was correctly detected in six and seven patients in the HS positive group (n = 7), and in four and three patients in the HS negative group (n = 8) by visual ATLAS evaluation and VBA of DIR-WM at ATL, respectively. Both of the evaluation methods had the same detectability of 10/15 (75%, κ = 0.70). Substantially the same result was attained.
Four patients had surgical resection on the clinically defined seizure focus sides, all of which were correctly detected by both DIR and FDG-PET. A good outcome of seizure control, Engel class Ia (Engel et al., 1993), was attained in all four patients after an average follow-up period of 24.5 months (range, 12–37 months).
This is the first study that has compared the detectability of DIR and FDG-PET regarding seizure focus laterality using a VBA in TLE patients. FDG-PET is widely used to define epileptogenic cortical areas, and also to guide the placement of intracranial electrodes, because of its high detection levels of seizure focus (Juhasz et al., 2000). On the other hand, MR imaging has been reported to have a lower detectability (King et al., 1998; Engel et al., 2008; Takaya et al., 2012). However, this study has successfully shown that the increased signal intensity areas on DIR-WM corresponded well to the hypometabolic areas on FDG-PET, especially in the anterior temporal lobe, and detection of seizure focus laterality was almost equivalent between DIR-WM and FDG-PET. In addition, visual analysis of ATLAS on DIR showed almost same detectability as that attained by VBA analysis of DIR-WM and FDG-PET, which objectively confirmed the high reliability of visual estimation ATLAS. It is simple and convenient for routine clinical practice, and gives MRI a more important role in image evaluation of TLE. There have been only a few reports regarding DIR images in epilepsy (Rugg-Gunn et al., 2006; Salmenpera et al., 2007; Morimoto et al., 2013). A previous study revealed that DIR achieved higher detectability of seizure focus laterality in TLE than conventional MR sequences by visual evaluation of ATLAS laterality (Morimoto et al., 2013). By a quantitative analysis, Salmenpera et al. (2007) demonstrated that 3D-DIR at 1.5T revealed signal abnormality (ipsilateral, 14%; bilateral, 7%; contralateral, 5%) in TLE patients using voxel-based whole brain analysis inclusive of both the GM and WM. In this study higher detectability was observed, which is considered attributable to an improved signal-to-noise ratio at higher magnetic strength (Wattjes et al., 2007) and the analysis focused only on the WM. Extremely high signal intensity of the GM obscures faint signal changes in the WM on quantitative analysis, because DIR nullifies signals from both the cerebrospinal fluid and WM (Turetschek et al., 1998).
Chassoux et al. (2004) reported that hypometabolic areas on FDG-PET in the patients with TLE were frequently detected in the mesial ATL and tended to spread to the lateral and dorsal parts of the TL, which is similar to what was observed in our study. They also noted that hypometabolic areas were more extended when ATLAS was detected (Chassoux et al., 2004). We found the same tendency with the observation that increased intensity areas in the WM on DIR were located in areas that corresponded to hypometabolic GM on FDG-PET, especially in the ATL (Figs. 1, 2 and S1). These results may mean that increased intensity areas in the DIR-WM show secondary changes caused by corresponding cortical abnormality; this is also suggested by the high κ values of agreement. In addition, they provide supportive evidence that evaluators should focus on the ATLAS, abnormal signal in the ATL-WM, when they want to visually determine the seizure focus laterality in the MR images of TLE patients (Kuzniecky et al., 1987; Choi et al., 1999; Meiners et al., 1999; Mitchell et al., 1999; Coste et al., 2002; Adachi et al., 2006; Morimoto et al., 2013). On the other hand, abnormal areas were less correlated in the dorsal part of the TL. Uncorrelated WM lesions would be degenerative changes like Wallerian degeneration, isolated areas vulnerable to the epileptic discharge, or incidental lesions unrelated to epilepsy.
Contralateral changes on DIR-WM and PET imaging were detected in some cases in our study as in former studies (Chassoux et al., 2004; Salmenpera et al., 2007). It might be induced in patients with TLE because of anatomic connections between bilateral TLs (Cavazos & Sutula, 1990). A patient in this study had contralateral changes and was found to have opposite laterality by both FDG-PET and DIR-WM. Two TLE patients having PET or MRI abnormal findings opposite to the depth EEG were reported (Benbadis et al., 1995). They had poor prognosis when resection was conducted on the side of image abnormality. However, such cases are relatively rare, and no reasoning was provided for causality (Benbadis et al., 1995).
Increased intensity areas on DIR-WM and hypometabolic areas on FDG-PET were larger in the LTLE group than that of RTLE, although no significant difference in the characteristics was found between them. Chassoux et al. (2004) found that VBA analysis of patient groups showed larger hypometabolic areas on FDG-PET, when RTLE images were reversed to the left than when LTLE images were flipped to the right in order to get epilepsy focus sides aligned to one side. One of the probable reasons is that the normal brain has metabolic laterality. If FDG uptake of the healthy volunteers was higher on the left side than on the right side, larger hypometabolic areas would be detected when images of the patients are analyzed by pooling the epileptogenic hemispheres on the left side. In fact, higher metabolism on the left TL than right TL was observed in the healthy brain using FDG-PET (Fujimoto et al., 2008). In addition, patients of LTLE had more prominent findings than those of RTLE for hypometabolism on FDG-PET (Kerr et al., 2013), GM volume reduction (Bonilha et al., 2007; Li et al., 2012), and abnormality of white matter fiber tracts (Kemmotsu et al., 2011). Some possible reasons have been presented. Temporal lobes are more susceptible to epileptic insults on the left (Bonilha et al., 2007; Kemmotsu et al., 2011; Li et al., 2012), and the dominant hemisphere possibly have more intensely connected to the rest of the brain and hippocampal damage may cause neuronal loss from deafferentation in a larger number of remote sites (Bonilha et al., 2007; Kerr et al., 2013). However, the precise mechanism of the asymmetrical and different intracranial damages in the LTLE and the RTLE remains unclear (Li et al., 2012).
Ipsilateral TL was affected on both FDG-PET and DIR-WM only at dorsolateral areas in a patient with lateral TLE, without significant involvement of the ATL. This result is consistent with a previous report (Bercovici et al., 2012) and may indicate difference between medial and lateral TLE patients. However, in our study, there was only this patient who had lateral TLE, and this result cannot be generalized.
In the ATL, hypometabolic areas and increased signal intensity areas were frequently observed in the MTG and PPo, which is in good agreement with a former finding that anterior parts of the STG and the MTG were usually involved in TLE patients, when measured using stereo-EEG with depth electrodes (Bartolomei et al., 1999). We also detected increased intensity areas in the IFOF in some patients. McDonald et al. (2008) reported that patients with TLE exhibited changes in fractional anisotropy and mean diffusivity in the IFOF, which were related to deteriorated verbal memory and naming performances, and predicted cognitive performances. DIR-WM not only has higher detectability of HS, cortical malformation, and other macrostructural changes than conventional MR sequences (Cotton et al., 2006; Rugg-Gunn et al., 2006; Zhang et al., 2011), but also is considered to have a potential to evaluate function in patients with TLE.
Other than DIR, diffusion-tensor imaging has advantage in detecting microstructural changes, and was used to evaluate white matter changes on the ipsilateral side to seizure foci (Kemmotsu et al., 2011). DIR reflects changes in T1 and T2 values and gives another contrast that will help to detect laterality of the seizure focus. Comparison of these two methods is out of focus of this study, but combined usage of these methods is expected to increase detectability and reliability.
Our study has some limitations. No histologic confirmation was obtained for the increased signal changes of DIR-WM. This was because some patients in this study were less severe TLE cases, so that they did not require surgical treatment in the end. The pathologic changes that correspond to ATLAS have yet to be clearly identified. Many possible causes of ATLAS have been discussed including lower myelin density (Meiners et al., 1999), gliosis (Mitchell et al., 1999; Adachi et al., 2006), corpora amylacea (Choi et al., 1999; Mitchell et al., 1999), dilated perivascular spaces (Mitchell et al., 1999), minor inflammatory changes (Mitchell et al., 1999), oligodendroglial cell clusters (Choi et al., 1999; Meiners et al., 1999; Mitchell et al., 1999), heterotopic neurons (Choi et al., 1999), and microdysgenesis (Adachi et al., 2006). However, they may not be specific for ATLAS because these findings were observed also in control specimens (Meiners et al., 1999; Mitchell et al., 1999). In addition, all patients who underwent surgical treatment had a good outcome of seizure control, which suggests a certain reliability of the clinical diagnosis in this study. Secondly, the present study included seven patients with conventional MRI-visible HS. However, in HS-negative eight patients, detection of seizure focus laterality on DIR-WM was comparable to that of FDG-PET, even though the detectability was lower. It suggests that DIR-WM and FDG-PET can provide additional information even for MR-negative patients. Although the number of patients is relatively small, this is the first study that investigated capability of DIR and found it equivalent to those of the FDG-PET. A large-scale patient study would reveal more advantages of DIR.
In conclusion, DIR is shown to have high-detection ability for seizure focus laterality in TLE using VBA. An important finding was that the laterality of increased signal intensity areas in the WM on DIR located concordantly with the hypometabolic areas on FDG-PET, especially when focused on the ATL. DIR would play an indispensable role to avoid radiation exposure, especially in children, and when FDG-PET examination is not available.
This study was partly supported by a Health and Labor Sciences Research Grant for the Comprehensive Research on Disability Health and Welfare from the Ministry of Health, Labor and Welfare, Japan to Dr. Shigetoshi Takaya. We thank Mr. Katsutoshi Murata of Siemens Japan K.K., for contributions to the optimization of scan parameters. We also thank Kiyotaka Nemoto MD, PhD (Department of Psychiatry, University of Tsukuba) for suggestions regarding the analysis using SPM8.
Dominik Paul is a Siemens Healthcare worker, and contributed to develop DIR sequences. Data acquisition and analysis were conducted by authors who have no conflicts and interest. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.