- Email: [email protected]
Ictal brain hemodynamics in the epileptic focus caused by a brain tumor using functional magnetic resonance imaging (fMRI)
Seizure 2000; 9: 585–589
doi:10.1053/seiz.2000.0464, available online at http://www.idealibrary.com on
Ictal brain hemodynamics in the epileptic focus caused by a brain tumor using functional magnetic resonance imaging (fMRI) †
†
†
†
FUMIO KUBOTA , SENICHIRO KIKUCHI , MAKOTO ITO , NOBUYOSHI SHIBATA , TAKUSHIRO † ‡ ‡ § § AKATA , AKIO TAKAHASHI , TOMIO SASAKI , NARIYUKI OYA & JUN AOKI †
Department of Neuropsychiatry, Graduate School of Medical Science, Gunma University, Gunma, Japan; ‡ Department of Neurosurgery, Graduate School of Medical Science, Gunma University, Gunma, Japan; § Department of Diagnostic Radiology and Nuclear Medicine, Graduate School of Medical Science, Gunma University, Gunma, Japan Correspondence to: Fumio Kubota, Department of Neuropsychiatry, Graduate School of Medical Science, Gunma University, 39-15, Showa-machi 3 Chome, Maebashi-shi, Gunma, 371-8511, Japan. E-mail: [email protected]
Using functional magnetic resonance imaging (fMRI) we were able to observe, in detail, ictal brain hemodynamics during epileptic seizure caused by a brain tumor. A 53-year-old man was experencing partial motor seizures of the left side of his face and neck. In a brain MR image a mass lesion was found in the subcortical area of the right frontal lobe. We found focal spikes in his right hemisphere, though dominantly in C4 and T4 regions. fMRI investigations were carried out at 1.5 T (GE Signa Horizon) using gradient-echo echo-planar neuroimaging. We were able to perform the ictal examination twice. The activated regions were focalized and clearly found only on the lateral side of the tumor base. The region was in agreement with the epileptic focus examined using an electrocorticogram (ECOG). The signal intensity in the seizure focus rapidly increased 30 seconds before the convulsion was observed. After the end of the convulsion it also took 30 seconds to restore the signal intensity to the baseline value. fMRI is a very useful tool for various studies such as the identification of the epileptic focus, the mechanism of epileptic seizure, and so on. c 2000 BEA Trading Ltd
Key words: brain tumor; epileptic focus; epileptic seizure; functional magnetic resonance imaging (fMRI); ictal hemodynamics.
INTRODUCTION Using functional magnetic resonance imaging (fMRI) the activated cortical regions can be found noninvasively by increasing the brain blood flow during the tasks1–3 . fMRI provides maps of human brain function with high spatial and temporal resolution. It is expected that it will be applied to various clinical and neurological fields4–6 . Presently SPECT and PET are used to measure brain hemodynamics7 . As a consequence of using these methods, it has become clear that blood flow during epileptic seizure increases in the epileptic focus. fMRI has better resolution than other methods used to measure brain hemodynamics, so the hemodynamics in the 1059–1311/00/080585 + 05
$35.00/0
seizure focus can be clarified in detail through its use. In this paper, we report in detail on ictal brain hemodynamics observed during epileptic seizure caused by a brain tumor, using fMRI. Moreover, we have ascertained that the activated regions in fMRI measurements correlate with the epileptic focus using ECOG.
CASE HISTORY A 53-year-old man had partial motor seizures of his left side. His seizures began at the age of 52 years, involving clonic convulsion of the left side of his face with no sensation. The seizure frequency was about once an hour. The seizure duration was 30 secc 2000 BEA Trading Ltd
586
F. Kubota et al.
Fig. 1: An fMRI activation image.
onds. Consciousness was never impaired during these events. He consulted a doctor who found abnormalities in a brain MR, and was consequently admitted to our hospital. On admission, a neurological examination revealed slight left facial never paralysis and a slight weakness in his orbicularis oris. In a brain MR a mass lesion was found in the subcortical area of the right frontal lobe (mainly in the primary motor area). Video EEG monitoring showed that spikes appeared 10 seconds before the convulsion began. The spikes gradually increased in amplitude and disappeared as soon as the convulsion finished. They spread widely in his right hemisphere, though dominantly in C4 and T4 regions. The convulsion lasted for about 30 seconds. Ictal SPECT showed an indistinct accumulation in the right frontal lobe. During the operation to remove the tumor, an electrocorticogram (ECOG) was performed. The spikes appeared in the primary motor area.
Fig. 2: A signal time course at the activated region colored red.
fMRI to observe Ictal brain hemodynamics
587
Fig. 3: Schematic diagram of the position of the electrodes in ECOG.
METHOD Informed consent was obtained for the fMRI study. MRI investigations were carried out at 1.5 T (GE Signa Horizon) using gradient-echo echo-planar neuroimaging. Foam padding was used to comfortably restrict head motion. The imaging protocol consisted of sagittal localizer images (repetition time TR = 43 milliseconds, echo time TE = 2 milliseconds, 256 × 128 matrix) and gradient-echo echo-planar images which were used for detecting susceptibility-based contract (TR = 2000 milliseconds, TE = 50 milliseconds, 256 × 128 matrix). All echo-planar images were acquired using a 7 mm slice thickness and 240 mm FOV for functional activation. Three contiguous slices were positioned to include the base of the tumor. Echoplanar images were obtained every 8 seconds, and acquisition continued for 12 minutes. From one slice, 90 images were obtained continuously. Raw image data was stored and processed off-line. Data processing was carried out using a SUN SPARK II utilizing the application Advantage Windows of Functool (GE Medical Systems). Urbit’ crosscorrelation analysis was used for 89 images excluding the first image in each slice of each trial. Foam padding was used to comfortably restrict head motion. We observed the patient while performing the fMRI
to assertain if a convulsion occurred. A clear view of the patient’s face was possible. An activation image was obtained by subtracting baseline data from data obtained during the epileptic seizure. The image data set was checked for motion by running an animated loop through the consecutive images. The activated region was determined and analysed during the period from the onset of the increased signal intensity to the end of the decreased signal intensity, though the seizure was judged by observing the convulsion.
RESULTS We were able to perform an ictal examination twice in several trials. The result of the first seizure was the same as that of the second. Figure 1 shows an activation image of the first seizure from the mid-slice. The activated regions are colored according to levels of significance. Regions with probability of Urbit’ cross-correlation analysis <0.001 are colored red. They are focalized and clearly found red only on the lateral side of the tumor base, though no activated regions appeared in other slices. The region of interest (5.4 × 5.4 × 7 mm3 ) was selected from the area of high signal intensity colored
588
red in Fig. 1. It was plotted against time with a resolution of 8 seconds, as shown in Fig. 2. The thick bar, marked ‘convulsion’, showed the period during which the convulsion was observed. The thin bar, marked ‘seizure’, by which the activated region was analysed, showed the period from the onset of the increased signal intensity to the end of the decreased signal intensity. The rapid increase of intensity was observed in connection with the convulsion. The signal intensity of the seizure focus rapidly increased 30 seconds before the convulsion was observed. After the end of the convulsion it also took 30 seconds to restore the signal intensity to the baseline value. Near the peak of the intensity curve, the convulsion occurred. The intensity increased by 7–8%. The relationship between the activated region and the epileptic focus in ECOG is shown in Fig. 3. The activated region correlates with the epileptic focus.
DISCUSSION In fMRI the data produced is unclear due to the slight motion of the subject. We were able to see no head motion during the seizure and, moreover, in animation. We did, however, employ an automated image realignment algorithm (Statistical Parametric Mapping 99) to minimize the effects of subnoxel interimage motionrelated artifacts in the brain. We were not able to find any differences between results using the Functool application alone and in combination with the Statistical Parametric Mapping 99 algorithm. Since ictal blood flow in the epileptic focus increased considerably before the convulsion was visible, we considered that the seizure had started before the convulsion. Consequently, in the analysis of the activated region we considered that the onset of the increased signal intensity marked the onset of the seizure. In addition, the end of the decreased signal intensity was considered to be the end of the seizure. The regions showing activation were those in which there was increased signal intensity during the seizure. Generally, increase of signal intensity in fMRI is thought to indicate blood flow. Using fMRI we can observe blood flow increase in cortical regions during epileptic seizures. The activated region in fMRI is focalized more clearly than in SPECT. It is a part of the primary motor area, controlling facial movement, which must be under pressure from the tumor. The region is reasonable for the clinical symptom of clonic convulsion of the face. It also correlated with the epileptic focus that was examined in ECOG. There are very few reports on ictal study of fMRI, except for two reports by Jackson et al.8 and Krings et al.9 No one has shown the relationship between ictal blood flow and the precise position of the epileptic fo-
F. Kubota et al.
cus. We have shown the precise position using ECOG. It is well known that fMRI is a blood flow test with greater spatial resolution than other blood flow tests. This study shows the activated region in the epileptic focus during the seizure focalized clearly. The region correlated with the epileptic focus, so an ictal fMRI test may identify the epileptic focus. This study demonstrates that ictal blood flow in the epileptic focus increased considerably before we were able to observe the convulsion. Ictal blood flow almost returns to the baseline value about 30 seconds after the convulsion finishes. The convulsion (the seizure which we can directly observe) occurs near the peak of the blood flow curve. We can make a detailed investigation of ictal blood flow from a point of time, because we have set up a short sampling time. We have found that the first significant increase of signal intensity preceded the onset of clinical symptoms by 30 seconds, which was also noted previously8, 9 . Video EEG monitoring showed that spikes appeared 10 seconds before the convulsion began. So, in our patient, the signal intensity in the seizure focus increased 20 seconds before the convulsion was observed. In an EEG-triggered function MRI study6 of interictal epileptiform activity, significant activation was defined several seconds after epileptiform discharges were observed in the scrap EEG. The time-lag mechanism taking place in ictal and interictal studies is not clear. The epileptiform activity often emerges in the brain considerably before we can find it in the scrap EEG. The time lag might be effected by the epileptiform activity which we cannot find in the scrap EEG. Ictal fMRI tests present difficult problems such as the data being unclear due to the slight motion of the patient during the examination and the appearance rate of epileptic seizures is usually not high enough to warrant the wait in the MR room. Therefore, many epileptic patients can not be tested using ictal fMRI. However, from the patients who can be examined, we can obtain much detailed information on ictal brain hemodynamics. fMRI is a very useful too for various studies such as identification of the epileptic focus, the mechanism of the epileptic seizure, and so on.
REFERENCES 1. Bandettini, P. A., Wong, E. C., Hinks, R. S., Tikofsky, R. S. and Hyde, J. S. Time course EPI of human brain function during task activation. Magnetic Resonance in Medicine 1992; 25: 390–397. 2. Bandettini, P. A., Jesmanowicz, A, Wong, E. C. and Hyde, J. S. Processing strategies for time-course data in functional MRI of the human brain. Magnetic Resonance in Medicine 1993; 30: 161–173. 3. Connelly, A., Jackson, G. D., Frackowiak, R. S. J., Belliveau, J. W., Vargha-Khadem, F. and Gadian, D. G. Functional map-
fMRI to observe Ictal brain hemodynamics ping of activated primary cortex with a clinical MR imaging system. Radiology 1993; 188: 125–130. 4. Rao, S. M., Binder, J. R., Bandettini, P. A. et al. Functional magnetic resonance imaging of complex human movement. Neurology 1993; 43: 2311–2318. 5. Morris, G. L. 3rd, Mueller, W. M., Yetkin, F. Z. et al. Functional magnetic resonance imaging in partial epilepsy. Epilepsia 1994; 35: 1194–1198. 6. Seeck, M., Lazeyras, F., Michel, C. M. et al. Non-invasive epileptic focus localization using EEG-triggered function
589 MRI and electromagnetic tomography. Electroencephalography and Clinical Neurophysiology 1998; 106: 508–512. 7. SPECT and PET in epilepsy (editorial). Lancet 1989; 1: 135– 137. 8. Jackson, G. D., Connelly, A., Cross, J. H., Gordon, I. and Gadin, D. J. Functional magnetic resonance imaging of focal seizure. Neurology 1994; 44: 850–856. 9. Kings, T., Topper, R. and Reinges, M. H. T. et al. Hemodynamic changes in simple partial epilepsy: A functional MRI study. Neurology 2000; 54: 524–527.
doi:10.1053/seiz.2000.0464, available online at http://www.idealibrary.com on
Ictal brain hemodynamics in the epileptic focus caused by a brain tumor using functional magnetic resonance imaging (fMRI) †
†
†
†
FUMIO KUBOTA , SENICHIRO KIKUCHI , MAKOTO ITO , NOBUYOSHI SHIBATA , TAKUSHIRO † ‡ ‡ § § AKATA , AKIO TAKAHASHI , TOMIO SASAKI , NARIYUKI OYA & JUN AOKI †
Department of Neuropsychiatry, Graduate School of Medical Science, Gunma University, Gunma, Japan; ‡ Department of Neurosurgery, Graduate School of Medical Science, Gunma University, Gunma, Japan; § Department of Diagnostic Radiology and Nuclear Medicine, Graduate School of Medical Science, Gunma University, Gunma, Japan Correspondence to: Fumio Kubota, Department of Neuropsychiatry, Graduate School of Medical Science, Gunma University, 39-15, Showa-machi 3 Chome, Maebashi-shi, Gunma, 371-8511, Japan. E-mail: [email protected]
Using functional magnetic resonance imaging (fMRI) we were able to observe, in detail, ictal brain hemodynamics during epileptic seizure caused by a brain tumor. A 53-year-old man was experencing partial motor seizures of the left side of his face and neck. In a brain MR image a mass lesion was found in the subcortical area of the right frontal lobe. We found focal spikes in his right hemisphere, though dominantly in C4 and T4 regions. fMRI investigations were carried out at 1.5 T (GE Signa Horizon) using gradient-echo echo-planar neuroimaging. We were able to perform the ictal examination twice. The activated regions were focalized and clearly found only on the lateral side of the tumor base. The region was in agreement with the epileptic focus examined using an electrocorticogram (ECOG). The signal intensity in the seizure focus rapidly increased 30 seconds before the convulsion was observed. After the end of the convulsion it also took 30 seconds to restore the signal intensity to the baseline value. fMRI is a very useful tool for various studies such as the identification of the epileptic focus, the mechanism of epileptic seizure, and so on. c 2000 BEA Trading Ltd
Key words: brain tumor; epileptic focus; epileptic seizure; functional magnetic resonance imaging (fMRI); ictal hemodynamics.
INTRODUCTION Using functional magnetic resonance imaging (fMRI) the activated cortical regions can be found noninvasively by increasing the brain blood flow during the tasks1–3 . fMRI provides maps of human brain function with high spatial and temporal resolution. It is expected that it will be applied to various clinical and neurological fields4–6 . Presently SPECT and PET are used to measure brain hemodynamics7 . As a consequence of using these methods, it has become clear that blood flow during epileptic seizure increases in the epileptic focus. fMRI has better resolution than other methods used to measure brain hemodynamics, so the hemodynamics in the 1059–1311/00/080585 + 05
$35.00/0
seizure focus can be clarified in detail through its use. In this paper, we report in detail on ictal brain hemodynamics observed during epileptic seizure caused by a brain tumor, using fMRI. Moreover, we have ascertained that the activated regions in fMRI measurements correlate with the epileptic focus using ECOG.
CASE HISTORY A 53-year-old man had partial motor seizures of his left side. His seizures began at the age of 52 years, involving clonic convulsion of the left side of his face with no sensation. The seizure frequency was about once an hour. The seizure duration was 30 secc 2000 BEA Trading Ltd
586
F. Kubota et al.
Fig. 1: An fMRI activation image.
onds. Consciousness was never impaired during these events. He consulted a doctor who found abnormalities in a brain MR, and was consequently admitted to our hospital. On admission, a neurological examination revealed slight left facial never paralysis and a slight weakness in his orbicularis oris. In a brain MR a mass lesion was found in the subcortical area of the right frontal lobe (mainly in the primary motor area). Video EEG monitoring showed that spikes appeared 10 seconds before the convulsion began. The spikes gradually increased in amplitude and disappeared as soon as the convulsion finished. They spread widely in his right hemisphere, though dominantly in C4 and T4 regions. The convulsion lasted for about 30 seconds. Ictal SPECT showed an indistinct accumulation in the right frontal lobe. During the operation to remove the tumor, an electrocorticogram (ECOG) was performed. The spikes appeared in the primary motor area.
Fig. 2: A signal time course at the activated region colored red.
fMRI to observe Ictal brain hemodynamics
587
Fig. 3: Schematic diagram of the position of the electrodes in ECOG.
METHOD Informed consent was obtained for the fMRI study. MRI investigations were carried out at 1.5 T (GE Signa Horizon) using gradient-echo echo-planar neuroimaging. Foam padding was used to comfortably restrict head motion. The imaging protocol consisted of sagittal localizer images (repetition time TR = 43 milliseconds, echo time TE = 2 milliseconds, 256 × 128 matrix) and gradient-echo echo-planar images which were used for detecting susceptibility-based contract (TR = 2000 milliseconds, TE = 50 milliseconds, 256 × 128 matrix). All echo-planar images were acquired using a 7 mm slice thickness and 240 mm FOV for functional activation. Three contiguous slices were positioned to include the base of the tumor. Echoplanar images were obtained every 8 seconds, and acquisition continued for 12 minutes. From one slice, 90 images were obtained continuously. Raw image data was stored and processed off-line. Data processing was carried out using a SUN SPARK II utilizing the application Advantage Windows of Functool (GE Medical Systems). Urbit’ crosscorrelation analysis was used for 89 images excluding the first image in each slice of each trial. Foam padding was used to comfortably restrict head motion. We observed the patient while performing the fMRI
to assertain if a convulsion occurred. A clear view of the patient’s face was possible. An activation image was obtained by subtracting baseline data from data obtained during the epileptic seizure. The image data set was checked for motion by running an animated loop through the consecutive images. The activated region was determined and analysed during the period from the onset of the increased signal intensity to the end of the decreased signal intensity, though the seizure was judged by observing the convulsion.
RESULTS We were able to perform an ictal examination twice in several trials. The result of the first seizure was the same as that of the second. Figure 1 shows an activation image of the first seizure from the mid-slice. The activated regions are colored according to levels of significance. Regions with probability of Urbit’ cross-correlation analysis <0.001 are colored red. They are focalized and clearly found red only on the lateral side of the tumor base, though no activated regions appeared in other slices. The region of interest (5.4 × 5.4 × 7 mm3 ) was selected from the area of high signal intensity colored
588
red in Fig. 1. It was plotted against time with a resolution of 8 seconds, as shown in Fig. 2. The thick bar, marked ‘convulsion’, showed the period during which the convulsion was observed. The thin bar, marked ‘seizure’, by which the activated region was analysed, showed the period from the onset of the increased signal intensity to the end of the decreased signal intensity. The rapid increase of intensity was observed in connection with the convulsion. The signal intensity of the seizure focus rapidly increased 30 seconds before the convulsion was observed. After the end of the convulsion it also took 30 seconds to restore the signal intensity to the baseline value. Near the peak of the intensity curve, the convulsion occurred. The intensity increased by 7–8%. The relationship between the activated region and the epileptic focus in ECOG is shown in Fig. 3. The activated region correlates with the epileptic focus.
DISCUSSION In fMRI the data produced is unclear due to the slight motion of the subject. We were able to see no head motion during the seizure and, moreover, in animation. We did, however, employ an automated image realignment algorithm (Statistical Parametric Mapping 99) to minimize the effects of subnoxel interimage motionrelated artifacts in the brain. We were not able to find any differences between results using the Functool application alone and in combination with the Statistical Parametric Mapping 99 algorithm. Since ictal blood flow in the epileptic focus increased considerably before the convulsion was visible, we considered that the seizure had started before the convulsion. Consequently, in the analysis of the activated region we considered that the onset of the increased signal intensity marked the onset of the seizure. In addition, the end of the decreased signal intensity was considered to be the end of the seizure. The regions showing activation were those in which there was increased signal intensity during the seizure. Generally, increase of signal intensity in fMRI is thought to indicate blood flow. Using fMRI we can observe blood flow increase in cortical regions during epileptic seizures. The activated region in fMRI is focalized more clearly than in SPECT. It is a part of the primary motor area, controlling facial movement, which must be under pressure from the tumor. The region is reasonable for the clinical symptom of clonic convulsion of the face. It also correlated with the epileptic focus that was examined in ECOG. There are very few reports on ictal study of fMRI, except for two reports by Jackson et al.8 and Krings et al.9 No one has shown the relationship between ictal blood flow and the precise position of the epileptic fo-
F. Kubota et al.
cus. We have shown the precise position using ECOG. It is well known that fMRI is a blood flow test with greater spatial resolution than other blood flow tests. This study shows the activated region in the epileptic focus during the seizure focalized clearly. The region correlated with the epileptic focus, so an ictal fMRI test may identify the epileptic focus. This study demonstrates that ictal blood flow in the epileptic focus increased considerably before we were able to observe the convulsion. Ictal blood flow almost returns to the baseline value about 30 seconds after the convulsion finishes. The convulsion (the seizure which we can directly observe) occurs near the peak of the blood flow curve. We can make a detailed investigation of ictal blood flow from a point of time, because we have set up a short sampling time. We have found that the first significant increase of signal intensity preceded the onset of clinical symptoms by 30 seconds, which was also noted previously8, 9 . Video EEG monitoring showed that spikes appeared 10 seconds before the convulsion began. So, in our patient, the signal intensity in the seizure focus increased 20 seconds before the convulsion was observed. In an EEG-triggered function MRI study6 of interictal epileptiform activity, significant activation was defined several seconds after epileptiform discharges were observed in the scrap EEG. The time-lag mechanism taking place in ictal and interictal studies is not clear. The epileptiform activity often emerges in the brain considerably before we can find it in the scrap EEG. The time lag might be effected by the epileptiform activity which we cannot find in the scrap EEG. Ictal fMRI tests present difficult problems such as the data being unclear due to the slight motion of the patient during the examination and the appearance rate of epileptic seizures is usually not high enough to warrant the wait in the MR room. Therefore, many epileptic patients can not be tested using ictal fMRI. However, from the patients who can be examined, we can obtain much detailed information on ictal brain hemodynamics. fMRI is a very useful too for various studies such as identification of the epileptic focus, the mechanism of the epileptic seizure, and so on.
REFERENCES 1. Bandettini, P. A., Wong, E. C., Hinks, R. S., Tikofsky, R. S. and Hyde, J. S. Time course EPI of human brain function during task activation. Magnetic Resonance in Medicine 1992; 25: 390–397. 2. Bandettini, P. A., Jesmanowicz, A, Wong, E. C. and Hyde, J. S. Processing strategies for time-course data in functional MRI of the human brain. Magnetic Resonance in Medicine 1993; 30: 161–173. 3. Connelly, A., Jackson, G. D., Frackowiak, R. S. J., Belliveau, J. W., Vargha-Khadem, F. and Gadian, D. G. Functional map-
fMRI to observe Ictal brain hemodynamics ping of activated primary cortex with a clinical MR imaging system. Radiology 1993; 188: 125–130. 4. Rao, S. M., Binder, J. R., Bandettini, P. A. et al. Functional magnetic resonance imaging of complex human movement. Neurology 1993; 43: 2311–2318. 5. Morris, G. L. 3rd, Mueller, W. M., Yetkin, F. Z. et al. Functional magnetic resonance imaging in partial epilepsy. Epilepsia 1994; 35: 1194–1198. 6. Seeck, M., Lazeyras, F., Michel, C. M. et al. Non-invasive epileptic focus localization using EEG-triggered function
589 MRI and electromagnetic tomography. Electroencephalography and Clinical Neurophysiology 1998; 106: 508–512. 7. SPECT and PET in epilepsy (editorial). Lancet 1989; 1: 135– 137. 8. Jackson, G. D., Connelly, A., Cross, J. H., Gordon, I. and Gadin, D. J. Functional magnetic resonance imaging of focal seizure. Neurology 1994; 44: 850–856. 9. Kings, T., Topper, R. and Reinges, M. H. T. et al. Hemodynamic changes in simple partial epilepsy: A functional MRI study. Neurology 2000; 54: 524–527.