- Email: [email protected]
PET does not eliminate need for extraoperative, intracranial monitoring in pediatric epilepsy surgery
Controversial Issues
PET Does Not Eliminate Need for Extraoperative, Intracranial Monitoring in Pediatric Epilepsy Surgery O. Carter Snead, HI, MD* and Marvin D. Nelson, Jr, MD+
From the Departments of *Neurology and *Radiology; University of Southern California School of Medicine; Divisions of *Neurology and tRadiology; Children's Hospital Los Angeles; Los Angeles, California.
The broad term, epilepsy surgery, when applied to children refers to either corpus callosotomy, hemispherectomy, or focal corticectomy. This report discusses the role of positron emission tomography (PET) with [18F]deoxyglucose (FDG) for planning surgery in children undergoing either hemispherectomy or focal corticectomy for uncontrolled seizures. Since the pioneering work by Goldring [1] and Goldring and Gregorie [2], epilepsy surgery has gained steady acceptance among the Neurology and Pediatric communities as a viable therapeutic modality for medically intractable epilepsy in children. Most centers that perform pediatric epilepsy surgery evaluations in children now agree that candidates for this procedure should be subjected to a rigorous, standardized, presurgical evaluation process to select those who stand the best chance of benefitting from epilepsy surgery and thus reduce the risk of operating unnecessarily. Typically, the evaluation process goes from the least to the most invasive. At our institution, phase one of this process consists of history, physical, antiepileptic drug profile, psychosocial evaluation, neuropsychologic evaluation, speech and hearing, psychiatric evaluation, electroencephalography (EEG) video recording to capture several typical seizures, enhanced and unenhanced computed tomography (CT) and magnetic resonance imaging (MRI), and when possible, FDG-PET. The latter, although helpful, is quite expensive and often not authorized by third-party payors; theretore, it is not always possible to obtain this study. At the end of phase one, the decision is made by the epilepsy surgery team whether enough evidence of potential benefit to the patient exists to proceed. Phase two consists of intracarotid amytal or Wada testing to determine cerebral dominance, neuroophthalmologic evaluation, and somatosensory evoked responses. During phase three, customdesigned subdural electrode grids are placed in order to determine the epileptogenic zone via extraoperative ictal recordings. In addition, eloquent areas of neurologic function are determined by stimulation through the grid. Phase
four consists of removal of the grids followed by focal corticectomy or hemispherectomy as indicated by all test results. This process of patient selection, as well as the epilepsy surgery itself, is expensive and labor intensive, but logical and data driven with the ultimate potential to provide dramatic improvement in seizure control in children who are uniformly devastated by epilepsy. The question before us is simple: Can one eliminate the need for ictal recordings from subdural grids (i.e., phase three testing) when one has an area of hypometabolism defined by FDG-PET, the surgical removal of which results in the same or better outcome as if phase three testing had been performed? If the answer to this question were yes, then by definition the epileptogenic zone, that is the area of seizure origin, which we currently define by ictal recordings from subdural grids, must be identical in anatomic boundaries to the area of hypometabolism observed on FDG-PET. If this finding were true then only interictal electrocorticography would need to be performed under anesthesia to localize the epileptogenic zone at the time of focal corticectomy/hemispherectomy. Our answer to this question is no. A "yes" answer assumes that the neurons in the epileptogenic zone always have the same functional deficit (i.e., decreased glucose utilization); however, no evidence exists to support this premise. Each diagnostic procedure performed during presurgical evaluation is designed to provide data concerning disparate areas of brain function. For example, each neuroimaging method provides different information based on different physical properties of brain tissue. CT measures densities of cerebral tissue by attenuation of an X-ray beam passed through the cranium. Standard MRI is based on the signal intensities given off by resonating water molecules. The signal intensity depends on both the amount of water present and how the water molecules interact with their surrounding environment. FDG-PET is based on cellular uptake and metabolism of glucose. Thus, PET provides a measure of only one aspect of neuronal metabolism which compliments, but does not replace CT and MRI. Similarly, FDG-PET visualizes properties of brain that are different from those elicited during ictal electrocorticography from subdural grids; therefore, it is fallacious to conclude that the area of abnormality of cellular metabolic function, as defined by FDG-PET, al-
PEDIATRIC NEUROLOGY
Vol. 9 No. 5
409
ways correlates exactly with the anatomic boundaries of the paroxysmally discharging neurons that comprise the epileptogenic zone. Although little doubt exists that the epileptogenic focus may present on the interictal FDG-PET scan as a hypometabolic area ipsilateral to the epileptogenic zone [3-6], many authors have documented that PET is not foolproof. Swartz et al. reported a patient in whom the epileptogenic zone was contralateral to the observed PET abnormality [7]. Engel et al. reported 2 patients with an abnormal PET on one side, but spikes recorded from a contralateral depth electrode [8]. Both patients were reportedly seizure free after the epileptogenic z o n e - as recorded by EEG, not the PET area of hypometabolism-was removed. Abou-Khalil et al. also reported a focal PET abnormality contralateral to temporal spikes [9]. Conversely, PET may be normal in patients with unilateral hippocampal onset of seizures [5,8]. Moreover, unilateral temporal lobe hypometabolism does not exclude bilateral-independenthippocampal-onset seizures and the absence of bilateral PET abnormalities does not exclude bilateral, independent epileptogenic zones [10]. Levesque et al. demonstrated that FDG-PET visualizes false lateralization in 2% and false localization in 6% of patients [tl]. These authors concluded that FDG-PET did not contribute significantly to decision-making in presurgical evaluation in 42% of patients. Theodore et al. recently reported the role of FDG-PET in a series of adults undergoing temporal lobectomy for intractable partial complex seizures [6]. These authors concluded that FDG-PET provided little additional information when a clear temporal ictal surface EEG focus was present, but was valuable when the surface recording was nonlocalizing. They also opined that quantitative measurements of asymmetry should be made rather than relying only upon visual analysis [6]. Finally, it should be noted that no one has apparently conducted sequential FDG-PET scans in the same patients in order to determine whether the defined area of hypometabolism is stable in terms of the anatomic borders. Most of the literature cited above deals with adult patients undergoing presurgical evaluation for temporal lobectomy; however, in our experience, most children with epilepsy who undergo surgical evaluation have extratemporal foci. The adult data also suggest that patients with extratemporal foci (e.g., in frontal lobe) may also have areas of hypometabolism on FDG-PET; these hypometabolic areas tend to be less frequent and more diffuse than in patients with temporal lobe foci [12]. There are few series of FDG-PET studies in children undergoing evaluation for epilepsy surgery for intractable partial seizures. During the past 2 years we obtained FDG-PET scans in 31 of 53 children who were evaluated for seizure surgery [13]. Twenty-two patients did not have the financial resources to obtain the study. FDG-PET was useful in 25 of 31 patients in making a surgical decision regarding focal corticectomy or hemispherectomy, determining the indi-
4/0
PEDIATRIC NEUROLOGY
Vol. 9 No. 5
cation tbr extraoperative subdural grid recordings, or cx cluding surgery, but was misleading or not helpful iu 6 patients. In 4 of these children FDG-PET was falsely localizing when compared to the ictal electrocorticographic monitoring and in two the FDG-PET was normal in the presence of a structural lesion which was also not visualized by CT and MRI, but which was ultimately discovered by ictal monitoring from subdural grids I141. In summary, our experience in children agrees with that reported for adults in that we found FDG-PET 1o be useful in the preoperative evaluation of the majority of pediatric epilepsy surgery candidates. PET appears to be most useful in two situations. First, when there is absolute agreement between the focal area of hypometabolism, a radiographic lesion, video-EEG, and clinical data, subdural ictal recordings are not indicated. Second, PET may be useful in guiding the use and placement of subdural grids when it demonstrates a single, local area of hypometabolism in a child in whom EEG does not provide localizing or even lateralizing information and when CT and MRI are normal, or when the area of hypometabolism exceeds or is different from that indicated by CT or MRI. PET is most often not helpful when it is normal or reveals multiple or diffuse areas of hypometabolism. In those patients, PET cannot be used alone to dictate surgical decisions. Therefore, we believe that PET is valuable, but only when considered with all other data upon which surgical decisions are made. It is useful only when taken in context with clinical, electrographic, and other radiologic data. FDG-PET obviates the need for ictal subdural recordings only when it provides localizing findings concordant with all others.
References
[l] Goldring S. A method for surgical management of focal epilepsy, especially as it relates to children. J Neurosurg 1978;49:344-56. [2] Goldring S, Gregorie EM. Surgical management of epilepsy using epidural recordings to localize the seizure focus. J Neurosurg 1984;60:457-66. [3] Kuhl DE, Engel J, Phelps ME, Selin C. Epileptic pattern,,, of local cerebral metabolism and perfusion in humans determined by emission computed tomography of tSFDG and ~3NH3. Ann Neurol 1980;8: 348-60. [4] Engel J Jr, Kuhl DE, Phelps ME, Mazziotta JC. lnterictal cerebral glucose metabolism in partial epilepsy and its relation to EEG changes. Ann Neurol 1982;12:510-7. [5] Theodore WH, Dorwart R, Holmes M, Porter RJ. DiChiro G. Neuroimaging in refractory partial seizures: Comparison of PET, CT, and MRI. Neurology 1986;36:750-9. [6] Theodore WH, Sato S, Kufta C, Balish MB, Bromfield EB. Leiderman DB. Temporal lobectomy for uncontrolled seizures: The role of positron emission tomography. Ann Neurol 1992;32:789-94. [7] Swartz BE, Tomiuasu U, Delgado-Escueta AV, Mandelkern M, Khonsari A. Neuroimaging in temporal lobe epilepsy: Test sensitivity and relationships to pathology and postoperative outcome. Epilepsia 1992;33:624-34. [81 Engel J Jr, Henry TR, Risinger MW, et al. Presurgical evaluation for partial epilepsy: Relative contributions of chronic depth-electrode recordings versus FDG-PET and scalp sphenoidal ictal EEG. Neurology 1990;40:1670-7.
[9] Abou-Khalil BW, Siegel GJ, Sackerllares C, Gilman S, Hichwa R, Marshall R. Positron emission tomography: Studies of cerebral glucose metabolism in chronic partial epilepsy. Ann Neurol 1987;22:480-6. [10] Labar D, Fraser R, Schaul N, Dhawan V, Eidelberg D. Wrong side PET lateralization in temporal lobe epilepsy. Epilepsia 1992;33 (suppl 3):48. [Ill Levesque M, Harkness W, Sutherling W, et al. SEEG localization in epileptic patients who fail localization after noninvasive studies. Epilepsia 1992;(suppl 3):90-1.
[12] Swartz BE, Halgren E, Delgado-Escueta AV, et al. Neuroimaging in patients with seizures of probable frontal lobe origin. Epilepsia 1989;30:547-58. [13] Snead OC, Kongelbeck S, Mitchell WG. Positron emission tomography in the evaluation of pediatric epilepsy surgery candidates. Epilepsia 1992;33(suppl 3):92. [14] Mitchell WG, Snead OC, Raffel C, Gilles F, Kongelbeck SR. Radiologically undetected tubers causing intractable partial seizures in children. Epilepsia 1992;33(suppl 3):96.
PET in Preoperative Evaluation of Intractable Epilepsy Harry T. Chugani, MD
From the Departments of Pediatric Neurology, and Radiology; Children's Hospital of Michigan and Wayne State University; Detroit, Michigan.
Surgical treatment of medically-refractory epilepsy in children has gained considerable popularity in recent years. As a result, pediatric epilepsy surgery programs have been established worldwide. Surgical procedures being performed include temporal lobectomy, extratemporal cortical resection, hemispherectomy, and corpus callosotomy. Although most of these procedures had been performed in a small number of epilepsy surgery centers for several decades, recent interest has focused more on infants and young children because of improved neurosurgical techniques and intensive care, increasing recognition and appreciation of considerable plasticity in the developing brain, and the desire to avoid progressive epileptic encephalopathy. Despite the belief of most pediatric neurologists and epileptologists that epilepsy surgery is effective in selected infants and children, there appears to be no consensus as to a standard preoperative assessment protocol; however, most would agree that following the confirmation of medical intractability of the seizures, EEG localization (or lateralization) of seizure onset should be correlated with localization from seizure semiology, clinical examination, and neuroimaging modalities, thereby providing enough convergence of data to indicate the epileptogenic region for resection. Epilepsy surgery owes its success to the identification of discrete epileptogenic foci that at least lateralize to one hemisphere. The clinical and electrographic features of seizures arising from a single focus in children often cannot be differentiated from seizures that arise from multiple bilateral and diffuse foci. In the absence of an MRI
lesion, interictal and ictal scalp EEG localization is insufficient to permit proceeding to surgery, and must be confirmed by chronic intracranial EEG monitoring, such as depth or subdural electrode recordings. The development of PET technology has enabled brain functional activity to be imaged noninvasively [1]. Although it can be used to image and acquire quantitative measurements of a wide variety of biochemical processes in brain, the clinical application of PET in pediatric epilepsy surgery has been limited to studies of cerebral glucose utilization with the tracer FDG. Even so, FDG-PET has proved to be extremely useful in infants and children who have intractable epilepsy without an MRI lesion. We found that in such instances, PET contributes significantly to the surgical decision-making process. By fulfilling a role similar to that of MRI when the latter fails to detect a lesion, PET provides some guide as to the type of resection to be performed. Furthermore, PET can provide an assessment of the functional integrity of brain regions outside the epileptogenic area. When data from PET are incorporated in such a manner into the presurgical evaluation, chronic, invasive electrographic monitoring can be avoided in the majority of infants and children. In a recent analysis of 84 infants and children undergoing surgery for intractable epilepsy at the University of California, Los Angeles [2], only 7 (i.e., 8.3%) required chronic, invasive monitoring as a result of the incorporation of PET into the surgical decision-making process. This less invasive (and less expensive) approach is particularly important in children because of the technical difficulties and morbidity associated with chronic intracranial monitoring [3]. The surgical procedures in our series at analysis consisted of hemispherectomy (n = 27), temporal lobectomy (n = 24), extended cortical resection (n = 26), and corpus callosotomy (n = 7). Postoperatively, seizures were eliminated or > 90% reduced in 75% of patients. These results are about
PEDIATRIC NEUROLOGY
Vol. 9 No. 5 411
PET Does Not Eliminate Need for Extraoperative, Intracranial Monitoring in Pediatric Epilepsy Surgery O. Carter Snead, HI, MD* and Marvin D. Nelson, Jr, MD+
From the Departments of *Neurology and *Radiology; University of Southern California School of Medicine; Divisions of *Neurology and tRadiology; Children's Hospital Los Angeles; Los Angeles, California.
The broad term, epilepsy surgery, when applied to children refers to either corpus callosotomy, hemispherectomy, or focal corticectomy. This report discusses the role of positron emission tomography (PET) with [18F]deoxyglucose (FDG) for planning surgery in children undergoing either hemispherectomy or focal corticectomy for uncontrolled seizures. Since the pioneering work by Goldring [1] and Goldring and Gregorie [2], epilepsy surgery has gained steady acceptance among the Neurology and Pediatric communities as a viable therapeutic modality for medically intractable epilepsy in children. Most centers that perform pediatric epilepsy surgery evaluations in children now agree that candidates for this procedure should be subjected to a rigorous, standardized, presurgical evaluation process to select those who stand the best chance of benefitting from epilepsy surgery and thus reduce the risk of operating unnecessarily. Typically, the evaluation process goes from the least to the most invasive. At our institution, phase one of this process consists of history, physical, antiepileptic drug profile, psychosocial evaluation, neuropsychologic evaluation, speech and hearing, psychiatric evaluation, electroencephalography (EEG) video recording to capture several typical seizures, enhanced and unenhanced computed tomography (CT) and magnetic resonance imaging (MRI), and when possible, FDG-PET. The latter, although helpful, is quite expensive and often not authorized by third-party payors; theretore, it is not always possible to obtain this study. At the end of phase one, the decision is made by the epilepsy surgery team whether enough evidence of potential benefit to the patient exists to proceed. Phase two consists of intracarotid amytal or Wada testing to determine cerebral dominance, neuroophthalmologic evaluation, and somatosensory evoked responses. During phase three, customdesigned subdural electrode grids are placed in order to determine the epileptogenic zone via extraoperative ictal recordings. In addition, eloquent areas of neurologic function are determined by stimulation through the grid. Phase
four consists of removal of the grids followed by focal corticectomy or hemispherectomy as indicated by all test results. This process of patient selection, as well as the epilepsy surgery itself, is expensive and labor intensive, but logical and data driven with the ultimate potential to provide dramatic improvement in seizure control in children who are uniformly devastated by epilepsy. The question before us is simple: Can one eliminate the need for ictal recordings from subdural grids (i.e., phase three testing) when one has an area of hypometabolism defined by FDG-PET, the surgical removal of which results in the same or better outcome as if phase three testing had been performed? If the answer to this question were yes, then by definition the epileptogenic zone, that is the area of seizure origin, which we currently define by ictal recordings from subdural grids, must be identical in anatomic boundaries to the area of hypometabolism observed on FDG-PET. If this finding were true then only interictal electrocorticography would need to be performed under anesthesia to localize the epileptogenic zone at the time of focal corticectomy/hemispherectomy. Our answer to this question is no. A "yes" answer assumes that the neurons in the epileptogenic zone always have the same functional deficit (i.e., decreased glucose utilization); however, no evidence exists to support this premise. Each diagnostic procedure performed during presurgical evaluation is designed to provide data concerning disparate areas of brain function. For example, each neuroimaging method provides different information based on different physical properties of brain tissue. CT measures densities of cerebral tissue by attenuation of an X-ray beam passed through the cranium. Standard MRI is based on the signal intensities given off by resonating water molecules. The signal intensity depends on both the amount of water present and how the water molecules interact with their surrounding environment. FDG-PET is based on cellular uptake and metabolism of glucose. Thus, PET provides a measure of only one aspect of neuronal metabolism which compliments, but does not replace CT and MRI. Similarly, FDG-PET visualizes properties of brain that are different from those elicited during ictal electrocorticography from subdural grids; therefore, it is fallacious to conclude that the area of abnormality of cellular metabolic function, as defined by FDG-PET, al-
PEDIATRIC NEUROLOGY
Vol. 9 No. 5
409
ways correlates exactly with the anatomic boundaries of the paroxysmally discharging neurons that comprise the epileptogenic zone. Although little doubt exists that the epileptogenic focus may present on the interictal FDG-PET scan as a hypometabolic area ipsilateral to the epileptogenic zone [3-6], many authors have documented that PET is not foolproof. Swartz et al. reported a patient in whom the epileptogenic zone was contralateral to the observed PET abnormality [7]. Engel et al. reported 2 patients with an abnormal PET on one side, but spikes recorded from a contralateral depth electrode [8]. Both patients were reportedly seizure free after the epileptogenic z o n e - as recorded by EEG, not the PET area of hypometabolism-was removed. Abou-Khalil et al. also reported a focal PET abnormality contralateral to temporal spikes [9]. Conversely, PET may be normal in patients with unilateral hippocampal onset of seizures [5,8]. Moreover, unilateral temporal lobe hypometabolism does not exclude bilateral-independenthippocampal-onset seizures and the absence of bilateral PET abnormalities does not exclude bilateral, independent epileptogenic zones [10]. Levesque et al. demonstrated that FDG-PET visualizes false lateralization in 2% and false localization in 6% of patients [tl]. These authors concluded that FDG-PET did not contribute significantly to decision-making in presurgical evaluation in 42% of patients. Theodore et al. recently reported the role of FDG-PET in a series of adults undergoing temporal lobectomy for intractable partial complex seizures [6]. These authors concluded that FDG-PET provided little additional information when a clear temporal ictal surface EEG focus was present, but was valuable when the surface recording was nonlocalizing. They also opined that quantitative measurements of asymmetry should be made rather than relying only upon visual analysis [6]. Finally, it should be noted that no one has apparently conducted sequential FDG-PET scans in the same patients in order to determine whether the defined area of hypometabolism is stable in terms of the anatomic borders. Most of the literature cited above deals with adult patients undergoing presurgical evaluation for temporal lobectomy; however, in our experience, most children with epilepsy who undergo surgical evaluation have extratemporal foci. The adult data also suggest that patients with extratemporal foci (e.g., in frontal lobe) may also have areas of hypometabolism on FDG-PET; these hypometabolic areas tend to be less frequent and more diffuse than in patients with temporal lobe foci [12]. There are few series of FDG-PET studies in children undergoing evaluation for epilepsy surgery for intractable partial seizures. During the past 2 years we obtained FDG-PET scans in 31 of 53 children who were evaluated for seizure surgery [13]. Twenty-two patients did not have the financial resources to obtain the study. FDG-PET was useful in 25 of 31 patients in making a surgical decision regarding focal corticectomy or hemispherectomy, determining the indi-
4/0
PEDIATRIC NEUROLOGY
Vol. 9 No. 5
cation tbr extraoperative subdural grid recordings, or cx cluding surgery, but was misleading or not helpful iu 6 patients. In 4 of these children FDG-PET was falsely localizing when compared to the ictal electrocorticographic monitoring and in two the FDG-PET was normal in the presence of a structural lesion which was also not visualized by CT and MRI, but which was ultimately discovered by ictal monitoring from subdural grids I141. In summary, our experience in children agrees with that reported for adults in that we found FDG-PET 1o be useful in the preoperative evaluation of the majority of pediatric epilepsy surgery candidates. PET appears to be most useful in two situations. First, when there is absolute agreement between the focal area of hypometabolism, a radiographic lesion, video-EEG, and clinical data, subdural ictal recordings are not indicated. Second, PET may be useful in guiding the use and placement of subdural grids when it demonstrates a single, local area of hypometabolism in a child in whom EEG does not provide localizing or even lateralizing information and when CT and MRI are normal, or when the area of hypometabolism exceeds or is different from that indicated by CT or MRI. PET is most often not helpful when it is normal or reveals multiple or diffuse areas of hypometabolism. In those patients, PET cannot be used alone to dictate surgical decisions. Therefore, we believe that PET is valuable, but only when considered with all other data upon which surgical decisions are made. It is useful only when taken in context with clinical, electrographic, and other radiologic data. FDG-PET obviates the need for ictal subdural recordings only when it provides localizing findings concordant with all others.
References
[l] Goldring S. A method for surgical management of focal epilepsy, especially as it relates to children. J Neurosurg 1978;49:344-56. [2] Goldring S, Gregorie EM. Surgical management of epilepsy using epidural recordings to localize the seizure focus. J Neurosurg 1984;60:457-66. [3] Kuhl DE, Engel J, Phelps ME, Selin C. Epileptic pattern,,, of local cerebral metabolism and perfusion in humans determined by emission computed tomography of tSFDG and ~3NH3. Ann Neurol 1980;8: 348-60. [4] Engel J Jr, Kuhl DE, Phelps ME, Mazziotta JC. lnterictal cerebral glucose metabolism in partial epilepsy and its relation to EEG changes. Ann Neurol 1982;12:510-7. [5] Theodore WH, Dorwart R, Holmes M, Porter RJ. DiChiro G. Neuroimaging in refractory partial seizures: Comparison of PET, CT, and MRI. Neurology 1986;36:750-9. [6] Theodore WH, Sato S, Kufta C, Balish MB, Bromfield EB. Leiderman DB. Temporal lobectomy for uncontrolled seizures: The role of positron emission tomography. Ann Neurol 1992;32:789-94. [7] Swartz BE, Tomiuasu U, Delgado-Escueta AV, Mandelkern M, Khonsari A. Neuroimaging in temporal lobe epilepsy: Test sensitivity and relationships to pathology and postoperative outcome. Epilepsia 1992;33:624-34. [81 Engel J Jr, Henry TR, Risinger MW, et al. Presurgical evaluation for partial epilepsy: Relative contributions of chronic depth-electrode recordings versus FDG-PET and scalp sphenoidal ictal EEG. Neurology 1990;40:1670-7.
[9] Abou-Khalil BW, Siegel GJ, Sackerllares C, Gilman S, Hichwa R, Marshall R. Positron emission tomography: Studies of cerebral glucose metabolism in chronic partial epilepsy. Ann Neurol 1987;22:480-6. [10] Labar D, Fraser R, Schaul N, Dhawan V, Eidelberg D. Wrong side PET lateralization in temporal lobe epilepsy. Epilepsia 1992;33 (suppl 3):48. [Ill Levesque M, Harkness W, Sutherling W, et al. SEEG localization in epileptic patients who fail localization after noninvasive studies. Epilepsia 1992;(suppl 3):90-1.
[12] Swartz BE, Halgren E, Delgado-Escueta AV, et al. Neuroimaging in patients with seizures of probable frontal lobe origin. Epilepsia 1989;30:547-58. [13] Snead OC, Kongelbeck S, Mitchell WG. Positron emission tomography in the evaluation of pediatric epilepsy surgery candidates. Epilepsia 1992;33(suppl 3):92. [14] Mitchell WG, Snead OC, Raffel C, Gilles F, Kongelbeck SR. Radiologically undetected tubers causing intractable partial seizures in children. Epilepsia 1992;33(suppl 3):96.
PET in Preoperative Evaluation of Intractable Epilepsy Harry T. Chugani, MD
From the Departments of Pediatric Neurology, and Radiology; Children's Hospital of Michigan and Wayne State University; Detroit, Michigan.
Surgical treatment of medically-refractory epilepsy in children has gained considerable popularity in recent years. As a result, pediatric epilepsy surgery programs have been established worldwide. Surgical procedures being performed include temporal lobectomy, extratemporal cortical resection, hemispherectomy, and corpus callosotomy. Although most of these procedures had been performed in a small number of epilepsy surgery centers for several decades, recent interest has focused more on infants and young children because of improved neurosurgical techniques and intensive care, increasing recognition and appreciation of considerable plasticity in the developing brain, and the desire to avoid progressive epileptic encephalopathy. Despite the belief of most pediatric neurologists and epileptologists that epilepsy surgery is effective in selected infants and children, there appears to be no consensus as to a standard preoperative assessment protocol; however, most would agree that following the confirmation of medical intractability of the seizures, EEG localization (or lateralization) of seizure onset should be correlated with localization from seizure semiology, clinical examination, and neuroimaging modalities, thereby providing enough convergence of data to indicate the epileptogenic region for resection. Epilepsy surgery owes its success to the identification of discrete epileptogenic foci that at least lateralize to one hemisphere. The clinical and electrographic features of seizures arising from a single focus in children often cannot be differentiated from seizures that arise from multiple bilateral and diffuse foci. In the absence of an MRI
lesion, interictal and ictal scalp EEG localization is insufficient to permit proceeding to surgery, and must be confirmed by chronic intracranial EEG monitoring, such as depth or subdural electrode recordings. The development of PET technology has enabled brain functional activity to be imaged noninvasively [1]. Although it can be used to image and acquire quantitative measurements of a wide variety of biochemical processes in brain, the clinical application of PET in pediatric epilepsy surgery has been limited to studies of cerebral glucose utilization with the tracer FDG. Even so, FDG-PET has proved to be extremely useful in infants and children who have intractable epilepsy without an MRI lesion. We found that in such instances, PET contributes significantly to the surgical decision-making process. By fulfilling a role similar to that of MRI when the latter fails to detect a lesion, PET provides some guide as to the type of resection to be performed. Furthermore, PET can provide an assessment of the functional integrity of brain regions outside the epileptogenic area. When data from PET are incorporated in such a manner into the presurgical evaluation, chronic, invasive electrographic monitoring can be avoided in the majority of infants and children. In a recent analysis of 84 infants and children undergoing surgery for intractable epilepsy at the University of California, Los Angeles [2], only 7 (i.e., 8.3%) required chronic, invasive monitoring as a result of the incorporation of PET into the surgical decision-making process. This less invasive (and less expensive) approach is particularly important in children because of the technical difficulties and morbidity associated with chronic intracranial monitoring [3]. The surgical procedures in our series at analysis consisted of hemispherectomy (n = 27), temporal lobectomy (n = 24), extended cortical resection (n = 26), and corpus callosotomy (n = 7). Postoperatively, seizures were eliminated or > 90% reduced in 75% of patients. These results are about
PEDIATRIC NEUROLOGY
Vol. 9 No. 5 411