Lochation of E,xtratemporal Epileptic Foci During Ictal lSinglePhoton Emission Computed Tomography David A. Marks, MD," Amiram Katz, MD," Paul Hoffer, MD,I and Susan S. Spencer, MD"
We obtained single photon emission computed tomography (SPECT) scans with technetium-9%-hexamethylpropylene-amine-oxime in 1 1 patients during 12 extratemporal partial seizures (9 simple partial, 3 complex partial). Ten ictal SPECT studies in 9 patients showed a focal region of hyperperfusion, which agreed with electrical seizure onset in 5 and with clinical seizure localization in 4 in whom ictal electroencephalography was not localized. Contralatera1 cerebellar and ipsilateral basal ganglia hyperperfusion was seen in 3 patients with a frontal lobe seizure focus. Ictal hyperperfusion was well circumscribed, unlike the diffuse hyperperfusion changes reported during temporal lobe seizures. This observation may indicate a different degree of seizure spread in temporal as opposed to extratemgoral epilepsy. Because electroencephalographic localization is often elusive in extratemporal seizures, ictal SPECT may be very helpful for the localization of extratemporal foci. Marks DA, Katz A, Hoffer P, Spencer SS. Localization of extratemporal epileptic foci during ictal single photon emission computed tomography. Ann Neurol 1992;.Z1:250-255
It has been known for more than 50 years that cerebral blood flow (CBF) increases during a s,eizure 111. h a 1 single photon emission computed tomography (SPECT) and positron emission tomography (PET) can exploit this phenomenon to localize the epibeptic focus, and are therefore useful methods in the presurgical evaluation of patients with epilepsy [Z-81. Several studies with both SPECT and PET have shown, however, that regions of hypermetabolism and hyperperfusion extend over a large region of the b.rain during complex partial ICP) seizures [2, 3, 5-71, This phenomenon was attributed to seizure spread [ S ] . Ictal SPECT was therefore often able to lateralize but not accurately localize the seizure focus. Postictal SPECT, performed within 20 minutes after seizure onset, is another useful method for seizure localization in patients with temporal lobe epilepsy; the region of hyperperfusion is usually limited to the mesial temporal lobe with associated lateral temporal hypoperfusion r2). No studies have specifically evaluated the use of ictal SPECT in patients with extratemporal foc.i, a difficult group of patients to localize. Regional cerebral variability in degree of seizure propagation and frequency of simple partial (9') versus C P seizures might also influence results of ictal blood flow. Seizure spread is more restricted during SP compared with C P and secondary generalized seizures, and widespread perfusion
From the Departmenrs of 'Neurology and +Diagnostic Radiology (Nuclear Medicine), Yale IJniversity School of Medicine, New Haven, CT.
changes might therefore not be so pronouncecl. We evaluated the use of ictal technetium-c)c)hi-hexamethvlpropylene-amiiie-oxime (Tc-HMPAO) SPECT in patients with extratemporal SP and C.P seizures
Methods Study Population Twelve ictal Tc-HMPAO SPECT scans wcre performed 011 11 patients with extratemporal partial seizures (1 patient was scanned twice); 8 patients were also subjected to an interic tal Tc-HMPAO SPECT study. Eight patients were being evaluated according to our epilepsy surgery protocol for possible surgery for medically uncontrolled epilepsy [ 9 ] . The other 3 patients had new onset focal motor seizures secondary to a stroke. Focal motor activity was a clinical accompaniment of seizures in all 11 patients. Nine ictal scans in 8 patients were obrained during an SP seizure, and 3 during a CP seizure. The Table summarizes the clinical seizure characteristics, electroencephalographic (EEG), SPECT, and magnetic rcsonance imaging (MRI) findings.
Determinution of the Epileptic Focus Seizure localization was determined using a combination o f ictal and interictal EEG, clinical seizure characteristics, and MRI. Ictal surface EEG recordings were obtained in all patients (16-channel scalp EEG using the International 10-20 system). Four of these patients also underwent intracranial EEG monitoring. In the 8 patients who underwent presurgi-
Address correspondence to Dr Spencer, Deparrmenr of Neurology, Yale Universky School of Medicine, 313 Cedar St, New Haven. CT Oh5 10.
Received Apr 17, 1991, and in revised form Jul 5. Accepted for publication Jul 28, 1991.
250
Copyright Q 1092 by the American Neurological Association
Clinical Feutures. Ictal E E G . SPECT. and MRI Findings SPECT Patient No. (AeeiSex)
Clinical Seizure
Ictal EEG
Interictal
Ictal
MRI
k g h t body motor (leg > arm > face)
Left SMAa
Not done
Left thalamic hyperdensity
Right body motor (face > arm > leg) Right body motor (arm > leg), eyes, head deviated right Left body motor (leg > arm), sensory aura Right arm tonic, forced right head, eye deviation Left face, arm clonic, sensory prodronieb
Left central
Not done
Left frontal
Not done
Inc left superior frontal, left basal ganglia, right cerebellum Inc left parietal occipital, inc left basal ganglia Inc left basal ganglia, dec left hemisphere
Righc anterior mesial frontala
Dec right hemisphere
Inc righr anterior mesial frontal
Left fron tal-cen tral
Dec left temporal
Dec left temporal
Right central, interictal, left central PLEDs Ictal unlocalized, interictal right orbitofrontal” Right frontal-polar
Dec right frontal temporal
Inc right parietal
Mild right hemisphere atrophy
Normal
Inc right mesial frontal
Normal
Normal
Inc right anterior frontal
No paroxysmal changes
Dec right parietal
Dec right parietal, inc right posterior frontal
Right frontal-polar infarct Right parietal infarct
Normal
Dec left hemisphere
Left frontal opercular infarct
Left mesial frontal”
Dec left central
Inc left lateral posterior frontal Inc left central
N o paroxysmal changes
Nor done
1nc right lateral frontal
Normal
Tonic, turns left, incontinent, no LOC Right face, arm clonic, confusedb Right facial twitching (EPC)
10 (27/M)
11 (4iM)
Right face > arm, clonic (EPC) Right body clonic generalizes‘ Left hand / face twitch, EPC
Left parietal occipital signal changes Left hemisphere atrophy
k g h t orbitofrontal, parasagittal enhancement Left parietal temporal dysplasia
Left SMA resection
aDepth EEG study. bCompIex partial seizure
SMA = supplementary motor area; inc = increase; dec = periodic lateralizing epileptiform discharge.
=
decrease; LOC
cal evaluation, EEG was recorded continuously with simultaneous audiovisual monitoring, and 6 of these patients were injected with reconstituted Tc-HMPAO during their EEG monitoring. For the remaining scans, Tc-HMPAO was not injected during EEG recording. In these patients, accurate descriptions of the clinical seizures during Tc-HMPAO injection were obtained from trained medical observers, and we confirmed that seizures studied during ictal SPECT were behaviorally identical to the patient’s habitual seizures.
Tc-HMPAO SPECT Imaging Patients were scanned with either a head-dedicated multicrystal SPECT imaging device (Model Number 810; Strichman Medical Equipment, Medfield, MA) or a triple-headed, dedicated head scanner (Prism; Picker International, Highland Heights, O H ) . Imaging with the Strichman 810, a 12-crystal focused collimator device, was done by sequential 1-cm spaced slice imaging from the vertex to the base of the brain
=
loss of consciousness; EPC = epilepsia partialis continua; PLEDs
(average, 10-15 slices) using a 576-hole collimator. The inplane resolution at the center of the field was measured at approximately 7-mm full width half maximum (FWHM) for T c - 9 9 ~ .The Z-axis resolution was approximately 1.6 cm FWHM. A 20% symmetrical energy window was used at the 140 keV T c - 9 9 ~peak. Approximately 1 million counts per slice were obtained using a 300-second-per-slice collection time. imaging with the Prism scanner, a triple-headed scanner using Anger logic for localization of events, was performed with the entire brain imaged at once. The camera was operated in the “stop and shoot” mode with acquisition at 3-degree intervals, with 1 minute of acquisition per interval. A low-energy, high-resolution system was measured at approximately 7 mm FWHM for T c - 9 9 ~in all planes at the center of the field. A 20% symmetrical energy window was used centered on the 140 keV peak of T c - 9 9 ~ .Imaging studies were performed using the same instrument for the same patient. Marks et al: Ictal SPECT
251
Patients were injected with 20 mCi of Tc-HM;PAO prepared with the Amersham Ceretec T M kit (Amersham Corporation, Arlington Heights, IL). The reconstituted formula, prepared with 20 mCi of technetium pertechnetate (99~Tc04 ), which was added to the vial containing lyophilized HMPAO, was used within 30 minutes of preparation. For all ictal studies, reconstituted Tc-HMPAO was injected during ongoing clinical seizure activity. Three patients had epilepsia partialis continua, and 2 patients experienced frequent seizures spaced approximately 30 minutes apiart; it was therefore relatively easy to obtain ictal scans in these 5 patients. During EEG monitoring, antiepileptic drugs were withdrawn in 3 other patients to precipitate seizures., and ictal studies were obtained during a seizure cluster. In these 3 patients, seizures were’ spaced approximately 2 hours apart. In the remaining 2 patients, seizures were spaced approximately 24 hours apart. and ictal scans were therefo3re fortuitously obtained. Clinical seizures lasted between 30 to 60 seconds in Patients 1, 2, and 6, between 1 to 2 minutes in Patients 3, 4 , 5 , 7, and 10, and were continuous in the 3 patients with epilepsia partialis continua. Tc-HMPAO was injected toward the end of the clinical seizures in all patients. Interictal scans were done following the ictal study and after a minimum period of 24 hours had elapsed since the last clinical seizure. Labeling efficiency was confirmed by chromatography and was always greater than 80%. Ictal and interictal SPECT scans were obtained within 2 hours after TcHMPAO injection. SPECT scans were done with eyes closed in a quiet room without sedation.
hypoperfusion, but both ictal EEG and clinical seizure characteristics strongly suggested a focus originating in the left sensorimotor cortex. H e r seizures were characterized by forced right head and eye deviation followed by right tonic-clonic motor activity that involved her arm more than her leg.
Interictal Tc-HMPAO Studies Interictal SPECT studies were obtained in 8 patients. Seizures were very closely spaced in Patients 1 and 2 and were continuous in Patient 11; therefore, we were unable to obtain interictal studies. A region of hypoperfusion, which was absent on the ictal scan, was seen in 2 of 8 interictal studies (Patients 3 and 5 ) . In both, the decreased perfusion was confined to one hemisphere but extended over a wide region. Intracranial EEG recording from Patient 3 demonstrated multifocal right hemisphere spikes, whereas ictal EEG showed a right mesial frontal focus, and the ictal hyperperfusion was localized to the right mesial frontal region (Fig 1). Patient 5 had continuous interictal right central periodic lateralizing epileptiform discharges (PLEDs); focal left-sided motor seizures were associated with rightcentral buildup. An area of focal hypoperfusion was seen on both ictal and interictal scans in 4 other patients but was secondary to previous surgical resection, infarct, or cortical dysplasia in these situations.
Analysis Transaxial, coronal, and sagittal scans were reconstructed and interpreted by a nuclear medicine physician (P.H.)without knowledge of each patient’s history. Both color- arid grayscale images were used for qualitative visual interpretation. Ictal and interictal SPECT scans were interpreted as positive when a focal region of hyperperfusion (ictal study) olr hypoperfusion (interictal study) was present on at least 3 contiguous slices. For both ictal and interictal scans, the homologous regions of the opposite hemisphere of the same scan were visually compared to determine a region of hyperperfusion or hypoperfusion. We use the terms focal hypoperjhfsion or hyperperfusion to imply involvement of a portion of a lobe by perfusion changes that were absent on the homologous region of the same scan. Ictal and interictal SPECT was correlated with seizure localization.
Results Location of the
R
Epileptic Focus
Because Patient 1 had 2 distinct clinical and electrical foci, a total of 12 extratemporal seizure foci were analyzed in 11 patients. A frontal lobe focus was documented by ictal EEG in 5 patients and by a cornbination of clinical, MRI, and interictal EEG findings in 3 others. A parietal focus was established by ictal EEG and clinical seizure characteristics in 2 patients and by clinical and radiological data in another. O n e patient with SP motor seizures was not well localized: MRI demonstrated left temporal parietal heterotopias, and ictal and interictal SPECT scans showed left temporal 252
F ig 1 . Sagittal and transaxial icta! single photon emission computed tomograph (upper) from Patient 3 shows focal right anterior mesial frontal hyperperfkion. lntracranial ictal electroencephalography confirmed a right mesial supplementary motor area epileptic focus. Note right frontal-temporal,hypoperjusion on the interictal coronal study (lower).
Annals of Neurology
Vol 31 No 3 March 1992
w
~
~~~~~
Fig 2. Ictal single photon emission computed tomographfrom Patient 2 shows increased left striatal hyperperfusion without corresponding cortical hyperpedusion. The large region of left hemisphere hypopevfusion i.r seconday to ldt hemiatrophy. Surface ictal electroencephalography documented leftfrontal buildup.
Ictal Tc-HMPAO Studies Twelve ictal SPECT scans were analyzed. Ten ictal scans (9 patients) showed focal cortical hyperperfusion. The ictal scan from Patient 2 (Fig 2) showed left striatal hyperperfusion, without corresponding cortical hyperperfusion. Surface EEG documented a left frontal focus in this patient. Patient 4 had both ictal and interictal left temporal lobe hypoperfusion. Subcortical hyperperfusion was seen in 3 patients with frontal lobe foci (see Fig 2; Fig 3), including ipsilateral basal ganglia (Patients 1 and 2) and contralateral cerebellar hyperperfusion (Patients I and 6 ) .
Fig 3. Transaxial ictal single photon emisJion computed tomograph (SPECT)from Patient 1 (j;rststudy, a ) shows left superior frontal hyperperfusion uith associated ldt basal ganglia and contrulateral cerebellar hyperperfusion. Intracranial ictal electroencephalography IEEG, demonstruted a prefrontal seizure focus. Seizures recurred follwing resection of a left frontal focus. Second ictal SPECT s t u g 1 year later 16) demonstrated left parietal-occipital hyperperfusion. and surface ictal EEG showed ldt central buildup.
Correlation of Ictal SPECT with Seizure Localization Ictal SPECT findings correlated with electrical seizure onset in 5 of 7 patients with localized ictal EEGs. In the 2 patients without EEG-SPECT correlation (Patients 2 and 4), ictal cortical hyperperfusion was not observed, despite electrical localization. In Patient 6, no paroxysmal changes were noted on EEG during clinical seizures, but prominent right orbitofrontal interictal spikes were present and ictal SPECT showed right inferior frontal hyperperfusion. In 3 other patients with epilepsia partialis continua secondary to a cortical infarct (Patients 8 and 9) and suspected chronic encephalitis (Patient l l ) , no paroxysmal EEG changes were seen, but ictal SPECT findings agreed with the clinical seizure focus and neuroimaging data. Patients 1, 2, 3, and 10 had surgery: extensive left frontal lobe resection (Patient I), left hemispherectomy (Patient 2), and limited frontal lobe resections (Patients 3 and 10). Patient 2 is seizure free, Patient 1 has experienced a marked reduction in seizures, and Patients 3 and 10 continue to have frequent seizures. Neuropathological findings from Patients 1, 2, and 10 showed nonspecific gliosis, whereas the biopsy from Patient 3, which showed gliosis and perivascular cuffing by lymphocytes, was compatible with a diagnosis of chronic encephalitis. Discussion Ictal and interictal SPECT may be useful noninvasive techniques to localize an epileptic focus and have been exploited for this purpose. Whereas ictal scans demonstrate a region of hyperperfusion, the interictal study may reveal hypoperfusion. Both correspond to the ictal focus, although interictal SPECT is less sensitive 12-8, 10, 111. In addition, postictal SPECT (within 20 min) following CP seizures of temporal lobe origin showed mesial temporal hyperperfusion and lateral temporal neocortical hypoperfusion in 50 of 70 scans 121. Several investigators have stressed that ictal SPECT and PET are difficult to interpret because the region of hyperperfusion involves a large cortical region, which was interpreted to indicate seizure spread. Lee and associates 131 used 123-iodoamphetamine and obtained
R
R
A
B Marks et
al:
Ictal SPECT
253
ictal SPECT scans in 16 patients with temporal lobe epilepsy. Patients were injected during a seizure, and the region of hyperperfusion involved more than one lobe in 11 patients, which included 1 patient with bilateral changes 13). Ictal studies during secondary generalized seizures were also not useful for localization because they often showed diffuse and bilateral regions of hyperperfusion 12, 8). The ictal scans obtained in our patients did not show these widespread perfusion changes. Instead, we observed a focal region of increased cere bra1 perfusion in 8 of 11 studies and more diffuse cha.nges in 1 study. There are nvo possible explanations for these different results. First, 9 scans were done during SP seizures, and eIertric:d seizure spread during SP seizures is presumably more limited than during C P seizures. In 3 of our patients with epilepsia partialis continua, no EEG changes were seen, which implies veq7 limited cortical involvement- by the seizure focus. In chew 3 patients, ictal SPECT showed focal hyperperfusion. Second, 3 patients were scanned during brief frontal lobe seizures. CP seizures of frontal lobe origin are often much briefer than temporal lobe seizures, and intracranial studies have shown that some frontal lobe seizures d o n o t spread beyond the frontal lobe { 1.21. In an ictal PET study with fluorodeoxyglucose (FDG). Engel and colleagues 151 reported widespread increases in cerebral metabolism during both SP and C P seizures. This finding is contrary to the focal ictal blood flow changes we observed with Tc-HMPAO SPECT; however, these two techniques, are fundamentally differenx. IJnlike FDG PET studies, Tc-HMPAO SPECT measures CHF, which is an indirect measutement of certlbral metabolism. Changes in CRF may therefore be less sensitive than changes in glucose metabolism during an epileptic seizure. Alternatively, whereas PET studies are interpreted with measurable units, no units are available for H M P A O SPECT studies, and each cerebral region is compared with respect to the homologous region of the opposite hemisphere. A diffuse bilateral increase in CBF during a seizure would be difficult to identify because there would be no difference in blood flow between ithe two hemispheres. Finaily, and perhaps most significantly, ictal PET is difficult to interpret, because a 20-minute steady state is needed for the study and seizures are too dynamic. I n contrast, Tc-HMPAO is a first-pass agent, and the majority of cerebral uptake occurs within 2 minutes after intravenous injection. The localization of the hyperperfusion seen o n ictal scans correlattxd with electrical, clinical, and radiological localization in 10 of 12 studies, whereas interictal SPECT was an insensitive method for extratemporal localization in our patients. Jn comparison, interictal SPECT for temporal lobe foci has been reported to show temporal lohe hypoperfusion in 50 to 80% of patients 12, 8, 10, 11). Interictal SPECT :jhowed hypo254
Annals of Neurology
Vol 31
No 3 March 1092
perfusion in only 2 of our studies and was widespread in both. Such widespread interictal hemispheric decrease in regional CBF (rCBF) has been previously noted in patients with both SP and C P seizures. One patient from our group had continuous PLEDs. P1,EDs have been reported to cause increased rCHF {13}, which suggests that PLEDs may represent a subclinical seizure; our single case, however. does not supporr this view. An unexpected finding in 2 patients wirh frontal lobe foci was cerebellar hyperperfusion contralateral to the frontal lobe hyperperfusion. The crossed cerebellar hyperperfusion was specific for frontal lobe foci in this group and has not been reported during temporal lobe seizures with motor involvement. The opposite phenomenon, crossed cerebellar diaschisis, has been reported following large acute cerebral hemisphere infarcts and is thought to represent functional disconnection of the opposite cerebellar hemisphere from afferent corticopontocerebellar connections [ 141. lncreased contralateral cerebellar rCBF has two possible explanations. One possibility is increased ipsilateral spinocerebellar afferent input from unilateral motor activity. If this explanation is correct, however, then crossed cerebellar hyperperfusion should occur both during CP seizures with motor involvement and during secondarily generalized seizures. A more likely explanation for the increased contralateral cerebellar rCBF is that it represents electrical seizure spread from a frontal lobe focus, which has extensive efferent projections to the contralateral ccrebellum [lS]. The results in Patient 1 are consistent with this explanation. H e r first ictal SPECT study showed left frontal and striatal hyperperfusion with contralateral cerebellar hyperperfusion. Her second scan, performed 1 year following resection of the left frontal lobe focus at a time when seizures had recurred, showed left parietal-occipital and striatal hyperperfusion without associated crossed cerebellar hyperperfusion. Right-sided focal motor activity was present during both scans. Subcortical structures, including cerebellum, are known to be involved during seizure propagation El, 16-18]. Ueno and colleagues f16] produced a penicillin fbcus in the motor cortex in monkeys and showed that during an electrical seizure, a marked increase in rCE$F in the epileptic focus, ipsilateral basal ganglia, and contralateral cerebellar hemisphere occurred. Gale [ 1 71 showed involvement of subcortical structures during seizure propagation in animals. Left hemisphere cerebellar atrophy was reported to develop in a patient afrer frequent bouts of right-sided focal motor status 1191. Ictal SPECT from this patient showed right frontal hyperperfusion with crossed cerebellar hyperperfiision. The authors suggested that the cerebellar atrophy was secondary to excitotoxic cell damage due to seizure spread.
Tc-HMPAO was used for all studies and proved to be an excellent agent for ictal SPECT. It is a lipophilic compound, and 7 5 % of the agent is extracted across the blood-brain barrier. It is then rapidly converted to a hydrophilic form and is trapped within cells, reaching peak concentration within 2 minutes after injection 120-221. Because it is stable within cells, scans may be done up to 6 to 8 hours after the compound is injected. Because this agent needs to be reconstituted from its freeze-dried form prior to injection and has a shelf life of only 30 minutes, it is technically difficult to obtain ictal SPECT scans. Although I-123-iodoamphetamine can also be used for such studies, this radiopharmaceutical is not available in kit form, redistributes after injection, and is more expensive than Tc-HMPAO [23}. Most studies with Tc-HMPAO or iodoamphetamines are therefore fortuitously obtained. Because our patients experienced multiple partial seizures, we were able to obtain ictal SPECT studies. This procedure would be extremely difficult to accomplish in patients with infrequent seizures. Postictally, however, rCBF remains elevated for up to 15 minutes in the region of the epileptic focus [2, 241. Therefore, until nuclear compounds with more stable shelf lives are developed, immediate postictal SPECT studies represent a technically practical alternative method for seizure localization. Although this technique has previously been reported to be useful for localization in temporal lobe epilepsy, it has not yet been systematically studied in extratemporal patients. In our study group, 5 patients were scanned during seizures that were spaced between 30 minutes to 2 hours apart. Therefore, images obtained in these 5 patients may reflect a combination of ictal and postictal events. Jctal scans performed during seizure clusters, however, showed similar flow characteristics compared with the ictal scans performed during more widely spaced seizures or during epilepsia partialis continua. Specifically, the region of hyperperfusion remained focal in all studies, regardless of the temporal relationship between seizures. Extratemporal seizure foci (without central nervous system lesions) are extremely difficult to localize, and surgery for extratemporal epilepsy is less successful compared with temporal lobe epilepsy {25, 261. Ictal SPECT in such patients should therefore be a high priority.
References I. Penfield W, von Santha K, Cipriani A. Cerebral blood flow during induced epileptiform seizures in animals and man. J Neurophysiol 1939;2:257-267 2. Rowe CC, Berkovic SF, Sia STB, et al. Localization of epileptic foci with posticral single photon emission computed tomography. Ann Neurol 1989;26:660-668 3. Lee B1, Markand O N , Wellman HN, et al. HIPDM-SPECT in patients with medically intractable complex partial seizures. Ictal study. Arch Neurol 1988;45:397-402
4. Lee BI, Markand ON, Wellman HN, et al. HIPDM single photon emission computed tomography brain imaging in partial onset secondarily generalized tonic-clonic seizures. Epilepsia 1987;28:305-311 5. Engel J Jr, Kuhl DE, Phelps ME, et al. Local cerebral metabolism during partial seizures. Neurology 1983;33:400-413 6. Theodore W H , Newmark ME, Sato S, et al. (18F) fluorodeoxyglucose positron emission tomography in refractory complex partial seizures. Ann Neurol 1983;14:429-437 7. Stefan H, Pawlik G , Bocher-Schwartz HG. Functional and rnorphological abnormalities in temporal lobe epilepsy: a comparison with ictal EEG, CT, MRI, SPECT and PET. J Neurol 1987;234:377-384 8. Denays R, Rubinstein M, Ham H, et al. Single photon emission computed tomography in seizure disorders. Arch Dis Child 1988;63:1184-1188 9. Spencer SS, Katz A. Arriving at the surgical options for intractable seizures. Semin Neurol 1990;10:422-430 0. Stefan H, Bauer J, Feistel H , et al. Regional cerebral blood flow during focal seizures of temporal and frontocentral onset. Ann Neurol l990;27: 162-166 1. Bierstack HJ, Stefan H, Reichman K, et al. HMPAO brain SPECT and epilepsy. Nucl Med Commun 1987;8: 5 13-5 18 2. Williamson PD, Spencer DD, Spencer SS, et al. Complex partial seizures of frontal lobe origin. Ann Neurol 1985;18:497-504 3. Lee BI, Schauwecker DS. Regional cerebral perfusion in PLEDs: a case report. Epilepsia 1988;29:607-611 14. Pantana P, Baron JC, Samson Y, et al. Crossed cerebellar diaschisis: further studies. Brain 1986;109:251-283 15. Brodal A. Cerebrocerebellar pathways. Anatomical and some functional implications. Acta Neurol Scand 1972;5l(supp1): 153-196 16. Ueno H, Yamashita Y, Caveness WF. Regional cerebral blood flow pattern in focal epileptiform seizures in the monkey. Exp Neurol 1976;47:81-96 17. Gale K. Progression and generalization of seizure discharge: anatomical and neurochemical substrates. Epilepsia 1988;29 (suppl 2):S15-S34 18. Collins RC. Use of cortical circuits during focal penicillin seizures: an autoradiographic study with ( 1 4 0 deoxyglucose. Brain Res 1978; 15O:487->01 19. Duncan R, Patterson J, Hadley DM, et al. Unilateral cerebellar damage in focal epilepsy. J Neurol Neurosurg Psychiatry 1990;53:436-442 20. Lassen N A , Blasberg RG. Technetium-99m-d,l-HM-PAO, the development of a new class of 99mTc-labelled tracers: an overview. J Cereb Blood Flow Metab 1988;8(suppl 1):SI-S3 21. Neirinckx RD, Canning LR, Piper IM, et al. Tc-99m d,l-HMPAO: a new radiopharmaceutical for cerebral blood flow imaging. Nucl Med Commun 1985;6:499-506 22. Anderson AR, Froberg H H , Schmidy JF, et a]. Quantitative measurements of cerebral blood flow using SPECT and (99mTc)-d, I-HM-PA0 compared to Xenon-I 33. J Cereb Blood Flow Metab 1988;8(suppI 1):S69-S81 23. Holman BL, Lee RGL, Hill TC, et al. A comparison of two cerebral blood flow tracers, n-isopropyl-I- 123-p-iodoamphetamine and I-123HIDPM. J Nucl Med 1984;25:25-30 24. Dymond AM, Crandall PH. Oxygen availability and blood flow in the temporal lobes during spontaneous epileptic seizures in man. Brain Res 1976;103:191-196 25. Swarm RE, Delgado-Escueta AV. Complex partial seizures of extratemporal origin: the evidence for. In. Wieser H G , Speckman EJ, Engel J Jr, eds. Current problems in epilepsy 3. The epileptic focus. London: John Libbey, 1987:137-174 26. Ludwig B, Ajmone-Marsan C, Van Buren J. Depth and direct cortical recording in seizure disorders of extratemporal origin. Neurology 1976;26: 108>-1099 M a r k s et al: Ictal SPECT
255
We obtained single photon emission computed tomography (SPECT) scans with technetium-9%-hexamethylpropylene-amine-oxime in 1 1 patients during 12 extratemporal partial seizures (9 simple partial, 3 complex partial). Ten ictal SPECT studies in 9 patients showed a focal region of hyperperfusion, which agreed with electrical seizure onset in 5 and with clinical seizure localization in 4 in whom ictal electroencephalography was not localized. Contralatera1 cerebellar and ipsilateral basal ganglia hyperperfusion was seen in 3 patients with a frontal lobe seizure focus. Ictal hyperperfusion was well circumscribed, unlike the diffuse hyperperfusion changes reported during temporal lobe seizures. This observation may indicate a different degree of seizure spread in temporal as opposed to extratemgoral epilepsy. Because electroencephalographic localization is often elusive in extratemporal seizures, ictal SPECT may be very helpful for the localization of extratemporal foci. Marks DA, Katz A, Hoffer P, Spencer SS. Localization of extratemporal epileptic foci during ictal single photon emission computed tomography. Ann Neurol 1992;.Z1:250-255
It has been known for more than 50 years that cerebral blood flow (CBF) increases during a s,eizure 111. h a 1 single photon emission computed tomography (SPECT) and positron emission tomography (PET) can exploit this phenomenon to localize the epibeptic focus, and are therefore useful methods in the presurgical evaluation of patients with epilepsy [Z-81. Several studies with both SPECT and PET have shown, however, that regions of hypermetabolism and hyperperfusion extend over a large region of the b.rain during complex partial ICP) seizures [2, 3, 5-71, This phenomenon was attributed to seizure spread [ S ] . Ictal SPECT was therefore often able to lateralize but not accurately localize the seizure focus. Postictal SPECT, performed within 20 minutes after seizure onset, is another useful method for seizure localization in patients with temporal lobe epilepsy; the region of hyperperfusion is usually limited to the mesial temporal lobe with associated lateral temporal hypoperfusion r2). No studies have specifically evaluated the use of ictal SPECT in patients with extratemporal foc.i, a difficult group of patients to localize. Regional cerebral variability in degree of seizure propagation and frequency of simple partial (9') versus C P seizures might also influence results of ictal blood flow. Seizure spread is more restricted during SP compared with C P and secondary generalized seizures, and widespread perfusion
From the Departmenrs of 'Neurology and +Diagnostic Radiology (Nuclear Medicine), Yale IJniversity School of Medicine, New Haven, CT.
changes might therefore not be so pronouncecl. We evaluated the use of ictal technetium-c)c)hi-hexamethvlpropylene-amiiie-oxime (Tc-HMPAO) SPECT in patients with extratemporal SP and C.P seizures
Methods Study Population Twelve ictal Tc-HMPAO SPECT scans wcre performed 011 11 patients with extratemporal partial seizures (1 patient was scanned twice); 8 patients were also subjected to an interic tal Tc-HMPAO SPECT study. Eight patients were being evaluated according to our epilepsy surgery protocol for possible surgery for medically uncontrolled epilepsy [ 9 ] . The other 3 patients had new onset focal motor seizures secondary to a stroke. Focal motor activity was a clinical accompaniment of seizures in all 11 patients. Nine ictal scans in 8 patients were obrained during an SP seizure, and 3 during a CP seizure. The Table summarizes the clinical seizure characteristics, electroencephalographic (EEG), SPECT, and magnetic rcsonance imaging (MRI) findings.
Determinution of the Epileptic Focus Seizure localization was determined using a combination o f ictal and interictal EEG, clinical seizure characteristics, and MRI. Ictal surface EEG recordings were obtained in all patients (16-channel scalp EEG using the International 10-20 system). Four of these patients also underwent intracranial EEG monitoring. In the 8 patients who underwent presurgi-
Address correspondence to Dr Spencer, Deparrmenr of Neurology, Yale Universky School of Medicine, 313 Cedar St, New Haven. CT Oh5 10.
Received Apr 17, 1991, and in revised form Jul 5. Accepted for publication Jul 28, 1991.
250
Copyright Q 1092 by the American Neurological Association
Clinical Feutures. Ictal E E G . SPECT. and MRI Findings SPECT Patient No. (AeeiSex)
Clinical Seizure
Ictal EEG
Interictal
Ictal
MRI
k g h t body motor (leg > arm > face)
Left SMAa
Not done
Left thalamic hyperdensity
Right body motor (face > arm > leg) Right body motor (arm > leg), eyes, head deviated right Left body motor (leg > arm), sensory aura Right arm tonic, forced right head, eye deviation Left face, arm clonic, sensory prodronieb
Left central
Not done
Left frontal
Not done
Inc left superior frontal, left basal ganglia, right cerebellum Inc left parietal occipital, inc left basal ganglia Inc left basal ganglia, dec left hemisphere
Righc anterior mesial frontala
Dec right hemisphere
Inc righr anterior mesial frontal
Left fron tal-cen tral
Dec left temporal
Dec left temporal
Right central, interictal, left central PLEDs Ictal unlocalized, interictal right orbitofrontal” Right frontal-polar
Dec right frontal temporal
Inc right parietal
Mild right hemisphere atrophy
Normal
Inc right mesial frontal
Normal
Normal
Inc right anterior frontal
No paroxysmal changes
Dec right parietal
Dec right parietal, inc right posterior frontal
Right frontal-polar infarct Right parietal infarct
Normal
Dec left hemisphere
Left frontal opercular infarct
Left mesial frontal”
Dec left central
Inc left lateral posterior frontal Inc left central
N o paroxysmal changes
Nor done
1nc right lateral frontal
Normal
Tonic, turns left, incontinent, no LOC Right face, arm clonic, confusedb Right facial twitching (EPC)
10 (27/M)
11 (4iM)
Right face > arm, clonic (EPC) Right body clonic generalizes‘ Left hand / face twitch, EPC
Left parietal occipital signal changes Left hemisphere atrophy
k g h t orbitofrontal, parasagittal enhancement Left parietal temporal dysplasia
Left SMA resection
aDepth EEG study. bCompIex partial seizure
SMA = supplementary motor area; inc = increase; dec = periodic lateralizing epileptiform discharge.
=
decrease; LOC
cal evaluation, EEG was recorded continuously with simultaneous audiovisual monitoring, and 6 of these patients were injected with reconstituted Tc-HMPAO during their EEG monitoring. For the remaining scans, Tc-HMPAO was not injected during EEG recording. In these patients, accurate descriptions of the clinical seizures during Tc-HMPAO injection were obtained from trained medical observers, and we confirmed that seizures studied during ictal SPECT were behaviorally identical to the patient’s habitual seizures.
Tc-HMPAO SPECT Imaging Patients were scanned with either a head-dedicated multicrystal SPECT imaging device (Model Number 810; Strichman Medical Equipment, Medfield, MA) or a triple-headed, dedicated head scanner (Prism; Picker International, Highland Heights, O H ) . Imaging with the Strichman 810, a 12-crystal focused collimator device, was done by sequential 1-cm spaced slice imaging from the vertex to the base of the brain
=
loss of consciousness; EPC = epilepsia partialis continua; PLEDs
(average, 10-15 slices) using a 576-hole collimator. The inplane resolution at the center of the field was measured at approximately 7-mm full width half maximum (FWHM) for T c - 9 9 ~ .The Z-axis resolution was approximately 1.6 cm FWHM. A 20% symmetrical energy window was used at the 140 keV T c - 9 9 ~peak. Approximately 1 million counts per slice were obtained using a 300-second-per-slice collection time. imaging with the Prism scanner, a triple-headed scanner using Anger logic for localization of events, was performed with the entire brain imaged at once. The camera was operated in the “stop and shoot” mode with acquisition at 3-degree intervals, with 1 minute of acquisition per interval. A low-energy, high-resolution system was measured at approximately 7 mm FWHM for T c - 9 9 ~in all planes at the center of the field. A 20% symmetrical energy window was used centered on the 140 keV peak of T c - 9 9 ~ .Imaging studies were performed using the same instrument for the same patient. Marks et al: Ictal SPECT
251
Patients were injected with 20 mCi of Tc-HM;PAO prepared with the Amersham Ceretec T M kit (Amersham Corporation, Arlington Heights, IL). The reconstituted formula, prepared with 20 mCi of technetium pertechnetate (99~Tc04 ), which was added to the vial containing lyophilized HMPAO, was used within 30 minutes of preparation. For all ictal studies, reconstituted Tc-HMPAO was injected during ongoing clinical seizure activity. Three patients had epilepsia partialis continua, and 2 patients experienced frequent seizures spaced approximately 30 minutes apiart; it was therefore relatively easy to obtain ictal scans in these 5 patients. During EEG monitoring, antiepileptic drugs were withdrawn in 3 other patients to precipitate seizures., and ictal studies were obtained during a seizure cluster. In these 3 patients, seizures were’ spaced approximately 2 hours apart. In the remaining 2 patients, seizures were spaced approximately 24 hours apart. and ictal scans were therefo3re fortuitously obtained. Clinical seizures lasted between 30 to 60 seconds in Patients 1, 2, and 6, between 1 to 2 minutes in Patients 3, 4 , 5 , 7, and 10, and were continuous in the 3 patients with epilepsia partialis continua. Tc-HMPAO was injected toward the end of the clinical seizures in all patients. Interictal scans were done following the ictal study and after a minimum period of 24 hours had elapsed since the last clinical seizure. Labeling efficiency was confirmed by chromatography and was always greater than 80%. Ictal and interictal SPECT scans were obtained within 2 hours after TcHMPAO injection. SPECT scans were done with eyes closed in a quiet room without sedation.
hypoperfusion, but both ictal EEG and clinical seizure characteristics strongly suggested a focus originating in the left sensorimotor cortex. H e r seizures were characterized by forced right head and eye deviation followed by right tonic-clonic motor activity that involved her arm more than her leg.
Interictal Tc-HMPAO Studies Interictal SPECT studies were obtained in 8 patients. Seizures were very closely spaced in Patients 1 and 2 and were continuous in Patient 11; therefore, we were unable to obtain interictal studies. A region of hypoperfusion, which was absent on the ictal scan, was seen in 2 of 8 interictal studies (Patients 3 and 5 ) . In both, the decreased perfusion was confined to one hemisphere but extended over a wide region. Intracranial EEG recording from Patient 3 demonstrated multifocal right hemisphere spikes, whereas ictal EEG showed a right mesial frontal focus, and the ictal hyperperfusion was localized to the right mesial frontal region (Fig 1). Patient 5 had continuous interictal right central periodic lateralizing epileptiform discharges (PLEDs); focal left-sided motor seizures were associated with rightcentral buildup. An area of focal hypoperfusion was seen on both ictal and interictal scans in 4 other patients but was secondary to previous surgical resection, infarct, or cortical dysplasia in these situations.
Analysis Transaxial, coronal, and sagittal scans were reconstructed and interpreted by a nuclear medicine physician (P.H.)without knowledge of each patient’s history. Both color- arid grayscale images were used for qualitative visual interpretation. Ictal and interictal SPECT scans were interpreted as positive when a focal region of hyperperfusion (ictal study) olr hypoperfusion (interictal study) was present on at least 3 contiguous slices. For both ictal and interictal scans, the homologous regions of the opposite hemisphere of the same scan were visually compared to determine a region of hyperperfusion or hypoperfusion. We use the terms focal hypoperjhfsion or hyperperfusion to imply involvement of a portion of a lobe by perfusion changes that were absent on the homologous region of the same scan. Ictal and interictal SPECT was correlated with seizure localization.
Results Location of the
R
Epileptic Focus
Because Patient 1 had 2 distinct clinical and electrical foci, a total of 12 extratemporal seizure foci were analyzed in 11 patients. A frontal lobe focus was documented by ictal EEG in 5 patients and by a cornbination of clinical, MRI, and interictal EEG findings in 3 others. A parietal focus was established by ictal EEG and clinical seizure characteristics in 2 patients and by clinical and radiological data in another. O n e patient with SP motor seizures was not well localized: MRI demonstrated left temporal parietal heterotopias, and ictal and interictal SPECT scans showed left temporal 252
F ig 1 . Sagittal and transaxial icta! single photon emission computed tomograph (upper) from Patient 3 shows focal right anterior mesial frontal hyperperfkion. lntracranial ictal electroencephalography confirmed a right mesial supplementary motor area epileptic focus. Note right frontal-temporal,hypoperjusion on the interictal coronal study (lower).
Annals of Neurology
Vol 31 No 3 March 1992
w
~
~~~~~
Fig 2. Ictal single photon emission computed tomographfrom Patient 2 shows increased left striatal hyperperfusion without corresponding cortical hyperpedusion. The large region of left hemisphere hypopevfusion i.r seconday to ldt hemiatrophy. Surface ictal electroencephalography documented leftfrontal buildup.
Ictal Tc-HMPAO Studies Twelve ictal SPECT scans were analyzed. Ten ictal scans (9 patients) showed focal cortical hyperperfusion. The ictal scan from Patient 2 (Fig 2) showed left striatal hyperperfusion, without corresponding cortical hyperperfusion. Surface EEG documented a left frontal focus in this patient. Patient 4 had both ictal and interictal left temporal lobe hypoperfusion. Subcortical hyperperfusion was seen in 3 patients with frontal lobe foci (see Fig 2; Fig 3), including ipsilateral basal ganglia (Patients 1 and 2) and contralateral cerebellar hyperperfusion (Patients I and 6 ) .
Fig 3. Transaxial ictal single photon emisJion computed tomograph (SPECT)from Patient 1 (j;rststudy, a ) shows left superior frontal hyperperfusion uith associated ldt basal ganglia and contrulateral cerebellar hyperperfusion. Intracranial ictal electroencephalography IEEG, demonstruted a prefrontal seizure focus. Seizures recurred follwing resection of a left frontal focus. Second ictal SPECT s t u g 1 year later 16) demonstrated left parietal-occipital hyperperfusion. and surface ictal EEG showed ldt central buildup.
Correlation of Ictal SPECT with Seizure Localization Ictal SPECT findings correlated with electrical seizure onset in 5 of 7 patients with localized ictal EEGs. In the 2 patients without EEG-SPECT correlation (Patients 2 and 4), ictal cortical hyperperfusion was not observed, despite electrical localization. In Patient 6, no paroxysmal changes were noted on EEG during clinical seizures, but prominent right orbitofrontal interictal spikes were present and ictal SPECT showed right inferior frontal hyperperfusion. In 3 other patients with epilepsia partialis continua secondary to a cortical infarct (Patients 8 and 9) and suspected chronic encephalitis (Patient l l ) , no paroxysmal EEG changes were seen, but ictal SPECT findings agreed with the clinical seizure focus and neuroimaging data. Patients 1, 2, 3, and 10 had surgery: extensive left frontal lobe resection (Patient I), left hemispherectomy (Patient 2), and limited frontal lobe resections (Patients 3 and 10). Patient 2 is seizure free, Patient 1 has experienced a marked reduction in seizures, and Patients 3 and 10 continue to have frequent seizures. Neuropathological findings from Patients 1, 2, and 10 showed nonspecific gliosis, whereas the biopsy from Patient 3, which showed gliosis and perivascular cuffing by lymphocytes, was compatible with a diagnosis of chronic encephalitis. Discussion Ictal and interictal SPECT may be useful noninvasive techniques to localize an epileptic focus and have been exploited for this purpose. Whereas ictal scans demonstrate a region of hyperperfusion, the interictal study may reveal hypoperfusion. Both correspond to the ictal focus, although interictal SPECT is less sensitive 12-8, 10, 111. In addition, postictal SPECT (within 20 min) following CP seizures of temporal lobe origin showed mesial temporal hyperperfusion and lateral temporal neocortical hypoperfusion in 50 of 70 scans 121. Several investigators have stressed that ictal SPECT and PET are difficult to interpret because the region of hyperperfusion involves a large cortical region, which was interpreted to indicate seizure spread. Lee and associates 131 used 123-iodoamphetamine and obtained
R
R
A
B Marks et
al:
Ictal SPECT
253
ictal SPECT scans in 16 patients with temporal lobe epilepsy. Patients were injected during a seizure, and the region of hyperperfusion involved more than one lobe in 11 patients, which included 1 patient with bilateral changes 13). Ictal studies during secondary generalized seizures were also not useful for localization because they often showed diffuse and bilateral regions of hyperperfusion 12, 8). The ictal scans obtained in our patients did not show these widespread perfusion changes. Instead, we observed a focal region of increased cere bra1 perfusion in 8 of 11 studies and more diffuse cha.nges in 1 study. There are nvo possible explanations for these different results. First, 9 scans were done during SP seizures, and eIertric:d seizure spread during SP seizures is presumably more limited than during C P seizures. In 3 of our patients with epilepsia partialis continua, no EEG changes were seen, which implies veq7 limited cortical involvement- by the seizure focus. In chew 3 patients, ictal SPECT showed focal hyperperfusion. Second, 3 patients were scanned during brief frontal lobe seizures. CP seizures of frontal lobe origin are often much briefer than temporal lobe seizures, and intracranial studies have shown that some frontal lobe seizures d o n o t spread beyond the frontal lobe { 1.21. In an ictal PET study with fluorodeoxyglucose (FDG). Engel and colleagues 151 reported widespread increases in cerebral metabolism during both SP and C P seizures. This finding is contrary to the focal ictal blood flow changes we observed with Tc-HMPAO SPECT; however, these two techniques, are fundamentally differenx. IJnlike FDG PET studies, Tc-HMPAO SPECT measures CHF, which is an indirect measutement of certlbral metabolism. Changes in CRF may therefore be less sensitive than changes in glucose metabolism during an epileptic seizure. Alternatively, whereas PET studies are interpreted with measurable units, no units are available for H M P A O SPECT studies, and each cerebral region is compared with respect to the homologous region of the opposite hemisphere. A diffuse bilateral increase in CBF during a seizure would be difficult to identify because there would be no difference in blood flow between ithe two hemispheres. Finaily, and perhaps most significantly, ictal PET is difficult to interpret, because a 20-minute steady state is needed for the study and seizures are too dynamic. I n contrast, Tc-HMPAO is a first-pass agent, and the majority of cerebral uptake occurs within 2 minutes after intravenous injection. The localization of the hyperperfusion seen o n ictal scans correlattxd with electrical, clinical, and radiological localization in 10 of 12 studies, whereas interictal SPECT was an insensitive method for extratemporal localization in our patients. Jn comparison, interictal SPECT for temporal lobe foci has been reported to show temporal lohe hypoperfusion in 50 to 80% of patients 12, 8, 10, 11). Interictal SPECT :jhowed hypo254
Annals of Neurology
Vol 31
No 3 March 1092
perfusion in only 2 of our studies and was widespread in both. Such widespread interictal hemispheric decrease in regional CBF (rCBF) has been previously noted in patients with both SP and C P seizures. One patient from our group had continuous PLEDs. P1,EDs have been reported to cause increased rCHF {13}, which suggests that PLEDs may represent a subclinical seizure; our single case, however. does not supporr this view. An unexpected finding in 2 patients wirh frontal lobe foci was cerebellar hyperperfusion contralateral to the frontal lobe hyperperfusion. The crossed cerebellar hyperperfusion was specific for frontal lobe foci in this group and has not been reported during temporal lobe seizures with motor involvement. The opposite phenomenon, crossed cerebellar diaschisis, has been reported following large acute cerebral hemisphere infarcts and is thought to represent functional disconnection of the opposite cerebellar hemisphere from afferent corticopontocerebellar connections [ 141. lncreased contralateral cerebellar rCBF has two possible explanations. One possibility is increased ipsilateral spinocerebellar afferent input from unilateral motor activity. If this explanation is correct, however, then crossed cerebellar hyperperfusion should occur both during CP seizures with motor involvement and during secondarily generalized seizures. A more likely explanation for the increased contralateral cerebellar rCBF is that it represents electrical seizure spread from a frontal lobe focus, which has extensive efferent projections to the contralateral ccrebellum [lS]. The results in Patient 1 are consistent with this explanation. H e r first ictal SPECT study showed left frontal and striatal hyperperfusion with contralateral cerebellar hyperperfusion. Her second scan, performed 1 year following resection of the left frontal lobe focus at a time when seizures had recurred, showed left parietal-occipital and striatal hyperperfusion without associated crossed cerebellar hyperperfusion. Right-sided focal motor activity was present during both scans. Subcortical structures, including cerebellum, are known to be involved during seizure propagation El, 16-18]. Ueno and colleagues f16] produced a penicillin fbcus in the motor cortex in monkeys and showed that during an electrical seizure, a marked increase in rCE$F in the epileptic focus, ipsilateral basal ganglia, and contralateral cerebellar hemisphere occurred. Gale [ 1 71 showed involvement of subcortical structures during seizure propagation in animals. Left hemisphere cerebellar atrophy was reported to develop in a patient afrer frequent bouts of right-sided focal motor status 1191. Ictal SPECT from this patient showed right frontal hyperperfusion with crossed cerebellar hyperperfiision. The authors suggested that the cerebellar atrophy was secondary to excitotoxic cell damage due to seizure spread.
Tc-HMPAO was used for all studies and proved to be an excellent agent for ictal SPECT. It is a lipophilic compound, and 7 5 % of the agent is extracted across the blood-brain barrier. It is then rapidly converted to a hydrophilic form and is trapped within cells, reaching peak concentration within 2 minutes after injection 120-221. Because it is stable within cells, scans may be done up to 6 to 8 hours after the compound is injected. Because this agent needs to be reconstituted from its freeze-dried form prior to injection and has a shelf life of only 30 minutes, it is technically difficult to obtain ictal SPECT scans. Although I-123-iodoamphetamine can also be used for such studies, this radiopharmaceutical is not available in kit form, redistributes after injection, and is more expensive than Tc-HMPAO [23}. Most studies with Tc-HMPAO or iodoamphetamines are therefore fortuitously obtained. Because our patients experienced multiple partial seizures, we were able to obtain ictal SPECT studies. This procedure would be extremely difficult to accomplish in patients with infrequent seizures. Postictally, however, rCBF remains elevated for up to 15 minutes in the region of the epileptic focus [2, 241. Therefore, until nuclear compounds with more stable shelf lives are developed, immediate postictal SPECT studies represent a technically practical alternative method for seizure localization. Although this technique has previously been reported to be useful for localization in temporal lobe epilepsy, it has not yet been systematically studied in extratemporal patients. In our study group, 5 patients were scanned during seizures that were spaced between 30 minutes to 2 hours apart. Therefore, images obtained in these 5 patients may reflect a combination of ictal and postictal events. Jctal scans performed during seizure clusters, however, showed similar flow characteristics compared with the ictal scans performed during more widely spaced seizures or during epilepsia partialis continua. Specifically, the region of hyperperfusion remained focal in all studies, regardless of the temporal relationship between seizures. Extratemporal seizure foci (without central nervous system lesions) are extremely difficult to localize, and surgery for extratemporal epilepsy is less successful compared with temporal lobe epilepsy {25, 261. Ictal SPECT in such patients should therefore be a high priority.
References I. Penfield W, von Santha K, Cipriani A. Cerebral blood flow during induced epileptiform seizures in animals and man. J Neurophysiol 1939;2:257-267 2. Rowe CC, Berkovic SF, Sia STB, et al. Localization of epileptic foci with posticral single photon emission computed tomography. Ann Neurol 1989;26:660-668 3. Lee B1, Markand O N , Wellman HN, et al. HIPDM-SPECT in patients with medically intractable complex partial seizures. Ictal study. Arch Neurol 1988;45:397-402
4. Lee BI, Markand ON, Wellman HN, et al. HIPDM single photon emission computed tomography brain imaging in partial onset secondarily generalized tonic-clonic seizures. Epilepsia 1987;28:305-311 5. Engel J Jr, Kuhl DE, Phelps ME, et al. Local cerebral metabolism during partial seizures. Neurology 1983;33:400-413 6. Theodore W H , Newmark ME, Sato S, et al. (18F) fluorodeoxyglucose positron emission tomography in refractory complex partial seizures. Ann Neurol 1983;14:429-437 7. Stefan H, Pawlik G , Bocher-Schwartz HG. Functional and rnorphological abnormalities in temporal lobe epilepsy: a comparison with ictal EEG, CT, MRI, SPECT and PET. J Neurol 1987;234:377-384 8. Denays R, Rubinstein M, Ham H, et al. Single photon emission computed tomography in seizure disorders. Arch Dis Child 1988;63:1184-1188 9. Spencer SS, Katz A. Arriving at the surgical options for intractable seizures. Semin Neurol 1990;10:422-430 0. Stefan H, Bauer J, Feistel H , et al. Regional cerebral blood flow during focal seizures of temporal and frontocentral onset. Ann Neurol l990;27: 162-166 1. Bierstack HJ, Stefan H, Reichman K, et al. HMPAO brain SPECT and epilepsy. Nucl Med Commun 1987;8: 5 13-5 18 2. Williamson PD, Spencer DD, Spencer SS, et al. Complex partial seizures of frontal lobe origin. Ann Neurol 1985;18:497-504 3. Lee BI, Schauwecker DS. Regional cerebral perfusion in PLEDs: a case report. Epilepsia 1988;29:607-611 14. Pantana P, Baron JC, Samson Y, et al. Crossed cerebellar diaschisis: further studies. Brain 1986;109:251-283 15. Brodal A. Cerebrocerebellar pathways. Anatomical and some functional implications. Acta Neurol Scand 1972;5l(supp1): 153-196 16. Ueno H, Yamashita Y, Caveness WF. Regional cerebral blood flow pattern in focal epileptiform seizures in the monkey. Exp Neurol 1976;47:81-96 17. Gale K. Progression and generalization of seizure discharge: anatomical and neurochemical substrates. Epilepsia 1988;29 (suppl 2):S15-S34 18. Collins RC. Use of cortical circuits during focal penicillin seizures: an autoradiographic study with ( 1 4 0 deoxyglucose. Brain Res 1978; 15O:487->01 19. Duncan R, Patterson J, Hadley DM, et al. Unilateral cerebellar damage in focal epilepsy. J Neurol Neurosurg Psychiatry 1990;53:436-442 20. Lassen N A , Blasberg RG. Technetium-99m-d,l-HM-PAO, the development of a new class of 99mTc-labelled tracers: an overview. J Cereb Blood Flow Metab 1988;8(suppl 1):SI-S3 21. Neirinckx RD, Canning LR, Piper IM, et al. Tc-99m d,l-HMPAO: a new radiopharmaceutical for cerebral blood flow imaging. Nucl Med Commun 1985;6:499-506 22. Anderson AR, Froberg H H , Schmidy JF, et a]. Quantitative measurements of cerebral blood flow using SPECT and (99mTc)-d, I-HM-PA0 compared to Xenon-I 33. J Cereb Blood Flow Metab 1988;8(suppI 1):S69-S81 23. Holman BL, Lee RGL, Hill TC, et al. A comparison of two cerebral blood flow tracers, n-isopropyl-I- 123-p-iodoamphetamine and I-123HIDPM. J Nucl Med 1984;25:25-30 24. Dymond AM, Crandall PH. Oxygen availability and blood flow in the temporal lobes during spontaneous epileptic seizures in man. Brain Res 1976;103:191-196 25. Swarm RE, Delgado-Escueta AV. Complex partial seizures of extratemporal origin: the evidence for. In. Wieser H G , Speckman EJ, Engel J Jr, eds. Current problems in epilepsy 3. The epileptic focus. London: John Libbey, 1987:137-174 26. Ludwig B, Ajmone-Marsan C, Van Buren J. Depth and direct cortical recording in seizure disorders of extratemporal origin. Neurology 1976;26: 108>-1099 M a r k s et al: Ictal SPECT
255
