- Top of page
- SUBJECTS AND METHODS
Summary: Purpose: Reoperation after failed cortical resection can alleviate seizures in patients with intractable neocortical epilepsy, provided that previously nonresected epileptic regions are accurately defined and removed. Most imaging modalities have limited value in identifying such regions after a previous surgery. Positron emission tomography (PET) using α−[11C]methyl-l-tryptophan (AMT) can detect epileptogenic cortical areas as regions with increased tracer uptake. This study analyzed whether increased cortical AMT uptake can detect nonresected epileptic foci in patients with previously failed neocortical resection.
Methods: Thirty-three young patients (age 3–26 years; mean age, 10.8 years) with intractable epilepsy of neocortical origin, and a previously failed cortical resection performed at various epilepsy centers, underwent further presurgical evaluation for reoperation. AMT-PET scans were performed 6 days to 7 years after the first surgery. Focal cortical areas with increased AMT uptake were objectively identified and correlated to ictal EEG data as well as clinical variables (age, postsurgical time, etiology).
Results: Cortical increases of AMT uptake were detected on the side of the previous resections in 12 cases. In two patients scanned shortly (within a week) after surgery, diffuse hemispheric increases were observed, without any further localization value. In contrast, in 10 (43%) of 23 patients scanned >2 months but within 2.3 years after surgery, focal cortical increases occurred, concordant with seizure onset on ictal EEG. Age, etiology (lesional vs. cryptogenic), epileptiform EEG activity during PET, or time of the last seizure were not significantly related to the presence of increased AMT uptake. All patients with localizing AMT-PET, who underwent reoperation, became seizure free (n = 5) or showed considerable improvement of seizure frequency (n = 2).
Conclusions: AMT-PET can identify nonresected epileptic cortex in patients with a previously failed neocortical epilepsy surgery and, with proper timing for the scan, can assist in planning reoperation.
Partial epilepsy is refractory to medical treatment in ≤20% of cases, and resective epilepsy surgery can effectively alleviate seizures in these patients. The outcome of epilepsy surgery varies widely depending on the type of epilepsy and operation performed, and the ultimate results with respect to seizures remain far from optimal. In particular, the success rate of cortical resections in epilepsies of neocortical origin (which are especially common among children), continues to be disappointingly low at 50 to 60% (1–6).
Further resection (reoperation) of epileptogenic brain regions after an initial failed epilepsy surgery can result in seizure freedom. Several studies have shown good results after reoperation (7–12), especially when additional resections included incompletely removed lesions or previous resections were extended based on ictal EEG recordings. However, Schwartz and Spencer (13) recently reported disappointingly low surgical success when the second surgery was performed in patients in whom seizures recurred after comprehensive preoperative evaluation and resection; only 19% of such patients remained seizure free after a 1-year follow-up. In that series, successful outcomes resulted from removal of recurrent tumors, completion of a functional hemispherectomy, or repeated invasive monitoring to correct a previous sampling error. Magnetic resonance imaging (MRI) can reveal missed or incompletely resected lesions, but it is not helpful in cryptogenic cases. Scalp ictal EEG can be difficult to interpret in patients with previous resection because of alterations of the original brain anatomy, shifting, intracranial cavities, as well as effects of skull and dural scarring.
A real need exists for innovative neuroimaging approaches to localize nonresected epileptic cortex after failed resective epilepsy surgery. Functional neuroimaging, including positron emission tomography (PET) and ictal single-photon emission computed tomography (SPECT) may be beneficial in this regard. However, functional imaging studies after epilepsy surgery are scarce, perhaps also reflecting some limitations of these imaging modalities after cortical resection. For example, interictal 2-deoxy-2-[18F]fluoro-d-glucose (FDG) PET can detect hypometabolism in nonresected cortex, but it is difficult to determine whether the observed hypometabolism indicates epileptogenic tissue rather than postsurgical tissue damage or deafferentation. The value of ictal SPECT also remains to be documented in such cases. One case report describes a recurrent epileptogenic oligodendroglioma detected by [11C]l-methionine PET, when it was not visible on MRI (14). Recently the PET tracer α−[11C]methyl-l-tryptophan (AMT) has been reported to be selectively accumulated in and around epileptogenic cortical regions (15–17). AMT was originally developed for the measurement of serotonin synthesis in vivo (18). Subsequent studies indicated that increased cortical AMT uptake can occur either because of increased rate of serotonin synthesis or of induction of the kynurenine pathway (19,20). Increased cortical AMT uptake appears to be highly specific for epileptogenic regions, although its sensitivity depends on the underlying pathology: it was found to be most sensitive (≤70%) in children with tuberous sclerosis or other developmental cortical malformations, and less sensitive (<40%) in cryptogenic epilepsy (15–17,21). A recent study (22) also showed that increased AMT uptake can lateralize the seizure focus in patients with temporal lobe epilepsy without hippocampal atrophy, thus further extending the clinical potential of this imaging method.
The goal of the present study was to determine the usefulness of AMT-PET in patients with recurrent seizures after resective epilepsy surgery. We hypothesized that increased AMT uptake will identify epileptogenic cortex in at least some patients with previously failed cortical resection. We also evaluated which factors (such as age, etiology, or timing of the PET scan) may help predict which patients will most likely benefit from AMT-PET, when reoperation is being contemplated after failed neocortical resective epilepsy surgery.
- Top of page
- SUBJECTS AND METHODS
This study demonstrates that AMT-PET can identify additional epileptogenic cortical areas and assist successful reoperation in patients with a previously failed cortical resection. When performed after the acute postsurgical period, but within 2 to 3 years after surgery, AMT-PET was able to identify well-localized cortical increases in ∼40% of the cases. These increases were concordant with ictal EEG localizations, and removal of the affected areas resulted in sustained seizure freedom or considerable improvement in all cases. The results are consistent with two recent studies documenting that increased cortical AMT uptake can detect epileptic regions in patients with neocortical epilepsy (16,17). Similar to these previous studies, the present findings indicate that abnormal increases of cortical AMT uptake are highly specific for epileptogenic brain regions and can occur in patients with and without structural lesions. Consistent with findings on patients with no previous resection (17), focal increase of AMT uptake did not appear to be related to epileptiform activity (including electrographic seizures) recorded during the tracer-uptake period, or to the time elapsed since the last clinical seizure before PET scanning. Thus it is unlikely that in these patients, increased AMT uptake in the epileptic focus was due to ictal or postictal activity, or even to active interictal spiking.
In addition to demonstrating the clinical utility of AMT-PET after failed cortical resections, our results provide clues regarding the optimal timing of postsurgical PET scanning. As demonstrated in two of our patients, early (within a week, or possibly even beyond) postoperative PET scanning with AMT may show widespread nonlocalized hemispheric increases, presumably due to transient postoperative inflammatory reactions. The reason for very high AMT uptake in such cases most likely stems from activation of the kynurenine pathway of tryptophan metabolism (27). This can lead to overproduction of toxic and convulsant tryptophan metabolites, particularly quinolinic acid (19,27). Previous studies showed increased levels of quinolinic acid in brain tissues resected from cortical regions showing increased AMT uptake in patients with epilepsy associated with tuberous sclerosis or brain tumors (19,20,28). Tissue quinolinic acid levels can increase under certain pathologic conditions, via induction of the key enzyme indoleamine 2,3-dioxygenase (by infections, viruses, or interferons), leading to significant metabolism of tryptophan along the kynurenine pathway (27,29,30). If the cortical increases on AMT-PET scans taken shortly after cortical resection reflect this pathway, it raises the possibility that increased quinolinic acid (or other convulsant tryptophan metabolites) levels may contribute to early postoperative seizures in such cases. Early postoperative seizures can occur in as many as 25% of patients after pediatric epilepsy surgery and, although these early seizures predict worse long-term outcome, approximately half of these patients become and remain seizure free in the prolonged postoperative period (31). It remains unclear, but would be useful to explore, whether similar widespread increases of AMT uptake occur shortly after resection in patients with no postoperative seizures.
AMT-PET did not provide localization information in nine patients in whom the PET scans were performed beyond 2.3 years after resection, although all of these patients had intractable seizures. The reason for the low sensitivity of AMT-PET in these patients remains unclear, and the limited number of patients precludes general conclusions regarding the clinical usefulness of AMT-PET performed beyond 2-3 years after epilepsy surgery. Nevertheless, our findings suggest that AMT-PET can be very helpful to localize epileptic foci in almost half of the patients with previously failed neocortical resection, if the PET is performed beyond the acute postsurgical period but within 2–3 years after resection. This finding represents an advance in seizure-focus localization in patients who have had a previously failed cortical resection. Currently no reliable imaging methods can detect missed epileptogenic cortex in these patients, and one must rely heavily on scalp-EEG data to guide subdural electrode placements, an approach that is not optimal.
The biologic basis of increased AMT uptake localized to the epileptic focus remains unclear. In the present study, increased quinolinic acid concentrations were found in the resected epileptogenic brain tissue of one reoperated-on patient with localized increase of AMT uptake. This is consistent with similar findings in children with tuberous sclerosis and intractable epilepsy (18). However, increased AMT uptake can also result from increased serotonin synthesis in or around the epileptogenic tissue. In support of this possibility, human epileptic tissue studies showed serotonergic hyperinnervation in dysplastic tissue but not in cortex from patients with cryptogenic epilepsy (32). Our recent PET study also found a significant association between presence of increased AMT uptake and cortical developmental malformations (17). In the present study, the number of patients with developmental abnormalities was too low, and such an association could not be confirmed. To address the role of serotonergic mechanisms in epileptogenic brain regions further, PET studies using a tracer for 5-HT1A receptors have recently become available (33).
Another important issue for further study is whether resection of cortex with increased AMT uptake leads to seizure-free outcome or, conversely, nonresection of it predicts seizure recurrence after surgery. Our previous study with intracranial EEG correlations demonstrated that cortical areas immediately adjacent to increased AMT uptake are often epileptogenic and should be addressed by subdural EEG monitoring. Analysis of surgical outcomes suggested that the best surgical results occurred when the resection included both cortex with increased AMT uptake and the surrounding epileptogenic region (17). With this consideration, increased AMT uptake can be a powerful guide for grid placement. In the present study, all but one reoperation in patients with increased AMT uptake was guided by intracranial EEG monitoring, and regions including the AMT-PET abnormality as well as intracranial seizure onset were resected. This resulted in seizure freedom in five patients and has led to a considerable improvement in the remaining two. One of these three patients had a hemimegalencephaly with mostly posterior quadrant involvement, and focal resection rather than hemispherectomy was performed in an attempt to spare the functioning motor cortex. A recent study described a series of 11 patients with a similar malformation (called “hemi-hemimegalencephaly”), in whom temporoparietooccipital resections (but not occipital or parietal resection alone) decreased but did not always abolish seizures (34). It is likely that, in such cases, complete, sustained seizure freedom could be achieved only by hemispherectomy. Nevertheless, all patients in the present series benefited greatly from reoperation assisted by AMT-PET. Thus the clinical utility of AMT-PET can be extended to the most challenging patients with previously failed neocortical resection; with proper timing, accurate focus localization facilitating successful reoperation can be achieved by applying AMT-PET in almost half of such patients. Conversely, the fact that more than half of patients did not show increased AMT uptake indicates that much work remains to be done in developing other neuroimaging approaches to define their remaining epileptogenic cortex.