19 Tailored Temporal Lobectomy Techniques
Temporal lobe epilepsy is a common cause of intractable epilepsy. In adults, temporal lobectomy is the most common surgery and has been proven effective with class I evidence for its efficacy.1 The disease in children has some features different from adults. In particular, the finding of mesial temporal (hippocampal) sclerosis is less common in children than in adults.2,3 Further, even when present in histopathology, sclerosis is evident on magnetic resonance imaging (MRI) in only approximately half the instances.4 The increased incidence of dual pathology5,6 and dysplasia, which can involve medial, anterolateral, or basal temporal structures, argues for an electrophysiological approach to temporal lobectomy that combines anatomical boundaries with electrophysiological information. Thus, a wide range of temporal lobectomy may be indicated—from highly selective hippocampal/basal temporal resections to full temporal lobectomy7 ( Fig. 19.1 ).
A tailored resection can refer to modification of the resection (larger, smaller, or different location from a standard anterior temporal lobectomy or amygdalohippocampectomy) based on interictal measurements or modification based on functional information, in particular language mapping.8–11
In adults with mesial temporal abnormalities, intraoperative electrocorticography (ECoG) has been used to guide the extent of hippocampal resection,12,13 and the presence or absence of residual epileptiform activity (e.g., spikes) has been associated with postoperative seizure control.13 Postoperative spikes outside the medial temporal lobe may be of less value,12 although some reports have found better outcome when neocortical spikes are eliminated with resection.14,15
Tailoring occurs de facto when invasive monitoring is used. The prolonged observation of interictal abnormalities and ictal onset will determine the resective strategy. This information can be augmented by intraoperative recordings. In this chapter, the strategies for using intraoperative ECoG are discussed, methods for hippocampal recordings are described, and approaches to functional considerations are reviewed.
Strategies
The advantages to tailoring a resection are seen in several circumstances. First, in the case of a clear lesion, intra-operative ECoG can identify particularly epileptogenic areas of cortex surrounding the lesion, and such an approach has been associated with excellent outcomes.15–19 A lesion can co-exist with functional cortex, even if grossly abnormal,20,21 and adjacent brain is, of course, vulnerable to functional loss; therefore, particularly in suspected language areas, a functional map is needed as well.
Other situations in which tailoring is useful are cases of suspected dual pathology. Two approaches can be taken. One option is to perform invasive monitoring, with electrodes covering both the suspected or evident lesion and other electrodes covering the medial temporal lesion. Either subtemporal electrodes placed medially or depth electrodes into the hippocampus should give similar results.22 Alternatively, if the lesions are in the same functional-anatomical region (e.g., basal and medial temporal lobes or anterior temporal lobe), intraoperative recordings can guide the relevant extents. Areas outside the hippocampus have been implicated in the genesis of anterior temporal lobe seizures,23 and their interictal abnormalities can be identified during intraoperative recordings ( Fig. 19.2 ).
Functional Considerations
Another role for tailoring is in the setting of minimizing resection to avoid postoperative deficits. In lateral temporal cortex, this may involve mapping language and avoiding, by at least 1 cm, areas where stimulation disrupts naming tasks.24 In dominant temporal lobe, this may involve mapping language8 and even memory,25 because lateral temporal cortex resection has been associated with memory difficulties,25 although the presence of a normal hippocampus on MRI and good preoperative memory are very strong risk factors for postoperative memory decline26 that may be impossible to escape.
Functional information can often be determined without direct cortical mapping. The location of language areas is similar in adults and children, although younger children may have a more limited distribution of language sites than adults.27 Thus, a preoperative assessment of the risk to language can be judged. Functional MRI (fMRI) continues to hold promise as a method to determine language sites non-invasively, but it remains a challenging tool.28,29 We have found language fMRI can be, with practice, performed to as young as age 6. Silent action generation from both pictures and words can be effective ( Fig. 19.3 ). The cerebral amytal (Wada) test may be necessary when features are present such as left-handedness, seizure semiology, or incongruent neuropsychological testing or when a proposed resection would be more radical if atypical dominance were confirmed. The Wada test is difficult in younger patients, and only very skilled teams are likely to be successful with children younger than age 12.
When direct cortical mapping is desired intraoperatively, preoperative practice is all the more critical in younger children. Below age 12 in the operating room may be impractical, but the use of dexmedetomidine has been described in very young children30 and appears to have minimal effect on the ECoG.31 Any anesthetic agent will influence the ECoG; benzo-diazepines in particular should be avoided, if possible. propofol is extensively used because of its effects on the depth of suppression of cortical activity, potentially masking epileptiform activity in the ECoG, can usually be titrated rapidly.32