Resections for Extratemporal Epilepsy

Semiology and Diagnostic Work-up

Although TLE remains the most common form of focal epilepsy, frontal lobe epilepsy (FLE) is the second most common and often also amenable to surgical resection to aid in seizure control. ^, The semiology of FLE is very diverse, and localization can be very challenging. Considerable overlap exists, but current literature has attempted to divide FLE into three main subtypes: partial motor, supplementary motor, and complex partial seizures (CPSs). ^, Partial motor seizures are characterized by focal clonic activity; however, consciousness is typically preserved in this subtype. With supplementary motor seizures, patients often experience speech arrest and asymmetric tonic posturing. These findings can often be seen in conjunction with partial motor seizures, suggesting spread of the electrical activity into the supplementary motor area (SMA). ^, Finally, CPSs in FLE often include staring spells, vocalization, depressed consciousness, and some bimanual or bipedal movements. ^,

Seizures arising from the frontal lobe are often short lived, and, as stated earlier, typically involve some complex motor behaviors or emotional symptoms. These seizures can be difficult to describe, especially in comparison to seizures arising from the temporal lobe, which have symptom patterns that are very well recognized and easy to identify. In addition to the difficulty in recognizing symptomatology, the frontal lobe is connected to other distant portions of the brain through a large network of efferent pathways. This often leads to widespread propagation of seizure activity to other sections of the frontal lobes or other lobes entirely. Although this connectivity can make seizure localization on EEG difficult, it is not the only factor that contributes to this challenge. The frontal lobe is the largest lobe of the human brain, accounting for approximately 35% to 45% of the cortical volume. Much of the cortical volume is buried, especially in the ventromedial portion of the frontal lobe. Because of the location of this cortical tissue, the use of scalp EEG and even electrodes placed over the cortical convexity directly do not often detect seizure activity in this deep tissue with any level of reliability. These factors are a large portion of the reason why surgical treatment of FLE leads to poorer outcomes when compared with other lobes.

It is easy to see that defining the onset zone of seizures in FLE has proven to be difficult for a number of reasons. However, with the increased use of stereo-electroencephalography (SEEG), the deeper areas of the frontal lobe have become accessible for electrical recordings ( Fig. 95.1 ). One group used information obtained from SEEG in addition to clinical symptoms to delineate the frontal lobe into groups. Group 1 seizures originated from the precentral and/or premotor regions. In this group, seizures were characterized by elementary motor signs (defined as tonic, clonic, or dystonic movement, as well as posturing) without evidence of gestural motor behavior (defined as more complex behaviors such as rhythmic movements, vocalization, autonomic signs, emotional expression, and so on). Group 2 seizures originated from the overlapping premotor and posterior prefrontal regions. Nonintegrated gestural motor behavior with some proximal tonic posturing and facial contraction are noted in this group. Finally, groups 3 and 4 both manifest as gestural motor behavior without elementary motor signs. However, the area from which these seizures originate is the differentiating factor. Group 3 seizures were found to arise from the lateral frontal cortex or frontal pole. In addition, seizures in this group were noted to be devoid of emotional content. On the other hand, group 4 seizures were found to arise from the ventromedial prefrontal cortex. Fearful emotional behavior was noted more frequently in this group. Although these groups nicely organize frontal lobe seizures in an anteroposterior direction, they give very little information, except for groups 3 and 4, about medial to lateral location.

Surgical Management

When discussing surgical treatment in the setting of FLE, the goal is to remove the seizure onset zone. Semiology can be helpful, but it is often not entirely indicative of the zone of seizure onset. Thus, before proceeding with surgical resection, care should be taken to identify the area of onset for frontal lobe seizures. Whereas frontal lobectomy is certainly feasible, resections in the frontal region can often be tailored to a more targeted resection. Broadly, frontal lobe resections can be divided into those involving the medial frontal lobe and those involving the lateral frontal lobe. The surgical approach for these locations varies, so it is important to distinguish between the two.

For medial approaches to the frontal lobe, a bicoronal incision is often required to achieve appropriate access. With this approach, access can be obtained to the SMA, the pre-SMA, and the anterior portion of the cingulate. When an interhemispheric approach is used to reach these regions, it must be noted that the falx does extend down to the corpus callosum, but instead stops at the superior border of the cingulate gyrus. Therefore dissection of the arachnoid adhesions in this region as well as protection of the en passage vessels, such as pericallosal and callosomarginal arteries, must be undertaken with great care. Anatomically, the posterior border of the SMA is the precentral sulcus, and the inferior border is the cingulate sulcus. These borders are well-defined anatomic landmarks; however, the anterior border is not very well defined. Understanding the basic anatomic region of the SMA is of critical importance because patients who undergo resections involving these regions may awake with an “SMA syndrome” that is likely transient in nature. This syndrome manifests as a deficit in volitional movement of the contralateral limbs, while maintaining a normal or increased level of tone. In addition, mutism is noted, most commonly when the dominant hemisphere is affected. These symptoms can last for weeks after surgery. In an effort to prevent this during surgery, motor stimulation can be used with grids to guide your resection. Postoperatively, an MRI should be obtained to ensure that no infarction has occurred. Although SMA syndrome may occur after surgery in the medial frontal region, it is of paramount importance to differentiate this finding from that of direct injury to the motor cortex, as these patients are less likely to improve from a motor standpoint.

Whereas medial approaches to the frontal lobe can be somewhat difficult, lateral approaches can often be performed through a more standard pterional or frontal craniotomy. These exposures provide access to both the anterior and the posterior lateral frontal lobes, respectively. When performing a resection of the lateral frontal regions in the dominant hemisphere, language mapping is typically required to define the Broca area. This can be done through an awake craniotomy or functional MRI prior to surgery. There is some variability with regard to the location of the Broca area, with most series noting this to be in the precentral gyrus or the pars opercularis. However, some have reported findings in the pars triangularis. In addition, as many of these patients have lesional findings that have been present for some time, the standard anatomic locations for the Broca area may be inaccurate, thereby further necessitating language mapping.

Outcomes

FLE epilepsy is treated surgically by way of either complete or partial frontal lobectomy. Unfortunately, seizure freedom rates (approximately 50%; Table 95.1 ) do not reach the same threshold as that of TLE (approximately 70%). ^{, ,} Although in some instances a focal lesion such as a tumor or cortical dysplasia is identified in FLE, imaging and EEG are often unrevealing with regard to a source of seizure onset. This obviously presents a diagnostic and surgical dilemma. Many patients with FLE undergo extensive presurgical work-up including SEEG and more advanced imaging techniques such as positron emission tomography (PET), single-photon emission computed tomography (SPECT), or magnetoencephalography (MEG). In addition to the difficulty with localization, some areas of the frontal lobe do not lend themselves to resection owing to concern for postoperative morbidity including the motor cortex, SMA, or expressive speech center (Broca area). These areas often limit the extent of resection and thus limit the ability to obtain seizure freedom. Many studies have attempted to predict long-term postoperative seizure freedom rates in FLE patients, but the diversity of these seizures from an anatomic, electrographic, and functional perspective leads to significant variability in long-term outcomes. ^{, ,} Rates of long-term seizure freedom in FLE may also be overestimated in some of the literature owing to inadequate length of follow-up. ^,

Several notable series have been published with regard to postoperative outcomes in cases of FLE. Jeha and colleagues published their outcomes in 70 patients who underwent frontal resection. At 1 year their seizure freedom rate was approximately 55%. Unfortunately, over time, their seizure freedom rate decreased to approximately 27% at 5 years. These investigators also looked at predictors of seizure recurrence, which included negative MRI, presence of extrafrontal pathologic features, generalized ictal EEG patterns, and incomplete resection. Lazow et al. reported a similar percentage of Engel class I patients at last follow-up (57%). In addition, this group was found to have favorable outcomes over a longer time period. A majority of patients in this study underwent some form of intracranial EEG monitoring (89%) prior to surgery, which was postulated to improve long-term outcomes. A meta-analysis by Englot and colleagues that included 21 studies found an overall seizure freedom rate of 45.1%. In this study, no change in seizure freedom rates were noted over time, which suggests that despite advances in neuroimaging, electrophysiologic, and surgical techniques, outcomes have not improved with regard to surgery for FLE. These investigators did, however, note that patients who were found to have a lesional cause of epilepsy or abnormal preoperative MRI findings or who underwent a focal frontal lobe resection had a better chance of achieving long-term seizure freedom. A more recent study noted younger age and shorter duration of epilepsy to also be predictors of a good outcome. Other studies have noted the same findings with regard to predictors of long-term seizure freedom in cases of FLE. ^{, ,} This highlights the importance of appropriate presurgical diagnostic work-ups in potential surgical candidates.

Semiology and Diagnostic Work-up

First described in the 19th century in a patient with a parietal-occipital lobe tumor and seizures with visual auras, occipital lobe epilepsy is less common than other extratemporal epilepsies. As the occipital lobe is the smallest lobe in the human brain, the reduced incidence of epilepsy in this region is understandable. Despite this, surgical intervention may provide significant benefit to those with this disorder.

Because the occipital lobe is the location of the primary visual cortex, visual function is important in both the semiology and management of occipital lobe epilepsy. Many studies have focused on the interesting semiology presented in these cases. Williamson and colleagues described the clinical characteristics of 25 patients with occipital lobe epilepsy. In this series, 60% experienced elementary visual hallucinations at ictal onset. This consisted of either flashing or steady light that was either white or colored in nature. The remaining 40% experienced amaurosis at the time of ictus. In a separate series, Kun Lee and colleagues found that 35% and 15%, respectively, experienced ictal symptoms as described earlier. Interesting to note, Jobst and colleagues published their findings from a series of 14 patients with intractable occipital lobe epilepsy. In contrast to the previous studies, only 50% experienced visual symptoms in this cohort. A majority of these symptoms were still elementary visual hallucinations. However, both the Kun Lee and Jobst groups noted an increased incidence of more formed visual hallucinations—21% and 27%, respectively. ^, These findings may be better explained in the context of the experience of Bien and colleagues. In this group of 20 patients, epileptic visual auras including elementary hallucinations, illusions, and visual field deficits were noted in those with occipital lobe epilepsy but also in those with seizures arising in the occipitotemporal region and the anteromedial temporal lobe. When the patient experienced complex hallucinations, these were always found to arise from the latter two areas, but never the occipital lobe alone. Finally, rapid bilateral blinking or eyelid flutter is common in occipital lobe epilepsy and may be helpful in directing attention to the occipital lobe when in search of the seizure onset zone. ^,

Understanding the spread of occipital lobe epilepsy is an important component of the diagnostic evaluation. These patients often experience complex partial epilepsy, which can appear very similar to that of temporal origin if there is a lack of visual symptoms associated with ictal onset. This is likely secondary to spread along the inferior longitudinal fasciculus. Spread along more superior pathways may lead to motor symptoms, similar to those seen in FLE. Understanding both the location of onset and the spread pattern associated with these seizures will presumably lead to improved surgical management and better long-term outcomes.

A number of underlying pathologic conditions, including tumor, cortical dysplasia, other developmental abnormalities, or glial scar, can be noted in the setting of occipital lobe epilepsy. In patients with classic semiology and a focal finding on imaging studies, less investigation is likely to be required in order to make a decision with regard to surgical management. With that being said, lack of a focal finding does not rule out the possibility of surgical intervention entirely. PET and SPECT, particularly with subtraction of the interictal imaging from the ictal imaging, as well as MEG can provide evidence for localization, or at the very least provide a rationale to continue with further diagnostic evaluation, such as SEEG ( Fig. 95.2 ). Unfortunately, scalp EEG in patients with occipital lobe epilepsy is often nonlocalizing in nature. Although the occipital lobe is anatomically small in comparison to other lobes in the brain, it does have deeper tissue that is difficult to evaluate with scalp EEG, which presents a likely explanation for its poor performance in these patients.

It is not surprising that visual field deficits are relatively common in this patient population. Documentation of a preoperative visual field test is an important part of the diagnostic work-up. If the patient already has a homonymous hemianopsia with structural and physiologic findings consistent with occipital lobe epilepsy, no further investigation may be required before proceeding with surgical management. On the other hand, if the patient does not have a focal pathology on imaging findings or has intact visual function, further investigation may be warranted. When considering intracranial studies in these patients, it is important to consider the goal of this intervention and the information that it could provide. In cases of negative imaging or a question of laterality, SEEG with multiple depth electrodes in the occipital lobes and other nearby regions may be of use. However, when the question regards the length of resection along the occipitotemporal or occipitoparietal border, subdural grid electrodes may be of some value. When image findings concerning for mesial temporal involvement are noted, resection of the hippocampus in addition to the occipital lobe focus should also be considered. In this case, depth electrode placement in the ipsilateral hippocampus at the time of SEEG plays a pivotal role in surgical planning. Finally, when considering occipital lobe resection in a patient with intact visual function, medial coverage of the calcarine cortex with either depth electrodes or an interhemispheric subdural grid electrode may be of value. Whereas it can provide important information about seizure onset, it also, through mapping, can provide information with regard to the localization of visual function.