Abstract
Maillard LG, Tassi L, Bartolomei F, Catenoix H, Dubeau F, Szurhaj W, Kahane P, Nica A, Marusic P, Mindruta I, Chassoux F, Ramantani G. Ann Neurol 2017;82:781–794.
OBJECTIVE: We aimed to (1) assess the concordance between various polymicrogyria (PMG) types and the associated epileptogenic zone (EZ), as defined by stereoelectroencephalography (SEEG), and (2) determine the postsurgical seizure outcome in PMG-related drug-resistant epilepsy. METHODS: We retrospectively analyzed 58 cases: 49 had SEEG and 39 corticectomy or hemispherotomy. RESULTS: Mean age at SEEG or surgery was 28.3 years (range, 2–50). PMG was bilateral in 9 (16%) patients and unilateral in 49, including 17 (29%) unilobar, 12 (21%) multilobar, 15 (26%) perisylvian, and only 5 (9%) hemispheric. Twenty-eight (48%) patients additionally had schizencephaly, heterotopia, or focal cortical dysplasia. The SEEG-determined EZ was fully concordant with the PMG in only 8 (16%) cases, partially concordant in 74%, and discordant in 10%. The EZ included remote cortical areas in 21 (43%) cases and was primarily localized in those in 5 (10%), all related to the mesial temporal structures. All but 1 PMG patient with corticectomy or hemispherotomy had a unilateral PMG. At last follow-up (mean, 4.6 years; range, 1–16), 28 (72%) patients remained seizure free. Shorter epilepsy duration to surgery was an independent predictor of seizure freedom. INTERPRETATION: PMG-related drug-resistant epilepsy warrants a comprehensive presurgical evaluation, including SEEG investigations in most cases, given that the EZ may only partially overlap with the PMG or include solely remote cortical areas. Seizure freedom is feasible in a large proportion of patients. PMG extent should not deter from exploring the possibility of epilepsy surgery. Our data support the early consideration of epilepsy surgery in this patient group.
Commentary
Defining the extent of surgical resection in patients with large lesions is a challenge. The predicament is between the acceptance that optimizing surgical outcomes requires removing the totality of the lesion, and the belief that limiting functional complications requires the smallest resections possible. Recent data challenge us to think beyond both of these convictions.
Achieving Seizure Freedom
At its core, achieving seizure freedom requires the removal of all brain tissue responsible for initiating and fully organizing the patient's seizures. When epileptic tissue is completely removed, seizures stop. Necessitating the complete removal of a lesion to achieve seizure freedom inherently then implies that the totality of the lesion, any lesion, is epileptic. However, data linking the completeness of resection with favorable seizure outcomes after surgery have been conflicting (1, 2). These discrepancies have always been attributed to variations in study design, surgical strategies, or some other technical explanations. Let's consider here an alternative hypothesis: not all lesions are created equal. In the study by Maillard et al. the authors evaluate the surgical outcomes of a large cohort of patients with a specific type of malformation of cortical development (polymicrogyria), relying on their centers’ large experience with stereo EEG (SEEG) explorations to investigate these multilobar and often bilateral lesions. Let's put this in context. In the most purist SEEG tradition, the anatomoelectro-clinical syndrome first and foremost drives the surgical strategy, the constellation of clinical semiology and electrical findings, without consideration of the lesion until after the epilepsy localization hypothesis is defined. When Bancaud and Talairach (3) were performing the first SEEG evaluations in the 1950s, all surgical planning was “lesion agnostic”: no preoperative neuroimaging was possible at the time, so the electrode implantation maps were generated purely based on semiology and EEG. This abstract separation between developing a surgical strategy and knowing the size or presence of a lesion was essential in allowing the SEEG centers involved in Mail-lard et al. to even conceptualize incompletely resecting the polymicrogyria as an acceptable surgical strategy. The SEEG-determined epileptogenic zone was fully concordant with the polymicrogyria in only 16% cases, partially concordant in 74%, and discordant in 10%. Inclusion of the mesial temporal structures in the epileptogenic zone, even when remote from the polymicrogyria, was often seen. Resections based on the SEEG-determined epileptogenic zone lead to seizure freedom in 72% of patients at last follow-up (mean of 4.6 years). One should note that these remarkably optimistic results were only seen in patients with unilateral polymicrogyria. In this specific pathology (polymicrogyria), the epilepsy may be restricted to a specific localization within a much larger lesion. This is in stark contrast with studies that have looked at other pathologies where removing the totality of the lesion was essential to achieve seizure freedom. These pathologies include type II cortical dysplasia (4), developmental tumors (5), cavernomas (6), or even the Maillard et al. patients with bilateral polymicrogyria where only one of nine patients was rendered seizure free by the SEEG-guided focal resection. Another recent example of epilepsy delicately overlapping with the lesion is depth of the sulcus dysplasia where invasive EEG often confirms ictal onset from the lesion, and a very focal resection of the lesion is enough to achieve seizure freedom (7).
The point is that the organization of the epileptogenic zone depends on pathology: in certain situations, like polymicrogyria, a distinction between “lesion” and epilepsy must be made (ideally with SEEG), whereas in other situations, like depth of the sulcus dysplasia, most malformations of cortical development (MCD) 2, developmental tumors, and others, removing the totality of the lesion is necessary.
Maintaining Seizure Freedom
To achieve seizure freedom, we have to remove the tissue that is causing the current seizures. For that, the considerations above related to completeness or incompleteness of resection are sufficient. However, maintaining seizure freedom is in a completely different ballpark. Seizures that recur after months or even years of postoperative seizure freedom in patients who were having frequent seizures before surgery are unlikely to be attributed to reactivation of an incompletely resected epileptic focus.
Alternatively, one should at least consider the possibility that some seizure recurrence can be attributed to postoperative maturation or development of epileptic tissue. For seizure freedom to be then maintained—and not simply achieved—we need two of the following strategies: 1) to either remove all tissue that has the potential to develop epilepsy in the future, or 2) to develop a mechanism to prevent this postoperative epileptogenesis. Data from Maillard et al. and others (8) linking short epilepsy duration with favorable seizure outcomes support the first strategy, and urge us to aggressively approach the surgical treatment of patients with drug-resistant epilepsy as soon as possible. When epilepsy is allowed to go on unchecked for a long period of time, secondary epileptogenesis, particularly in the hippocampus, will expand the extent of potentially epileptic tissue and limit the chances for success with a small resection. It is not surprising that the hippocampus was actually the most common “remote” epileptogenic zone in this polymicrogyria series. This line of thinking can similarly explain the more favorable surgical outcomes with removal of the hippocampus in patients with temporal lobe cavernomas or low-grade developmental tumors (6). In cases with longstanding epilepsy, a large resection now may actually prevent a much larger resection in the future. A “remote” extent of the epileptogenic zone in multiple epileptic pathologies is actually supported by recent volumetric imaging data showing multifocal and extensive gray matter volume loss, mirroring the extent of the epileptic networks (9). As connectivity measures gain in sophistication and accuracy, we are getting closer to measuring the extent of the epileptogenicity. We have speculated that some inherent patient characteristics, such as genetic predisposition, may underlie some of the mechanisms of postoperative epileptogenesis, but this hypothesis still needs further exploration (10).
Epilepsy is a disease of electrical networks in the brain. Surgery removes a piece of the brain. It is therefore no surprise that seizure freedom after surgery depends on a lot more than the piece that was removed.