Radiofrequency lesioning for epileptogenic periventricular nodular heterotopia: A rational approach


Address correspondence to Friedhelm C. Schmitt, Department of Neurology, University of Magdeburg, Leipziger Str. 44, 39120 Magdeburg, Germany. E-mail:


Periventricular nodular heterotopias (PNHs) are frequently associated with pharmacoresistant epilepsy. They are considered part of a dysfunctional network, connected to the overlying cortex. Therefore, removal of the PNHs and additional cortectomy or lobectomy seem to be essential for significant and long-lasting seizure reduction. These procedures, however, can have considerable limitations, especially in patients with functional eloquent cortex adjacent to the PNH. Alternatively, stereotactic neurosurgery can reduce the surgical trauma. Presented is a 56-year-old man who became seizure-free after stereotactically guided radiofrequency lesioning of a solitary PNH.

Pharmacologic treatment of patients with periventricular nodular heterotopias (PNHs) is oftentimes unsatisfactory (Raymond et al., 1994), even if the extent of PNH is limited to one brain region and neurologic examination is normal (D’orsi et al., 2004). Standard resective neurosurgery also does not necessarily result in cessation of seizures (Li et al., 1997). In case of adjacent eloquent cortex, this approach must be abandoned or is further complicated by the need for invasive electroencephalography (EEG) monitoring, such as implementation of grid electrodes. As far back as 60 years ago, stereotactically guided radiofrequency lesioning (RFL) using depth electrode recordings in the rhinencephalon was thought to circumvent this problem (Marossero et al., 1980). Generally accepted indications and target zones for this procedure, however, have been missing since (Ojemann & Ward, 1975).

The presented case illustrates that RFL can be used to deactivate effectively an epileptogenic zone associated with PNH if it is performed in combination with high-resolution magnetic resonance imaging (MRI) and electrophysiologic selected target detection.

Patient and Method

On admission, this 56-year-old, right-handed male patient presented with poorly controlled epilepsy manifested as frequent unspecific auras, daily clonic, aphasic seizures and one bilateral tonic–clonic seizure per week. In addition, he complained about memory and concentration difficulties. Current treatment was lamotrigine 850 mg/day and carbamazepine 800 mg/day.

Since onset of epilepsy at the age of 3, several antiepileptic drugs have been used. From childhood to adolescence, the patient took phenobarbital, phenytoin, and valproic acid. Later oxcarbazepine and topiramate treatment was discontinued because of poor seizure control and intolerable side-effects. Birth and child development was unremarkable other than for measles encephalitis in early childhood. A third-degree relative of the patient allegedly had epilepsy. He had no other affliction apart from medically controlled hyperlipidemia.

The main findings on cranial MRI (Fig. 1A) were a solitary PNH dorsal to the left caput caudatum at the level of the ventricular wall and atrophy of the left-sided superior insular cortex and the cerebellum. Moreover, fluid-attenuated inversion recovery (FLAIR)–MRI sequences indicated possible connection of the PNH to the insular cortex. During scalp video-EEG monitoring we documented five habitual seizures: Three unspecific auras (type I) did not show a seizure pattern. One clonic seizure (type II) in the right half of the face was preceded with a habitual aura (type I). The same aura was followed by an aphasic seizure (type III), which secondarily generalized. During each of the latter two seizures, an ictal EEG pattern beginning in the left frontotemporal region was recorded. In addition to the right sided cloni and the aphasia, a “sign of four” lateralized clinically to the left hemisphere. To evaluate the epileptogenicity of the PNH we implanted three electrodes (AD-TECH, Racine, WI, U.S.A.) in the left frontal lobe. The electrodes were temporarily externalized for continuous depth electrode recording. Using computed tomography (CT)/MRI–guided and computer-assisted treatment planning software (Voges et al., 2002) the trajectories were deliberately arranged to obtain three-dimensional (3D) information on seizure extension. In addition, single electrode contacts of the 8-contact or 12-contact electrodes, respectively, were placed either inside or in close vicinity to the PNH (Fig. 1B).

Figure 1.

(A) T1-weighted MRI illustrating the preoperative status with the PNH near the ventricular wall and its connection to the insular cortex. (B) T1-weighted MRI illustrating individual electrode position in relation to the PNH (white arrow). Locations of electrode A (green), electrode B (yellow), and electrode C (red). Note the proximity of electrode B and C to the PNH. (C) Definition of the tissue volume for RFL (red) on coronal T1-weighted MRI coregistered in the stereotactic coordinate system. (D) T1-weighted MRI illustrating surgical approaches for stereotactically guided RFL at different target points in projection onto the volume of interest (yellow) containing the PNH. (E) T1-weighted MRI illustrating the result 6 months after RFL. A small area of the heterotopia has deliberately not been coagulated in order to not disrupt the ventricular wall by the RFL procedure.

During invasive video-EEG monitoring we obtained two of the three seizures already recorded during scalp video-EEG (type I and type III). The isolated aura started with increased clusters of low-amplitude fast activity (Fig. 2A) at the tip of electrode C. The ictal EEG pattern of the aphasic seizure began with localized polyspike activity at the tip of electrode C with a propagation to a diffuse seizure pattern within several minutes (Fig. 2B), also detectable on the surface EEG. The regular detections of the ictal EEG pattern onset (type I and type III) from electrode C, with subsequent propagation to the overlying cortex, supported the hypothesis of the PNH as the relevant structure responsible for the patient’s seizures. In this respect the aphasic semiology of seizure type III represents the clinical manifestation of the seizure pattern’s propagation in the overlying cortex. Because partial cortectomy of the insula was not possible, we offered to the patient stereotactically guided RFL of the PNH, which was performed in general anesthesia. Using special software for brachytherapy of brain tumors (Precisis Brachytherapy Tool, release 3.1, Heidelberg, Germany) we delineated the PNH and its connection to the frontal cortex on axial MRI (proton-weighted-T2-weighted and T1-weighted sequences) as target volume (0.3 ml) (Fig. 1C,D). Taking into consideration the volume of ablated tissue resulting from heating for 60 s with 80°C (cone with the dimensions 4 × 6 mm) when a particular bipolar RF probe is used (TCB013, Neurostimulator N30; Inomed, Freiburg, Germany), we considered each coagulation as virtual radioactive source defining then the number of trajectories and treatment points necessary to fully cover the target volume. In total, we applied 20, partially overlapping coagulations using eight trajectories (Fig. 1D). A single, already existing burr-hole trephination (8 mm) was used to gain access for the different target points. The postoperative course was uncomplicated. As expected, immediate postoperative MRI displayed a circumscribed coagulation necrosis (volume 0.4 ml), which was surrounded by a localized, non–space-occupying brain edema.

Figure 2.

White arrows indicate electroencephalographic manifestations; black arrows indicate clinical seizure onset. (A) Intracerebral recording of seizure type 1 with repetitive bursts of low-amplitude–fast-activity beginning before clinical onset predominantly at electrode contacts C6–C7 and B10–B11 (both near PNH). During the clinical manifestation LAFAs are predominantly at electrode contacts C2–C3. (B) Intracerebral recording of seizure type 3, which secondarily generalizes: 15 s before clinical onset, there is a diffuse seizure pattern slowly spreading toward the cortex. This seizure pattern is preceded by polyspike activity at electrode contacts C6–C7, located near PNH.


Three months after RFL there was one “major” seizure after accidental nonadherence to antiepileptic medication. The patient also reported regular occurrence of approximately one habitual aura per week. Six months after surgery seizure frequency further reduced to one habitual aura every second week. After 1 year the patient reported <1 aura per month. Therefore, seizure outcome was classified as Engel class Ib (seizure free besides isolated aura). With medication of lamotrigine 900 mg/day, the patient did not complain anymore about major cognitive deficits. Follow-up MRI displayed a small defect zone with a medially located fragment of the PNH (Fig. 1E).


The patient presented in this case report had a solitary PNH and intractable epilepsy, and experienced significant improvement of seizure activity after stereotactically guided RFL of the epileptogenic focus. We argue that the solitary PNH of this patient is the dominant epileptogenic structure and conclude that minimally invasive lesional stereotactic surgery is a highly effective treatment option minimizing the risk for postoperative deficits.

The extent of PNHs has a great variability and there is only a loose relationship to the manifestation of seizures (Raymond et al., 1994). In general, clinical (Wagner et al., 2009), histologic (Hannan et al., 1999), electrophysiologic (Kothare et al., 1998; Scherer et al., 2005), and functional MRI studies (Dubeau & Tyvaert, 2010) imply a widespread network connecting PNHs to distant cerebral structures. In addition, PNHs can apparently participate in higher cerebral functions (Wagner et al., 2009), traditionally thought to be restricted to the cortex. There is, however, a considerable dispute about the role of the cortex and the PNH in terms of epileptogenicity: It is discussed that seizure activity starts from the PNH (Kothare et al., 1998; Scherer et al., 2005) or the PNH is part of a malfunctioning network comprising of the PNH and the overlying cortex (Dubeau & Tyvaert, 2010).

In the presented case, seizures consisted either of auras (type I) or were preceded by an aura (type II and III). All seizure patterns started in proximity to the heterotopia according to depth electrode recording. The fact that seizure type II could not be obtained during depth recording is in our view negligible, since it was regularly preceded by a habitual aura and the surface EEG had previously shown the same ictal pattern, as in seizure type III. Therefore, the electroclinical findings confirmed a neuronal connectivity between the PNH and overlying cortex, as morphologically indicated on MRI.

Notably, the patient still experienced habitual auras in a reduced frequency after RLF. Concerning the role of epileptogenesis of the PNH, three mechanisms appear plausible:

  • 1 Prior to surgery, the PNH acted as a pacemaker (Kothare et al., 1998; Scherer et al., 2005) and the pathologic network enabled the ictal activity to propagate to the overlying cortex. After RFL, the residual fragment has further pacemaker function to a limited extent.
  • 2 The PNH is essential for amplifying (Dubeau & Tyvaert, 2010) the ictal activity inside a given network. RFL has disabled this epileptogenic network so that cortically initiated auras can no longer propagate.
  • 3 A participation of PNH in experiencable cerebral function has been reported (Wagner et al., 2009). It seems conceivable that the postoperative auras in the presented case originate from the remaining fragmented PNH.

The analysis of the temporal evolution of the seizure pattern with simultaneous intracerebral and cortical recording could have resolved this issue. Nonetheless, there was no straightforward indication for this procedure for our patient, because resective surgery was not feasible.

In a series of 41 patients with pharmacoresistant epilepsy of various etiologies, stereo-EEG-guided RFL of the epileptogenic zone led to a responder rate of 48.7% (at least 50% seizure reduction) (Catenoix et al., 2008). Similar to our patient, the two patients with PNH in this series became seizure-free in a delayed time course. A delayed treatment response has been a longstanding clinical observation after RFL for epilepsy patients and a similar observation was coined as “running down phenomenon” in resective epilepsy surgery.

The good clinical outcome supports the hypothesis that the PNH—as the dominant epileptogenic structure—and its vaguely visualized connection to the cortex were valid targets for RFL by deactivating the PNH. Scherer et al. (2005) have published a case report that supports the primary role of a solitary PNH both with the beginning of the ictal activity in PNH and seizure freedom after resective surgery. In comparison to resective surgery, RFL has the advantage of preserving the overlying cortex in toto. Moreover, because of its minimal invasiveness this technique can be applied repeatedly as a procedure in stages, which could be a strategy for patients with multiple PNH. In addition, RFL does not render later resective surgery impossible. It could, therefore, be considered as first-line treatment in appropriately selected patients. Yet, a sufficient number of depth recordings near a solitary PNH is advised (Scherer et al., 2005), as long as the role of PNHs for the epileptogenicity is not fully understood.




None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.