The Role of Robots in Epilepsy Surgery Open Access

Epilepsy is a chronic neurologic disease that affects upwards of 50 million individuals globally [1]. Historically, it has been feared as demonic curses but also revered as a sacred disease, reflecting its stigmatizing aspects and complex cultural significance [2]. The treatment of epilepsy in the 20th century has mostly centered around anti-seizure medications and while they have greatly improved the course of the disease, approximately one-third of patients continue to suffer from seizures despite adequate pharmacological therapy [3, 4]. For these patients, epilepsy surgery offers the possibility of substantial seizure reduction and sometimes even a complete cure.

Epilepsy surgery – a cornerstone subfield of functional neurosurgery – has made great strides in the era of modern neurosurgery from early case reports and case series to large randomized controlled trial establishing its efficacy and role as a definitive treatment [5, 6]. It now encompasses a wide spectrum of surgical approaches: from invasive diagnostic modalities such as stereotactic EEG (SEEG), which helps determine the epileptogenic zone (i.e., the area of the brain that needs to be removed to abate seizures) to resective and disconnective procedures, as well as to neuromodulatory techniques, which can reduce seizure frequency without major tissue destruction.

In parallel to the increasing complexity of cases treated by many established epilepsy centers [7], surgical methods have progressed to provide safe and effective therapy options. One of many advancements but certainly a milestone in surgery has been the introduction of robotic technology [8]. Slowly but steadily even entire surgical steps have been reported to being performed autonomously by robotic machines [9]. The field of epilepsy surgery has been enriched by these technical innovations mostly through the use robotic guidance in stereotactic procedures. This narrative review explores the many current applications of robotic epilepsy surgery today and aims to provide an outlook of the innovations currently emerging.

Localizing the origin of seizure activity has long been the cornerstone for the presurgical evaluation of epilepsy. Accurate interpretation of seizure semiology, findings on scalp EEG recordings, and high-resolution imaging serve as the foundation for delineating the epileptogenic zone (EZ), which needs to be removed to free patients from seizures [10]. The resulting hypothesis can be further refined through neuropsychologic assessments, metabolic investigations, and detection of slight magnetoencephalographic alterations. Despite these advanced technologies, the full extent of the EZ can remain elusive and invasive EEG recordings are used to pinpoint the onset of seizure activity. These recordings can be critical in determining whether a patient is candidate for resective surgery and they can furthermore be used to tailor the surgical strategy.

Stereotactic EEG (SEEG), the currently most widely used method for invasive monitoring, was developed and pioneered by Talairach and Bancaud in the 1950s and 1960s [11, 12]. Talairach developed his own stereotactic frame, which allowed the placement of strictly orthogonal electrodes. Over time, technical adaptations of the method have streamlined the procedure, reducing the logistical demand on surgical staff thus increasing its use across the globe [13, 14]. Especially, the lower morbidity of SEEG over subdural electrodes led to its widespread adoption [15, 16]. Additionally, the development of frameless methods has decreased the duration of each operation and allowed more liberal planning of the trajectories including oblique trajectories [17‒20]. However, these advancements may have come at the cost of the procedure’s accuracy with errors margins more than twice of frame-based stereotactic methods [17‒19, 21]. To retain the benefits of a stereotactic frames but reducing the time needed for each trajectory adjustment, robotic solutions were developed.

Early reports of robotically assisted stereotactic systems for biopsies were published at the end of the last century [22, 23]. The first commercially available robotic system was the Neuromate system (Renishaw plc., Wotton-under-Edge, UK) introduced in 1987 and has received North American and European regulatory approval [24]. In the early 2000s, the first reports of SEEG series using Neuromate as a stereotactic tool holder were published [25, 26]. In 2007 another system, ROSA (Zimmer Biomet, Warsaw, USA) was released and has since become the most widely used and studied robotic platform for SEEG implantations with over 4,000 electrodes assessed for accuracy in the literature [27]. In the following years, multiple robotic systems with their own specific advantages and limitations were developed and are described in a recent review [24].

Over the past 2 decades, numerous reports on the accuracy of these systems have emerged. Because they can be used with or without a stereotactic frame, using intraoperative CT or MRI for registration, and rely on different navigational software, the results can vary between centers [28‒32]. Many publications show similar if not improved accuracy as compared to conventional frame-based operations with the average accuracy ranging between 1 and 2 mm [27, 30, 33‒36]. Additionally, the time needed for the implantation of each individual electrode is reduced by using robotic guidance [28, 30, 35]. This is especially advantageous in procedures with many planned electrodes and can easily offset the initial set up time of the robotic guidance system.

Innovations in robotic stereotactic surgery, which allow safe and fast SEEG implantation even without a strong foundation in stereotactic coordinate systems, have thus helped the method to gain wide approval globally. Especially in the USA, where electrophysiological explorations with subdural grids were the gold standard, we now see an increase of SEEG implantations [14]. Today the role of robotic SEEG is firmly established and current innovations mostly deal with making systems smaller, more user-friendly, and faster.

When a clear EZ has been defined, its removal can decrease the seizure frequency and even lead to seizure freedom. Traditionally, this has required open craniotomies and neurosurgical resections or disconnections of the affected brain region. However, in cases where the EZ is restricted to a small area and especially in deep seated lesions, ablations using radiofrequency thermocoagulation (RFTC) or laser interstitial thermotherapy (LITT) have become minimally invasive treatment options [37, 38]. Both methods rely on the accurate stereotactic implantation of either a multicontact electrode for RFTC or a laser fiber for LITT into the lesion. One advantage of RFTC is that it can be performed using the same electrodes that were previously implanted for SEEG recordings allowing a seamless transition from diagnostics to a therapeutic step [39]. Consequently, its accuracy and applicability is the same as during conventional robotic SEEG. This strategy has proven particularly useful in periventricular nodular heteropias and hypothalamic harmatomas but had lower rates of seizure freedom than resective surgeries [40, 41]. One limitation of RFTC is the lack of direct visualization of the ablation extent.

In contrast, LITT is performed under MRI guidance where real-time MRI thermography is processed through an algorithm calculating the estimated area of ablation every few seconds [42]. This high degree of control has led to the procedure being widely regarded as minimally invasive alternative to open epilepsy surgery and its usage frequency is currently climbing [43]. Despite the accurate ablation control provided by MRI thermography, the precision of the procedure strongly relies on the correct stereotactic implantation of one or multiple laser fibers. Robotic guidance has been adopted in many centers for LITT-based procedures in epilepsy surgery thus achieving laser fiber implantation accuracies ranging around one to 2 mm [44].

Not only the implantation but also positional adjustments of the laser probe are facilitated through robotic technology. The NeuroBlate system (Monteris Medical, Minnetonka, USA) contains a small robotic probe driver, which allows the surgeon to remotely finetune the position and rotation of the side-firing probe [45].

This high degree of accuracy and control has helped widen the indication for ablative procedures. Besides its use in lesional epileptogenic foci [44, 46], ablation has also gained traction as an alternative to open disconnective surgery with multiple reports describing robotic assisted LITT callosotomy [47‒49]. Recently, groups around the world have started to expand the spectrum of stereotactic, ablative disconnections. One Spanish case report describes a Neuromate assisted LITT-based temporoparietooccipital disconnection [50]. One case series reports 5 patients who underwent completion of previously done disconnective surgeries using the ROSA system [51]. A North American report details a complete functional hemispherotomy performed with the same system [52], while in India a group has published their series of functional hemispherotomies performed through multiple RFTC ablations [53]. These impressive results show the growing role of robotic guidance in ablative surgeries.

When presurgical investigations indicate an epileptogenic zone that is either widespread or can only be poorly located, patients are generally not well suited for resective surgeries. Neuromodulation, once solely regarded as a palliative procedure, has gained widespread approval after multiple studies have demonstrated its success. The three most common techniques are deep brain stimulation (DBS), responsive neurostimulation (RNS), and vagus nerve stimulation. As vagus nerve stimulation is a purely extracranial procedure with no reported robotic adjuncts yet, it falls outside the scope of this review.

DBS has been developed as an alternative to stereotactic lesioning. High frequency stimulation essentially acts as a reversible lesion around the stimulating electrodes [54]. Increasing the voltage of stimulation enlarges the electric field, which potentially increases its effectiveness by enlarging the “reversible lesion,” but it can also induce side effects through off-target stimulation.

Currently, the anterior nucleus of the thalamus is the only DBS target approved for the treatment of focal epilepsy by medical regulation authorities. The SANTE trial, a prospective multicenter effort with a period of double-blinded randomized stimulation (5 volts vs. 0 volts) demonstrated a seizure reduction of more than 50% after 2 year [55]. ANT DBS is most effective in patients with temporal lobe seizure foci, and long-term follow-up data indicate that its efficiency increases with time reducing the median seizure frequency by around 75% from baseline [56].

Two additional thalamic targets currently under investigation for DBS in epilepsy are the centromedian nucleus and the pulvinar, both of which are frequently targeted with robotic assistance [57, 58]. Especially, generalized epilepsies like Lennox Gastaut Syndrome responded to centromedian stimulation [57‒59] while temporal lobe and posterior quadrant epilepsies were ameliorated by stimulating the pulvinar [60‒62]. The most common extrathalamic target is the hippocampus, which can be stimulated in refractory temporal lobe epilepsy [63, 64].

The small size of thalamic targets underscores the need for accurate implantation strategies [65, 66]. As a result, many centers utilize robotic implantation systems, which have been reported to reduce target point errors [67, 68]. Despite only implanting two trajectories robotic guidance was also reported to shorten surgical times in DBS procedures as compared to conventional stereotactic methods [67].

RNS as well has been proven beneficial in medically refractory epilepsy in multicenter prospective trials [69, 70]. It allows a sensing electrode to detect and record ictal seizure activity and a stimulation electrode. Both strip electrodes and depth electrodes can be used with RNS making stereotactic implantation appealing method for the latter. For example, in long trajectories such as hippocampal electrodes implanted in a prone position, the rigidity of robotic systems provides a high degree of accuracy [71].

The shorter operating times when using robotic systems as compared to conventional stereotactic approaches are offset by the higher preparation time, which makes both methods equally long when only one RNS depth electrode is implanted [72]. In cases with multiple electrodes, however, using a robotic system can save time [73].

Currently, the majority of robotic systems are used for the stereotactic placement of electrodes or laser fibers. Yet, innovators are already expanding the potential applications of robotic adjuncts. One group from India reported using the ROSA system to secure an endoscope during transcallosal hemispherotomy, effectively employing it as both a holding arm and a neuronavigational device [74].

Despite the robustness of robotic systems, certain error sources remain. Drilling through a held tube can become imprecise due to tugging forces exerted by soft tissues or because of instrument slippage along the curvature of the skull. Especially in long trajectories, these minor deviations to the angle of the drilled hole can lead to increased inaccuracies. To address this problem, robotically guided laser-based craniotomy systems have recently been developed but are still awaiting use in clinical practice. These systems can perform automatic small craniotomies that stop as soon as the skull is perforated [75, 76].

The advent of robotic technology also has the potential to transform conventional, open microsurgery. Multiple companies are currently developing and producing robotically guided 3D exoscopes, which promise improved ergonomics during neurosurgical procedures [77‒79]. These tools could lower the physical strain on the surgeon in procedures that demand oblique view angles with a microscope, such as vertical hemispherotomies.

Robotic technologies in epilepsy surgery have gained an increasing role, evolving from an experimental adjunct to a widely accepted tool that increases accuracy improves treatment efficiency and lowers the procedural risk and time demands. It has helped spread the utilization of SEEG across the globe even to centers, which traditionally relied on subdural grids. The role of robotics has now gone beyond diagnostic applications into the therapeutic procedures, especially in ablative treatments and neuromodulation. These advancements have helped further diversify the treatments options we can offer to patients with a disease as diverse as epilepsy. Currently, new developments are emerging with the promise to further refine the role of robotic systems in epilepsy surgery including automated surgical steps such as robotically performed craniotomies. Robotic systems will take a central role in the future progress of epilepsy surgery.

The authors have no conflicts of interest to declare.

This review was not supported by any sponsor or funder.

C.D. conceptualized this work. M.T. and S.N.S. screened the literature for matching publications. M.T. drafted the initial version of the manuscript. C.D. critically revised the manuscript.