Image Guided Surgery

Epilepsy Navigator (EpiNavTM) is an interactive software platform we are developing in conjunction with our clinical partners at the National Hospital for Neurology and Neurosurgery to assist in planning and guiding surgical interventions for epileptic patients [1].


Epilepsy affects about 500,000 people in the UK. 20-30% of epilepsy patients continue to experience seizures after medication. Such patients are candidates for curative resection, where the brain region responsible for seizure onset is surgically removed in an attempt to eliminate seizures. Planning for curative resection involves identifying the brain region to be removed as well as ensuring that the resection avoids eloquent regions involved in motor, language, and vision.

The goal of EpiNavTM is to develop a software platform to assist in planning and guiding surgical interventions for patients at every stage of the clinical process. EpiNavTM allows for the creation of multi-modal 3D maps of brain structures which can be used to direct neurosurgery. We are actively developing automated segmentation and planning tools in this platform to assist in electrode implantation and resection. The current EpiNavTM team at TIG consists of: Sebastien Ourselin, Rachel Sparks, Sjoerd Vos, Martin Schweiger, Alejandro Granados Martinez, Matteo Mancini, Phil Noonan.


Automated Vessel Segmentation

Flowchart of vessel extraction algorithm

Figure 1. Flowchart of the vessel extraction algorithm used for SEEG Electrode Planning.

The ability to accurately identify and delineate structures to be avoided is essential to provide guidance for interventional neurosurgery. The most critical structures to be avoided are blood vessels, which may cause hemorrhage if punctured during surgery.


We have developed a novel multi-modal, multi-scale vessel segmentation algorithm for intracranial vessels [2]. Our method extracts a set of tensors for each input image over a range of scales. Tensors over images and scales are fused with a tensor voting scheme. A final segmentation is obtained using marching cubes on the fused image. Figure 1 illustrates a flowchart of our algorithm.




Assisted SEEG Electrode Placement

Figure 2. SEEG electrode trajectory determined via computer assisted planning algorithm. Red and light blue surfaces represent arteries and veins, respectively while the white surface represents the scalp. The colored patch represents the risk for potential entry points.

Stereoelectroencephalography (SEEG) depth electrodes are placed within the brain to help identify the brain region responsible for seizure onset. This invasive investigation carries the risk of hemorrhage, infection, and neurologic deficit. Careful planning of electrode placement can reduce risks by avoiding critical structures (e.g. blood vessels, cerebrospinal tract) within the brain.

We have developed assisted planning tools to aid in determining SEEG electrode placement [3]. Electrode trajectories, the path between entry on the scalp surface and the target in the cortex, are assessed via an automated entry point search and risk evaluation.

Entry point search first finds all points on the surface of the scalp. Potential entry points are removed from consideration if they meet one of three hard criteria: trajectories intersect critical structures, have too sharp an angle with respect to the skull surface, are longer than the SEEG electrode.

Risk evaluation is performed on the remaining trajectories by calculating a risk score that integrates the distance to critical structures along the entire trajectory. The lowest risk score trajectory is selected (pink line in Figure 2) and a risk map displays the relative risk score for the other potential entry points (colored patch in Figure 2 where red represents the highest risk and green represents the lowest risk).

EpiNavTM allows the user to interact with the selected trajectory, enabling manual modification of the trajectory to other low risk trajectories if the automatically defined trajectory is still unsuitable. Additionally EpiNavTM allows the user to navigate through the imagery along the path of the trajectory to identify the anatomic location of the trajectory throughout the cortex.

 Resection Planning

To minimise damage to the brain during  surgery  while still removing the seizure area, we are integrating neuroimaging techniques into the process of planning how to do the resection. One such technique is diffusion tractography which can reveal the brain’s wiring and connections between different functional areas.

The most common surgery for epilepsy is in the temporal lobe, where resection of its anterior part can be an effective long-term treatment option [4]. In this surgery, a key brain connection in the visual system, the optic radiation, is at risk [5]. Figure 3 shows an example  the optic radiation fibre bundle. We have shown that  tractography can help prevent damage to the nerve fibres of the visual system, effectively preserving the vision of patients undergoing surgery [6].

Website EpiNav OR FT Figure 3: Tractography of the optic radiation, seen from above (left image, axial view) and the side (right image, sagittal view). This fibre bundle starts from a central area, goes anteriorly and inferior, and then loops back to the visual cortex at the back of the brain. This most anterior loop, called Meyer’s loop, is the area most at risk of damage during resection.

 Ongoing work is focused on considering multiple brain connections to avoid in order to to reduce side effects, and extend tractography-guided resections to surgery of different brain areas. The complication is that this rapidly becomes an increasing complex surgical procedure, however, sophisticated software tools can aid in identify safe resection margins while still providing an effective cure to the patient’s epilepsy.



[1] J.S. Duncan et al., “Brain imaging in the assessment for epilepsy surgery”, Lancet Neurology 2016;15, 420-33.

[2] M.A. Zuluaga, et al. “SEEG Trajectory Planning: Combining Stability, Structure and Scale in Vessel Extraction”. In: Medical Image Computing and Computer-Assisted Interventions - MICCAI 2014

[3] Zombori, G., et al. "A computer assisted planning system for the placement of sEEG electrodes in the treatment of epilepsy." In: (Proceedings) Information Processing in Computer-Assisted Interventions. Fukuoka, Japan, 2014.

[4] J. de Tisi et al., “The long-term outcome of adult epilepsy surgery, patterns of seizure remission, and relapse: a cohort study”, Lancet 2011; 378, 1388-95.

[5] G.P. Winston et al., “Optic radiation tractography and vision in anterior temporal lobe resection”, Annals of Neurology 2012; 71, 33-41.

[6] G.P. Winston et al., “Preventing visual field deficits from neurosurgery”, Neurology 2014; 83, 604-11.