Abstract
Utility of 3D Multimodality Imaging in the Implantation of Intracranial Electrodes in Epilepsy
Nowell M, Rodionov R, Zombori G, Sparks R, Winston G, Kinghorn J, Diehl B, Wehner T, Miserocchi A, McEvoy AW, Ourselin S, Duncan J. Epilepsia 2015;56:403–413.
OBJECTIVE: We present a single-center prospective study, validating the use of 3D multimodality imaging (3 DMMI) in patients undergoing intracranial electroencephalography (IC-EEG). METHODS: IC-EEG implantation preparation entails first designing of the overall strategy of implantation (strategy) and second the precise details of implantation (planning). For each case, the multidisciplinary team made decisions on strategy and planning before the disclosure of multimodal brain imaging models. Any changes to decisions, following disclosure of the multimodal models, were recorded. RESULTS: Disclosure of 3 DMMI led to a change in strategy in 15 (34%) of 44 individuals. The changes included addition and subtraction of electrodes, addition of grids, and going directly to resection. For the detailed surgical planning, 3 DMMI led to a change in 35 (81%) of 43 individuals. Twenty-five (100%) of 25 patients undergoing stereo-EEG (SEEG) underwent a change in electrode placement, with 158 (75%) of 212 electrode trajectories being altered. SIGNIFICANCE: The use of 3 DMMI makes substantial changes in clinical decision making.
Commentary
Surgical evaluation of medically intractable epilepsy often entails implantation of intracranial electrodes for the purpose of localizing the epileptogenic zone and mapping the cortex (1). Invasive monitoring should be based on a solid presurgical hypothesis that is generated after careful analyses of the seizure semiology, ictal and interictal EEG, seizure-protocol brain MRI, as well as other tests, including neuropsychological profile, intracarotid amobarbital test, functional neuroimaging (such as functional MRI, FDG–PET, and ictal SPECT), and magnetoecephalography. Various studies have shown how different presurgical test results correlate with the surgical outcome. For example, not only does regional hypometabolism on FDG-PET correlate with good seizure outcome after resection even with normal, nonlesional brain MRI (2), but the superimposition of brain MRI with PET increases the chances of detecting subtle abnormalities such as cortical dysplasias (3, 4). Similarly, coregistration of the subtraction ictal–interictal SPECT studies with the brain MRI (SISCOM) has demonstrated great utility (5). Thus, there is a definite role for superimposition of different imaging studies in identification of the seizure focus and in defining its anatomical relation to surrounding functional brain regions. Postimplantation of intracranial electrodes, superimposition of postoperative CT on presurgical studies is also important to identify the location of intracranial electrodes in relation to gyral anatomy; this would help the clinician put the ictal and mapping data in the proper surgical context as the resection margins are being determined. In addition, displaying the integrated studies on neuronavigation systems can facilitate planning the implantation of intracranial electrodes and deciding the coordinates of their trajectories, avoiding blood vessels and increasing the safety of implantation.
Nowell et al. studied the role of 3D multimodal imaging (3DMMI) on surgical strategy and planning in individuals with intractable epilepsy. Their aim was to demonstrate how the use of 3DMMI could change implantation strategy or surgical planning. The authors defined implantation strategy as the decision whether to abandon surgery, go directly to surgical resection, or plan invasive monitoring. In the case of the latter, further specification of subdural versus depth electrodes is also decided. Surgical planning entails the actual details of implantation, including electrode numbers, types, and trajectories. The 3DMMI concomitantly displays data from the T1-weighted Stealth MRI scan with gadolinium and 3D phase-contrast MRI scan, in addition to those from other studies, such as PET, functional MRI, tractography, and CT angiogram when SEEG implantation was intended. The outcome image was a 3D-volume-rendered brain representation, with volume- or surface-rendered clusters representing different imaging modalities. For each case, the multidisciplinary team initially decided on surgical strategy and planning before the disclosure of 3DMMI and then recorded any changes dictated by disclosing the 3DMMI. The authors found that 3DMMI resulted in changing the surgical strategy in 15 (34%) of 44 individuals. This included changes in the numbers of electrodes to either improve coverage of brain areas as indicated by functional and neurophysiologic data or minimize the surgical risk, addition of grids, and skipping intracranial monitoring altogether and going directly to resection. Regarding detailed surgical planning, 3DMMI resulted in changing the plans in 35 (81%) of 43 individuals. All 25 patients undergoing SEEG underwent a change in electrode placement, including altering 75% of the electrode trajectories. The authors concluded that 3DMMI was influential in determining surgical strategy through displaying structural and functional data, as well as altering precise planning through displaying vascular and gyral anatomy. The surgical outcome data are insufficient but did not seem superior to the ones in the literature. Additionally, the surgical complication rate appeared to be comparable to that in the literature, with one case of hemorrhage during SEEG implantation, two cases of subdural hematomas necessitating evacuation, and two cases of infection.
The study adds to other studies that have suggested the utility of superimposition of multimodal data in the surgical planning of intractable epilepsy. An important study showed that SISCOM images were remarkably superior to visual inspection of ictal and interictal SPECT images because they localized the seizure onset zone in 45 of 51 (88.2%) patients compared with 20 of 51 (39.2%) for traditional SPECT inspection (5). Similarly, a recent study found that coregistration of PET with MRI identified hypometabolic areas that were not detected on PET alone (3). There are known difficulties in designing a controlled study to demonstrate the utility of multimodal imaging, but it would be interesting if the technological advances in brain imaging—including accurate image coregistration and neuronavigation—can improve surgical outcomes or decrease the rate of surgical complications.