Multimodal image-guided transcranial magnetic stimulation in the delineation of eloquent cerebral cortex in the neurosurgical patient and the treatment of refractory focal epilepsy
The goal of this project was to clinically validate motor and language functional mapping with transcranial magnetic stimulation (TMS) and to treat refractory focal epilepsy with multimodal image-guided repetitive TMS (rTMS).
Each year, more than 65,000 people are diagnosed with cancer in Belgium. Of those, 10-25% will experience brain metastasis at some point in the disease course. Part of those brain lesions need surgery, especially if only one brain metastasis is present, as is the case in half of the patients.18 Primary brain tumors are rarer with an age-adjusted incidence rate of 21.42 per 100,000.19
In addition, some forms of epilepsy are amendable to surgery and up to 1,600 new patients each year in Belgium could in theory be evaluated for this option. With surgery, the aim is not only to render the patient seizure-free by removal of the epileptogenic zone, but equally important, to preserve normal functioning, including movement and speech. Anatomical knowledge and fMRI are used to gauge risk of inducing deficits with surgery. However, due to the tumor changing both the anatomy and the response of the brain that is measured with fMRI, those methods are far from perfect. Therefore, if surgery is deemed possible, direct mapping of the brain with DCS can be performed. Especially for language, DCS is done during awake surgery. This is not possible in all patients, it can complicate surgery and has false positive and false negative findings. Moreover, the information is only available during the surgery. TMS-based mapping is an attractive, non-invasive alternative to map brain function. This has been developed further for motor and language function.
In epilepsy, if surgery is not an option and seizures continue despite adequate anti-epileptic drug treatments, alternatives are needed. Since rTMS can change the excitation-inhibition balance in the brain by pattern application of repeated TMS-pulses, it could prove to be a valuable treatment option. Conflicting evidence was available on this prior to our experiments. To advance the knowledge on rTMS for the treatment of epilepsy even further, FDG-PET scans were performed at baseline and after each treatment condition.
rTMS for the treatment of epilepsy
We have conducted a randomized sham-controlled crossover trial that included 11 patients with well-defined focal epilepsy. rTMS (0.5 Hz) was targeted to the focus during three treatment conditions consisting of 1500 pulses/day for 10 weekdays at 90% of resting motor threshold (rMT) followed by a 10-week observation period. None of the patients achieved an overall 50% seizure reduction. Side effects in the study were a rebound in seizure frequency after initial reduction in seizure frequency (one patient) and increase in seizure frequency during and shortly following active stimulation (overall change in seizure frequency over 1 month and 12 weeks observation period was not significant). FDG-PET scans showed that rTMS caused measurable changes in brain metabolism in patients with epilepsy. No single pattern of change emerged however. Only three out of eight subjects showed a relative decrease in brain metabolism around the stimulation target after stimulation with the figure-8 coil, and five out of nine after the round coil. Increased metabolism around the stimulation target was seen in one patient after using the figure-8 coil and in four after round coil sstimulation. FDG-PET metabolic changes in the area of stimulation were only apparent after active stimulation, not sham treatment, and a carry-over effect after sham treatment following active treatment was observed, pointing to a lasting change in metabolism after rTMS of at least 12 weeks. Widespread changes at a distance of the stimulation target were also observed in all patients. These observations indicate that rTMS induces plastic changes in the brain, but a clear and predictable pattern was not present. This could in part explain the lack of efficacy of rTMS as a treatment of refractory focal epilepsy in the study.
TMS-based motor and language mapping
In our study on TMS-based motor cortex mapping, we created and analyzed different models.
To delineate the whole area of the motor cortex representation, the model based on the weighted average of the induced electric fields calculated with a realistic head model performed best.
The optimal single threshold to visualize the field-based maps was 40% of the maximal value for the anisotropic model and 50% for the isotropic model. For clinical purposes, we suggest to use different thresholds, which is a unique benefit of these maps: a high threshold highlights the center of the motor area and a lower threshold is able to capture the motor representation completely.
The interpretation of those maps is that regions with high values are those where high induced electric field strengths are likely to result in high MEP amplitudes.
For language mapping, we have developed and tested new ways to set up the mapping and analyze the data. Our automated speech analysis algorithm was able to discern both speech onset and speech content, even in the presence of TMS noise. Our algorithm was easy-to-use, fast, reliable and had high accuracy (90%) and specificity (96%) on the data derived in healthy volunteers and with a lower accuracy (61%) in patients. A TMS-RT-based probability map of the language cortex in the three patients was consistent with anatomical knowledge, DCS data and was able to predict postsurgical transient language decline in one patient with negative DCS-mapping. TMS-induced increases in RT of correct responses during a confrontational naming task may be an important biomarker of language cortex. Further studies might want to focus on TMS-RT-based mapping of the language cortex with lower stimulation intensities, which would make the procedure less painful. Patients’ comfort should be maximized for the technique to become a first-line test in the preoperative delineation of language cortex.
Contributions have been made in TMS-based motor mapping and current performance seems reliable enough that implementing it into clinical care seems reasonable. Our contributions in TMS-based language mapping should make future trials on the topic easier due to the development of an automated speech analysis algorithm and a new direction for future research has been explored, namely the use of TMS-induced increases in RT of correct responses during a confrontational naming task as a relevant marker to map the language cortex.
In the development of rTMS as treatment for epilepsy, we have published a negative trial. Future directions of research based on these findings were discussed, especially since the clinical data were corrugated by changes in brain metabolism as measured with FDG-PET.