Probabilistic Mapping Reveals Optimal Stimulation Site in Essential Tremor.
Nowacki A., Barlatey S., Al-Fatly B., Dembek T., Bot M., Green AL., Kübler D., Lachenmayer ML., Debove I., Segura-Amil A., Horn A., Visser-Vandewalle V., Schuurman R., Barbe M., Aziz TZ., Kühn AA., Nguyen TAK., Pollo C.
OBJECTIVE: The objective of this study was to obtain individual clinical and neuroimaging data of patients undergoing deep brain stimulation (DBS) for essential tremor (ET) from 5 different European centers to identify predictors of outcome and to identify an optimal stimulation site. METHODS: We analyzed retrospectively baseline covariates, pre- and postoperative clinical tremor scores (for 12 months) as well as individual imaging data from 119 patients to obtain individual electrode positions and stimulation volumes. Individual imaging and clinical data were used to calculate a probabilistic stimulation map in normalized space using voxel-wise statistical analysis. Finally, we used this map to train a classifier to predict tremor improvement. RESULTS: Probabilistic mapping of stimulation effects yielded a statistically significant cluster that was associated with a tremor improvement >50%. This cluster of optimal stimulation extended from the posterior subthalamic area to the ventralis intermedius nucleus and coincided with a normative structural connectivity-based cerebellothalamic tract (CTT). The combined features "distance between the stimulation volume and the significant cluster" and "CTT activation" were used as a predictor of tremor improvement. This correctly classified a >50% tremor improvement with a sensitivity of 89% and a specificity of 57%. INTERPRETATION: Our multicenter ET probabilistic stimulation map identified an area of optimal stimulation along the course of the CTT. The results of this study are mainly descriptive until confirmed in independent datasets, ideally through prospective testing. This target will be made openly available and may be used to guide surgical planning and for computer-assisted programming of DBS in the future. ANN NEUROL 2022;91:602-612.