Probabilistic Subthalamic Nucleus Stimulation Sweet Spot Integration Into a Commercial Deep Brain Stimulation Programming Software Can Predict Effective Stimulation Parameters

Objectives: Subthalamic nucleus (STN) deep brain stimulation (DBS) programming in patients with Parkinson disease (PD) may be challenging, especially when using segmented leads. In this study, we integrated a previously validated probabilistic STN sweet spot into a commercially available software to evaluate its predictive value for clinically effective DBS programming.Materials and Methods: A total of 14 patients with PD undergoing bilateral STN DBS with segmented leads were included. A nonlinear co-registration of a previously defined probabilistic sweet spot onto the manually segmented STN was performed together with lead reconstruction and tractography of the corticospinal tract (CST) in each patient. Contacts were ranked (level and direction), and corresponding effect and side-effect thresholds were predicted based on the overlap of the volume of activated tissue (VTA) with the sweet spot and CST. Image-based findings were correlated with postoperative clinical testing results during monopolar contact review and chronic stimulation parameter settings used after 12 months.Results: Image-based contact prediction showed high interrater reliability (Cohen kappa 0.851-0.91). Image-based and clinical ranking of the most efficient ring level and direction of stimulation were matched in 72% (95% CI 57.0-83.3) and 65% (95% CI 44.9-81.2), respectively, across the whole cohort. The mean difference between the predicted and clinically observed effect thresholds was 0.79 +/- 0.69 mA (p = 0.72). The median difference between the predicted and clinically observed side-effect thresholds was -0.5 mA (p < 0.001, Wilcoxon paired signed rank test). Conclusions: Integration of a probabilistic STN functional sweet spot into a surgical programming software shows a promising capability to predict the best level and directional contact(s) as well as stimulation settings in DBS for PD and could be used to optimize programming with segmented lead technology. This integrated image-based programming approach still needs to be evaluated on a bigger data set and in a future prospective multicenter cohort.

Automated summary

This study integrated a previously validated probabilistic STN sweet spot into a commercially available software to evaluate its predictive value for clinically effective DBS programming in patients with Parkinson disease. The results showed a certain level of match between the image-based prediction and clinical evaluation in terms of stimulation effect and side-effect thresholds, indicating the potential of this integrated image-based programming approach in optimizing DBS programming for PD patients.