A prediction model integrating synchronization biomarkers and clinical features to identify responders to vagus nerve stimulation among pediatric patients with drug-resistant epilepsy
Ma J, Wang Z, Cheng T, Hu Y, Qin X, Wang W, Yu G, Liu Q, Ji T, Xie H, Zha D, Wang S, Yang Z, Liu X, Cai L, Jiang Y, Hao H, Wang J, Li L, Wu Y.
CNS Neurosci Ther. 2022 Jul 27.
https://doi.org/10.1111/cns.13923
#Prediction of responders for vagus nerve stimulation (#VNS) before implantation is essential but challenging.
This work by Ma et al. (Jul. 2022) established the first prediction model based on #preoperative clinical characteristics and synchronization features of electroencephalography (#EEG) to identify VNS responders among pediatric patients with drug-resistant epilepsy (#DRE).
The model used a support vector machine classifier technology, an algorithm widely used in the epileptic area to predict treatment outcomes (Ref.1-3).
The clinical and synchronization features of scalp EEG data from 88 pediatric patients (VNS follow-up time: >1 year) were collected, and 70 of them were used to build the model. The accuracy was tested in the remaining 18 patients, showing a 61.1% accuracy.
The article provided clinical evidence that a prediction model based on a machine learning algorithm can facilitate patient selection by enabling the prediction of treatment outcomes for VNS. In future studies, better performance and precision of the prediction model is likely to be achieved with a larger sample size (from multiple centers) and by the use of other biomarkers from MRI, DTI, and MEG data.
Ref. 1. doi:10.1016/j.nicl.2017.09.015 2. doi:10.1002/ana.25574 3. doi:10.1109/5254.708428