Prediction of Radiofrequency-induced Heating of Spinal Fixation System Using Mesh-based Convolutional Neural Network

In this paper, a convolutional neural network (CNN) model is proposed to predict the RF-induced heating of Spinal Fixation System (SFS). Device mesh information is combined with the incident electric field information as the input for the network model. Principal Component Analysis (PCA) is performed to estimate the appropriate number of training data. 1500 SFS models are created and 296 of thoseare used as training data. Results show good convergence for network training. The R2 score of rod and screw RF-induced heating prediction all reaches above 0.95.