Multidata Analysis Based on an Artificial Neural Network Model for Long-Term Pain Outcome and Key Predictors of Microvascular Decompression in Trigeminal Neuralgia

Abstract

Objective: To investigate use of multidata analysis based on an artificial neural network (ANN) to predict long-term pain outcomes after microvascular decompression (MVD) in patients with trigeminal neuralgia (TN) and to explore key predictors.

Methods: Perioperative and long-term follow-up multidata of 1041 patients with TN who received MVD surgery at Hangzhou First People's Hospital from March 2013 to May 2018 were collected to construct an ANN model for prediction. The prediction results were compared with the actual follow-up outcomes, and the variables in each input layer were changed to test the effectiveness of ANN and explore the factors that had the greatest impact on prediction accuracy.

Results: The ANN model could predict the long-term pain outcomes after MVD in patients with TN with an accuracy rate of 95.2% and area under the curve of 0.862. Four factors contributed the most to the predictive performance of the ANN: whether the neurovascular offending site of the trigeminal nerve corresponded the region of facial pain, immediate postoperative pain remission after MVD, degree of nerve compression by culprit vessels, and the type of culprit vessels. After these factors were sequentially removed, the accuracy of the ANN model decreased to 74.5%, 78.6%, 87.2%, and 90.1%, while the area under the curve was 0.705, 0.761, 0.793, and 0.810.

Conclusions: The ANN model, constructed using multiple data, predicted long-term pain prognosis after MVD in patients with TN objectively and accurately. The model was able to assess the importance of each factor in the prediction of pain outcome.

Keywords: ANN; Artificial neural networks; MVD; Microvascular decompression; TN; Trigeminal neuralgia.