探讨人工神经网络（ANN）模型预测三叉神经痛（TN）患者显微血管减压术（MVD）术后长期临床疗效的价值。

收集自2013年3月至2018年5月于浙江大学医学院附属杭州市第一人民医院神经外科接受MVD手术的1041例TN患者的围术期多数据资料，构建ANN预测模型。比较预测结果与实际随访结果，改变输入层变量，寻找对ANN预测精度影响最大的因素。

ANN模型可以准确预测TN患者MVD术后的长期疗效，准确率为94.2%。血管压迫三叉神经的解剖位置与面部疼痛部位是否符合、MVD术后即刻疼痛缓解情况、责任血管压迫程度、责任血管类型4个因素在ANN预测性能中贡献最大，将其依次删除后模型预测的准确率下降为75.3%、79.8%、86.6%、89.2%。

ANN模型可客观准确地预测TN患者MVD术后的长期临床疗效，并且该模型可以评估临床疗效预测中每个因素的重要性。

To explore the value of artificial neural network (ANN) model in predicting the long-term clinical efficacy of microvascular decompression (MVD) in patients with trigeminal neuralgia (TN).

The perioperative data of 1041 TN patients who underwent MVD surgery in Neurosurgery Department of Affiliated Hangzhou First People’s Hospital, Zhejiang University School of Medicine from March 2013 to May 2018 were collected to construct the ANN prediction model. The prediction results are compared with the actual follow-up results to evaluate the performance of the model, then changing the variables of the input layer to find the factors which have the greatest impact on the prediction accuracy of the model.

The ANN model can accurately predict the long-term efficacy of TN patients after MVD with an accuracy rate of 94.2%. These four factors (whether the anatomical location of offending vessels on the trigeminal nerve and the location of facial pain meet, the pain relief in the short term after MVD, the degree of the offending vessels compression, and the type of offending vessels) contribute the most to the prediction performance of ANN. After deleting them in turn, the prediction accuracy of the model dropped to 75.3%, 79.8%, 86.6%, 89.2%.

The ANN model can objectively and accurately predict the long-term clinical efficacy of TN patients after MVD, and the model can evaluate the importance of each factor in the prediction of clinical efficacy.