脑动静脉畸形病人栓塞术后发生神经功能障碍的风险预测模型构建与验证

Construction and validation of a risk prediction model for neurological dysfunction in patients with cerebral arteriovenous malformations after embolization surgery

目的脑动静脉畸形病人栓塞术后发生神经功能障碍的风险预测模型构建与验证。方法选取2020年1月至2021年12月梅州市人民医院收治的72例脑动静脉畸形病人为研究对象,均行介入栓塞术。根据是否发生神经功能障碍分为发生神经功能障碍组(19例)及未发生神经功能障碍组(53例)。利用logistic多因素回归分析脑动静脉畸形病人栓塞术后发生神经功能障碍的危险因素。使用R3.5.3软件绘制预测脑动静脉畸形病人栓塞术后发生神经功能障碍的列线图模型。应用受试者操作特征曲线(ROC曲线)下面积对列线图模型预测效能进行检验,利用Bootstrap法检验模型的准确性,并采用决策曲线分析(DCA)评价模型的临床实用性。结果单因素分析显示发生神经功能障碍组术前颅内出血、病灶长径、病灶位置、动脉瘤、Spetzler-Martin分级、深部引流静脉与未发生神经功能障碍组比较差异有统计学意义(P<0.05)。多因素logistic回归分析结果显示,颅内出血、病灶长径、病灶位置、动脉瘤、Spetzler-Martin分级、深部引流静脉均是脑动静脉畸形病人栓塞术后发生神经功能障碍的危险因素(P<0.05)。将logistic多因素分析的结果建立预测脑动静脉畸形病人栓塞术后发生神经功能障碍的风险预警模型,Bootstrap法内部验证结果显示,C-index指数为0.82[95%CI:(0.76,0.88)]。ROC曲线下面积、灵敏度、特异度分别为0.79、81.42%、82.69%。决策曲线可选阈概率为13%~87%,净获益值较高。结论术前颅内出血、病灶长径、病灶位置、动脉瘤、Spetzler-Martin分级、深部引流静脉均是脑动静脉畸形病人栓塞术后发生神经功能障碍的危险因素,基于以上因素构建的列线图风险模型对脑动静脉畸形病人栓塞术后发生神经功能障碍具有较好的预测效能。

Objective To establish and verify a risk prediction model for neurological dysfunction in patients with cerebral arteriovenous malformations after embolization.Methods A total of 72 patients with cerebral arteriovenous malformation treated in Meizhou People's Hospital from January 2020 to December 2021 were selected as the study objects,all of whom underwent interventional embolization.According to whether the neurological dysfunction occurred,the patients were divided into the neurological dysfunction group(19 cases)and the non-neurological dysfunction group(53 cases).Multivariate logistic regression was used to analyze the risk factors of neurological dysfunction in patients with cerebral arteriovenous malformations after embolization.R3.5.3 software was used to create a nomogram model for predicting neurological dysfunction in patients with cerebral arteriovenous malformations after embolization.The area under receiver operating characteristic curve(ROC curve)was used to test the prediction efficiency of the nomogram model,and the accuracy of the model was tested by Bootstrap method,and a decision curve analysis(DCA)was drawn to evaluate the clinical practicality of the model.Results Univariate analysis showed that there were significant differences in preoperative intracranial hemorrhage,lesion size,lesion location,aneurysm,Spetzler-Martin grade and deep drainage vein between the group with neurological dysfunction and the group without neurological dysfunction(P<0.05).Multivariate logistic regression analysis showed that intracranial hemorrhage,lesion size,lesion location,aneurysm,Spetzler-Martin grade,and deep drainage vein were all risk factors for neurological dysfunction in patients with cerebral AVM after embolization(P<0.05).The results of logistic multivariate analysis were used to establish a risk warning model for predicting neurological dysfunction in patients with cerebral arteriovenous malformation after embolization.The internal verification results of Bootstrap method showed that the C-index index was 0.82[95%CI:(0.76,0.88)].The area,sensitivity and specificity under ROC curve were 0.79,81.42%and 82.69%,respectively.The optional threshold probability of the DCA curve is 13%-87%,and the net benefit value is relatively high.Conclusions Preoperative intracranial hemorrhage,lesion size,lesion location,aneurysm,Spetzler-Martin grade and deep drainage vein are all risk factors for neurological dysfunction after embolization in patients with cerebral arteriovenous malformations.Based on the above factors,the nomogram risk model has a good predictive effect on nerve dysfunction in patients with cerebral arteriovenous malformations after embolization.