摘要
目的探究甲状腺根治术后喉返神经损伤的风险因素,并建立相应的风险预测模型以指导临床实践。方法选取甲状腺根治术的500例患者作为研究对象,根据术后随访统计喉返神经损伤情况,通过分析患者年龄、性别、甲状腺病理类型、术前喉镜检查结果、手术次数、VI区淋巴结清扫、术中出血量等临床资料,利用多因素logistic回归分析确定喉返神经损伤的危险因素。采用R软件构建风险预测模型,并通过受试者工作特征(ROC)曲线评估模型的区分度。使用Bootstrap方法进行1000次重复采样以验证模型的预测效能。结果喉返神经损伤患者为10例,单因素结果显示患者发生喉返神经损伤与手术次数、不同的手术入路、术中出血量及VI区淋巴结清扫等有关;多因素logistic回归分析显示,多次手术、甲状腺下动脉入路、术中出血量>100 mL及进行VI区淋巴结清扫是喉返神经损伤的危险因素(P<0.05);模型预测喉返神经损伤风险的AUC面积为0.905,模型的偏差校准曲线与理想曲线吻合良好,MAE为0.036。结论通过分析患者手术次数、不同的手术入路、术中出血量及VI区淋巴结清扫等因素来构建甲状腺根治术后喉返神经损伤风险的预测模型是可行的,该模型对于指导临床实践、降低喉返神经损伤风险具有重要价值。
Objective This study aims to explore the risk factors of recurrent laryngeal nerve injury after radical thyroidectomy and establish corresponding risk prediction models to guide clinical practice.Methods This study selected 500 patients who underwent radical thyroidectomy.Based on postoperative follow-up statistics,the risk factors for recurrent laryngeal nerve injury were determined using multivariate logistic regression analysis by analyzing clinical data such as patient age,gender,thyroid pathological type,preoperative laryngoscopy results,number of surgeries,lymph node dissection in zone VI,and intraoperative blood loss.Using R software to construct a risk prediction model,and evaluating the discriminability of the model through receiver operating characteristic(ROC)curves.Use the Bootstrap method for 1000 repeated samples to validate the predictive performance of the model.Results There were 10 patients with recurrent laryngeal nerve injury in this study,the univariate results showed that the occurrence of recurrent laryngeal nerve injury in patients was related to the number of surgeries,different surgical approaches,intraoperative blood loss,and lymph node dissection in zone VI;Multivariate logistic regression analysis showed that multiple surgeries,inferior thyroid artery approach,intraoperative bleeding>100 mL,and lymph node dissection in zone VI were risk factors for recurrent laryngeal nerve injury(P<0.05);The ROC curve analysis of the risk prediction model constructed based on these indicators shows that the AUC area of the model predicting the risk of recurrent laryngeal nerve injury is 0.905;Bootstrap validation shows that the deviation calibration curve of the model is in good agreement with the ideal curve,with a MAE of 0.036.Conclusion It is feasible to construct a predictive model for the risk of recurrent laryngeal nerve injury after radical thyroidectomy by analyzing factors such as patient surgery frequency,different surgical approaches,intraoperative blood loss,and lymph node dissection in zone VI.This model has important value in guiding clinical practice and reducing the risk of recurrent laryngeal nerve injury.
作者
任俊杰
王成
寇乐
REN Junjie;WANG Cheng;KOU Le(Surgical Oncology,the First People′s Hospital of Yinchuan,Yinchuan 750001,China;Department of Dermatology,Ningxia Combined Traditional Chinese and Western Medicine Hospital,Yinchuan 750001,China)
出处
《宁夏医学杂志》
CAS
2024年第10期867-870,I0001,共5页
Ningxia Medical Journal
关键词
甲状腺根治术
喉返神经损伤
危险因素
预测模型
甲状腺病理
Radical thyroidectomy
Injury of recurrent laryngeal nerve
Risk factors
Prediction model
Thyroid pathology
