重型颅脑创伤急诊去骨瓣减压手术临床预后预测模型的开发及验证评价  

Development and Validation Evaluation of Clinical Prognosis Prediction Model for Emergency Decompressive Craniectomy in Severe Traumatic Brain Injury

作　　者：唐知己[1] 李卫[1] 黄乾亮[1] 钟鸣[2] 何晓明 王乐临 曾瑞亮 TANG Zhi-ji;LI Wei;HUANG Qian-liang;ZHONG Ming;HE Xiao-ming;WANG Le-lin;ZENG Rui-liang(Department of Neurosurgery,Ganzhou People's Hospital,Ganzhou 341000,Jiangxi,China;Department of Neurosurgery,Huichang County People's Hospital,Huichang 342600,Jiangxi,China;Department of Neurosurgery,Shangyou County People's Hospital,Shangyou 341200,Jiangxi,China;Department of Neurosurgery,Ningdu County People's Hospital,Ningdu 342800,Jiangxi,China;Department of Neurosurgery,Ruijin People's Hospital,Ruijin 342500,Jiangxi,China)

机构地区：[1]赣州市人民医院神经外科,江西赣州341000 [2]会昌县人民医院神经外科,江西会昌342600 [3]上犹县人民医院神经外科,江西上犹341200 [4]宁都县人民医院神经外科,江西宁都342800 [5]瑞金市人民医院神经外科,江西瑞金342500

出　　处：《医学信息》2024年第18期25-30,共6页Journal of Medical Information

基　　金：江西省科技厅重点研发计划——一般项目(编号:20203BBGL73174)。

摘　　要：目的构建重型颅脑损伤患者急诊行去骨瓣减压的预后预测模型并进行评价和验证。方法对2015年1月-2021年1月在赣州市人民医院等5家医院的重型颅脑损伤患者临床资料进行回顾性分析。观察患者术后6个月的结局,以GOS评分1~2分为不良预后,通过对建模数据集进行单因素和多因素Logistic回归分析,明确去骨瓣减压术后6个月不良预后的独立预测因素。利用R语言编程构建术后6个月内不良预后的早期预测模型,并对模型进行评价和验证。结果最终有386例患者纳入研究,随机分为建模组286例,验证组100例。建模数据集6个月内总体不良预后率为55.59%。年龄、GCS、失血量>750 ml、环池消失、术中低血压,以及APTT>36 s是重型颅脑损伤患者行急诊去骨瓣减压术后6个月不良预后的独立影响因素(P<0.05)。利用预测因素构建了预测模型,经评价和验证,提示模型均具备良好的区分能力,校准度。结论基于颅脑损伤去骨瓣减压术后6个月总体不良预后的预测因素构建的预测模型,以预测列线图的方式展示,验证结果优良,可应用于临床。预测模型有助于指导医师临床决策,也可作为个体化治疗方案的参考。Objective To construct a predictive model for the prognosis of emergency decompressive craniectomy in patients with severe craniocerebral injury and to evaluate and verify it.Methods The clinical data of patients with severe craniocerebral injury in 5 hospitals including Ganzhou People's Hospital from January 2015 to January 2021 were retrospectively analyzed.The outcome of patients at 6 months after operation was observed,and the GOS score 1-2 was divided into poor prognosis.Univariate and multivariate Logistic regression analysis was performed on the modeling data set to identify the independent predictors of poor prognosis at 6 months after decompressive craniectomy.The early prediction model of poor prognosis within 6 months after operation was constructed by R language programming,and the model was evaluated and verified.Results A total of 386 patients were included in the study and randomly divided into the modeling group(286 patients)and the validation group(100 patients).The overall poor prognosis rate of the modeling data set within 6 months was 55.59%.Age,GCS,blood loss>750 ml,disappearance of ambient cistern,intraoperative hypotension,and APTT>36 s were independent influencing factors for poor prognosis at 6 months after emergency decompressive craniectomy in patients with severe craniocerebral injury(P<0.05).The prediction model was constructed by using the predictive factors.After evaluation and verification,it was suggested that the model had good discrimination ability and calibration.Conclusion The predictive model based on the predictive factors of overall poor prognosis at 6 months after decompressive craniectomy for craniocerebral injury is displayed in the form of predictive nomogram,and the verification results are excellent and can be applied to clinical practice.The prediction model is helpful to guide the doctor's clinical decision-making and can also be used as a reference for individualized treatment plan.

关 键 词：颅脑创伤 去骨瓣减压 危险因素 预测模型 

分 类 号：R651.15[医药卫生—外科学]