基于Lasso-Logistic回归构建机器人辅助经皮椎弓根螺钉内固定术患者术中低体温的预测模型

打开文本图片集

中图分类号 R473.6 文献标识码A 文章编号 2096-7721（2026）01-0085-08

AbstractObjective：Toestablishapreditivemodelforntraoperativehypothemiainpatientsundergoingrobot-asistedperutaneous pediclescrewinteralfixation （RPPSIF）usingLasso-Lgisticregresion.Methods：234RPPSIFpatients reatedatBeijingTogren Hospital，CapitalMedicalUniversityfroJanuary22toJune4 wrenrolled.LasoLogisticegesionwasusedtoaalye factorsflueningitraoperativehypoteria，andanomogam-basedpredictivemodelasonstructed.Results：Patientseredided into the intraoperative hypothermia group （n=91） and the non-hypothermia group （n=143 ）.Lasso-Logistic regression analysis indicated that age ⩾70 years，BMI <24kg/m2 ，surgery duration >2 h，intraoperative blood loss >150mI ，operating room temperature ⩽24% ， intraoperative infused fluid volume >2Ooo mL， intraoperative irrigation fluid volume ， absence of intraoperative warming measures，anesthesia duration >2 h，and adverse actions from operators were independent risk factors for intraoperative hypothermia in RPPSIF patients. The AUC value was 0.862（ 95% CI： 0.832-0.893），with a sensitivity of 77.73% and specificity of 81.97% ： Conclusion：Multipleiskfactorscontribute tointraoperativehypothermiainRPPSIFpatients.Thepredictivemodelconstructedusing Lasso-Logistic regression has significant clinical utility.

KeyWordsLassoLogisicRegreionModel;Robo-asistdPereutanousPedicleSrewixation;Hypothia;Predicidel

脊柱骨折占所有骨折类型的 5%～6% ，其中胸腰椎骨折尤为常见[1。（剩余10837字）