基于logistic-Nomogram构建门诊急救急性肺损伤患者预后预测模型

Construction of a prognosis prediction model for outpatient emergency acute lung injury patients based on a logistic-Nomogram

目的:基于logistic-Nomogram构建门诊急救急性肺损伤(ALI)患者预后预测模型。方法:选取2022年3月至2023年1月航空总医院门诊部收治的308例ALI患者。分析ALI患者入院28 d预后情况,比较不同预后患者的临床资料,通过logistic分析筛查出的预测因素与预后不良的关联性,根据预测因素构建预后不良Nomogram预测模型,并对Nomogram预测模型进行外部验证。结果:死亡组年龄、呼气末正压、肺血管阻力指数(PVRI)、急性生理学与慢性健康状况评分系统Ⅱ(APACHEⅡ)评分,血浆高迁移率族蛋白B1(HMGB1)、单核细胞趋化蛋白-1(MCP-1)、溶性髓系细胞表达的触发受体-1(sTREM-1)水平,血清miR-300、miR-221表达水平及葡萄糖调节蛋白78(GRP78)、趋化因子受体4(CXCR4)、粒细胞集落刺激因子(G-CSF)水平均高于存活组,平均动脉压低于存活组(均P<0.05)。年龄、PVRI、APACHEⅡ评分,血浆HMGB1、sTREM-1、MCP-1,血清miR-300、miR-221表达水平及GRP78、CXCR4、G-CSF水平均为ALI患者预后不良的独立危险因素(均P<0.05),基于以上因素构建ALI患者预后不良的Nomogram预测模型,受试者工作特征(ROC)曲线下面积(AUC)为0.947,灵敏度、特异度分别为91.45%、86.53%;校准曲线显示该预测模型预测结果与实际观测结果一致性较好。结论:基于年龄、APACHEⅡ评分、HMGB1、sTREM-1、MCP-1、miR-300、miR-221、PVRI、GRP78、CXCR4、G-CSF因素构建的预测模型对于ALI具有良好的预测效能。

Objective:To construct a prognosis prediction model for outpatient emergency acute lung injury(ALI)patients based on a logistic-Nomogram.Methods:A total of 308 ALI patients in Aviation General Hospital from March 2022 to January 2023 were selected.The prognosis of ALI patients admitted for 28 days was analyzed,the clinical data of patients with different prognosis were compared,a Nomogram prediction model of poor prognosis was established based on the predictive factors by logistic analysis,and it was validated.Results:In the death group,age,positive end-expiratory pressure,pulmonary vascular resistance index(PVRI),acute physiology and chronic health status score systemⅡ(APACHEⅡ)score,plasma high mobility group protein B1(HMGB1),monocyte chemoattractant protein-1(MCP-1),soluble myeloid cell expression trigger receptor-1(sTREM-1),expression levels of serum miR-300 and miR-221,glucose-regulatory protein 78(GRP78),chemokine receptor 4(CXCR4)and granulocyte colony-stimulating factor(G-CSF)were higher than those in the survival group and the mean arterial pressure was lower than that in the survival group(all P<0.05).Age,PVRI,APACHEⅡscore,plasma HMGB1,sTREM-1,MCP-1,expression levels of serum miR-300 and miR-221 and serum levels of GRP78,CXCR4 and G-CSF were all independent risk factors for poor prognosis in ALI patients(all P<0.05).Based on the above factors,a Nomogram prediction model for ALI patients with poor prognosis was constructed.The area under receiver operating characteristics(ROC)curve(AUC)was 0.947,and the sensitivity and specificity were 91.45%and 86.53%,respectively.The calibration curve showed that the predicted results of the model were in good agreement with the actual observation results. Conclusion: The prediction model based on age, APACHEⅡ score, HMGB1,sTREM-1, MCP-1, miR-300, miR-221, PVRI, GRP78, CXCR4 and G-CSF has good prediction efficiency forALI.