摘要
目的探讨基于多组学分析构建的列线图模型对重症肺炎支原体肺炎(SMPP)并发塑形性支气管炎(PB)的早期预测价值。方法回顾性分析2020年1月至2024年12月住院的216例SMPP患儿临床资料,其中116例并发PB(PB组),100例未并发PB(非PB组)。比较两组患儿的临床特征、影像学表现、肺泡灌洗液病原学检测及实验室指标。采用最小绝对收缩和选择算子(LASSO)回归和基于梯度提升树改进的机器学习(XGBoost)算法进行变量筛选,并通过多因素Logistic回归分析筛选独立影响因素。基于筛选出的关键因素构建列线图预测模型,并采用校准曲线、决策曲线分析及受试者工作特征(ROC)曲线评估模型的校准度、临床实用性和区分度。结果PB组病程、发热天数更长,首次就诊前最高体温较高,出现肠胃症状、低氧血症、呼吸困难、呼吸音减弱、热峰>40℃、肺外损伤、≥2/3肺叶实变、胸腔积液、肺不张、肺坏死和混合感染的患儿较多,白细胞介素-4(IL-4)、干扰素-γ(IFN-γ)、白细胞介素-4与干扰素-γ的比值(IL-4/IFN-γ)、中性粒细胞百分比(NEUT%)、降钙素原(PCT)、丙氨酸氨基转移酶(ALT)、天门冬氨酸氨基转移酶(AST)、乳酸脱氢酶(LDH)、C反应蛋白(CRP)、D-二聚体(D-D)、肌酸激酶(CK)、白细胞介素-6(IL-6)水平较高,巨噬细胞百分比(Mφ%)、淋巴细胞计数(LYM)、血小板计数(PLT)、白蛋白(ALB)水平较低(P<0.05)。LASSO回归和XGBoost算法筛选出病程、混合感染、LDH、肺泡灌洗液中IL-4和Mφ%为关键变量,多因素Logistic回归分析证实各变量为独立影响因素(P<0.05)。基于多因素Logistic回归分析构建列线图预测模型,预测模型的校准曲线显示模型拟合度较好,决策曲线分析表明模型具有较高的临床净收益,ROC曲线显示模型的曲线下面积(AUC)为0.926(95%CI:0.882~0.970),准确率为94.9%,敏感度为89.0%,特异度为100%。结论基于病程、混合感染、LDH、肺泡灌洗液中IL-4和Mφ%构建的列线图预测模型有较高的准确率、敏感度和特异度,可为SMPP患儿PB的早期识别和干预提供量化工具。
Objective To explore the early predictive value of a nomogram model constructed based on multi-omics analysis for plastic bronchitis(PB)complicating severe Mycoplasma pneumoniae pneumonia(SMPP)in children.Methods Clinical data of 216 children with SMPP hospitalized at the Affiliated Hospital of Binzhou Medical University from January 2020 to December 2024 were retrospectively analyzed,including 116 cases complicated with PB(PB group)and 100 cases without PB(non-PB group).Clinical characteristics,imaging findings,bronchoalveolar lavage fluid(BALF)etiological tests,and laboratory indicators were compared between the two groups.LASSO regression and XGBoost algorithms were used for variable selection,and independent influencing factors were screened by multivariate Logistic regression analysis.A nomogram prediction model was constructed based on the selected key factors,and calibration curves,decision curve analysis,and receiver operating characteristic(ROC)curves were used to evaluate the calibration,clinical utility,and discrimination of the model.Results Compared with the non-PB group,the PB group had longer disease duration and fever duration,higher maximum body temperature before the first visit,and more children with gastrointestinal symptoms,hypoxemia,dyspnea,diminished breath sounds,fever peak>40℃,extrapulmonary injury,≥2/3 lobar consolidation,pleural effusion,atelectasis,pulmonary necrosis,and mixed infection.The levels of IL-4,IFN-γ,IL-4/IFN-γ,neutrophil percentage,PCT,ALT,AST,LDH,CRP,D-D,CK,and IL-6 in the PB group were higher,while the levels of macrophage percentage,lymphocyte(LYM),platelet(PLT),and albumin(ALB)were lower(all P<0.05).LASSO regression and XGBoost algorithms identified disease duration,mixed infection,LDH,IL-4 in BALF,and macrophage percentage in BALF as key variables,which were confirmed as independent influencing factors by multivariate Logistic regression analysis(all P<0.05).The nomogram prediction model constructed based on multivariate Logistic regression analysis showed good fit by calibration curve,high clinical net benefit by decision curve analysis,and an area under the ROC curve(AUC)of 0.926(95%CI:0.882-0.970),with an accuracy of 94.9%,sensitivity of 89.0%,and specificity of 100%.Conclusion The nomogram prediction model constructed based on disease duration,mixed infection,LDH,IL-4 in BALF,and macrophage percentage in BALF has high accuracy,sensitivity,and specificity,which can provide a quantitative tool for the early identification and intervention of PB in children with SMPP.
作者
韩婷婷
周甜
马莲美
孙建建
樊洪静
HAN Tingting;ZHOU Tian;MA Lianmei(Department of Pediatric Respiratory Medicine,Binzhou Medical College Affiliated Hospital,Binzhou,Shandong Province 256600,P.R.China)
出处
《临床放射学杂志》
北大核心
2026年第1期137-144,共8页
Journal of Clinical Radiology
基金
滨州医学院科研计划与科研启动基金项目(编号:BY2022KJ35)。
