Clinical Diagnostic Prediction Model For Mycoplasma pneumoniae Pneumonia in Children

Xiaohua Liu,1,* Hongfei Du,2,* Linyan Zhang,1,* Yang Yu,1 Guixiang Wang,1 Fan Zhong,1 Mingjun Zhang,3 Huacui Huang,1 Ying Xu2 1Department of Medical Laboratory, Xindu District People’s Hospital of Chengdu, Chengdu, 610500, People’s Republic of China; 2Department of Clinical Laboratory, the First Affiliated Hospital of Chengdu Medical College, Chengdu, 610500, People’s Republic of China; 3Department of Medical Laboratory, Jiulongpo District People’s Hospital of Chongqing, Chongqing, 400050, People’s Republic of China*These authors contributed equally to this workCorrespondence: Ying Xu, Department of Clinical Laboratory, the First Affiliated Hospital of Chengdu Medical College, No. 278, Baoguang Road, Xindu District, Chengdu, Sichuan, 610500, People’s Republic of China, Email [email protected] Huacui Huang, Department of Clinical Laboratory, Xindu District People’s Hospital of Chengdu, No. 199, Yuying Road South, Xindu District, Chengdu, Sichuan, 610500, People’s Republic of China, Email [email protected]: The aim of this research is to ascertain the risk determinants associated with Mycoplasma pneumoniae pneumonia (MPP) in pediatric patients diagnosed with community-acquired pneumonia (CAP), as well as to construct predictive models to forecast the incidence of MPP.Methods: This study was conducted at Xindu District People’s Hospital of Chengdu from August 2023 to March 2024. A total of 1030 children aged 0 to 14 years with CAP were enrolled and divided into MPP (n=414) and non-MPP (NMPP, n=616) groups based on diagnostic criteria including MP antibody and MP RNA. Data were collected within 24 hours of admission, including peripheral blood counts, inflammatory markers, and other biochemical parameters. The Logistic+Stepwise, Logistic+Lasso, Logistic+Elastic-net, and Logistic+Ridge were employed to identify risk factors, and were used for variable selection with penalization algorithms. Model performance was evaluated using C-index, sensitivity, specificity, accuracy, recall, and F1 score.Results: The results of prediction model showed that four models had good performance. The area under the ROC curve revealed good predictive ability (AUC > 0.8 in both Logistic model and Experience model), the results of calibration curves indicated a good consistency consistency. Logistic+Lasso model selected 9 key variables for further analysis.Conclusion: We have developed and validated a clinical prediction model in children with Mycoplasma pneumoniae pneumonia (MPP). The model identifies NEUT%, EOS%, HSCRP, ADA, Crea, Urea, HDL, P, and ESR as significant independent predictors. It demonstrated robust discriminative ability and good calibration, offering a practical tool for clinicians to stratify risk and guide early intervention in pediatric patients.Keywords: mycoplasma pneumoniae pneumonia, prediction model, clinical characteristics, children