Risk prediction model construction and validation for postoperative pneumonia in patients with lung cancer
Lu Siyu
Dai Qi
Cao Juan
Xia Yang
Wang Aimei
Liu Hairong
Peng Xue
Ma Tiantian
Abstract：Objective To investigate the incidence of postoperative pneumonia in patients with lung cancer and to develop a risk prediction model based on logistic regression analysis to inform clinical decision-making.Methods A total of 300 pa-tients who underwent radical surgery for lung cancer in the Department of Thoracic Surgery at the First Affiliated Hospital with Nanjing Medical University from March 2022 to June 2024 were enrolled and randomly assigned in a 7∶3 ratio to a modeling group(n=210)and a validation group(n=90).In the modeling group,patients were classified into pneumonia and non-pneumonia subgroups according to whether postoperative pneumonia occurred.Clinical data were collected and lo-gistic regression analysis was used to identify independent risk factors for postoperative pneumonia;a risk prediction model was then constructed based on these factors.Model performance was evaluated by receiver operating characteristic(ROC)curve,calibration curve,and decision curve analysis.Results The overall incidence of postoperative pneumonia in this cohort was 27.67％.Multivariate logistic regression analysis identified older age(OR=1.235,95％CI:1.146-1.332),longer operative time(OR=1.028,95％CI:1.005-1.050),open thoracotomy(OR=8.677,95％CI:2.359-31.912),and postoperative invasive pro-cedures(OR=2.796,95％CI:1.057-7.395)as independent risk factors for postoperative pneumonia,while use of dexmedetomidine was a protective factor(OR=0.229,95％CI:0.079-0.666;P＜0.05).The derived risk prediction model was:P=1[1+e-19.593+0.211×(age)+2.161×(surgical approach)-1.475×(use of dexmedetomidine)+0.027×(operative time)+1.028×(postoperative invasive procedures)],and a nomogram was plotted ac-cordingly.ROC analysis showed an area under curve(AUC)of 0.897(95％CI:0.849-0.945)for the model;calibration curve analy-sis indicated good calibration,and clinical decision curve analysis demonstrated favorable clinical net benefit.Conclusion Age,opera-tive time,surgical approach,dexmedetomidine use,and postoperative invasive procedures are independent factors influen-cing postoperative pneumonia in patients with lung cancer.The validated risk prediction model based on these factors can ef-fectively predict the risk of postoperative pneumonia.
Keywords：Lung cancerSurgeryPneumoniaRisk factorsLogistic regressionPrediction model
Publication Date：2025-09-20
Online Publishing Date：2026-01-06(First online date of this platform, not the publication date of the document)
Pages：8( 67-74 )