Abstract

Background: To construct the risk prediction model for postoperative pulmonary infection in elderly patients with lung cancer.

Methods: A retrospective study of 240 elderly patients who underwent lung cancer resection from January to December 2018 at the Sun Yat-sen University Cancer Center in Guangzhou was conducted using a self-designed questionnaire. Basis information was collected. The risk factors were identified by univariate analysis and logistic regression analysis. Prediction model was also constructed using logistic regression analysis. ROC curve was plotted to evaluate the predictive performance of the model.

Results: The incidence of postoperative pulmonary infection in elderly patients with lung cancer was 10.0%. Four independent risk factors including heart disease, need for sputum suctioning, BMI and postoperative WBC count, were entered into the regression equation. The risk prediction equation was Z = 2.562 × heart disease + 2.322 × need for sputum suctioning + 2.963 × emaciation + 1.472 × overweight/ obesity + 0.148 × postoperative WBC count-3.747. The area under the ROC curve was 0.827, the Youden index was 0.532, the sensitivity was 70.8%, and the specificity was 82.4%. C

onclusion: Heart disease, need for sputum suctioning, emaciation, overweight/obesity, and increased postoperative WBC count were the risk factors for postoperative pulmonary infection in elderly patients with lung cancer. The risk prediction model constructed in this study had an excellent predictive effect, which was of certain significance for guiding clinical observation and early screening.