RESUMO
Objective:To investigate predictive value of model based on pre-surgical 18F-FDG PET/CT metabolic parameters for mediastinal lymph node metastasis (LNM) in lung adenocarcinoma. Methods:A total of 288 patients with lung adenocarcinoma (135 males, 153 females, age (61.6±8.5) years) who diagnosed and treated in the Fourth Hospital of Hebei Medical University from January 2016 to February 2021 were enrolled retrospectively. All patients underwent 18F-FDG PET/CT examination within 1 month before operation, and underwent complete resection of primary lung tumor and standard lymph node dissection. PET/CT parameters were extracted (PET metabolic parameters: minimum SUV(SUV min), SUV max, SUV mean, SUV standard deviation (SUV std), metabolic tumor volume (MTV) and total lesion glycolysis (TLG); CT parameters: minimum CT value (HU min), maximum CT value (HU max), mean CT value (HU mean), CT value standard deviation (HU std)). Multivariate logistic regression analysis was used for screening parameters and establishing model to predict LNM. ROC curves analyses were used to evaluate the predictive performance of models. Results:Among 288 patients, 90 had LNM, and 361 metastatic lymph nodes (N1: 186, N2: 175) were reported by pathology. SUV min (odds ratio ( OR)=1.859, 95% CI: 1.074-3.220, P=0.027), SUV max ( OR=2.255, 95% CI: 1.306-3.893, P=0.004), SUV mean ( OR=0.277, 95% CI: 0.115-0.665, P=0.004) were predictors of LNM. The AUC of PET/CT model was 0.849 (95% CI: 0.804-0.893), and the sensitivity, specificity, accuracy, and positive and negative predictive values were 87.8%(79/90), 72.2%(143/198), 77.1%(222/288), 59.0%(79/134) and 92.9%(143/154), respectively. Conclusion:The model based on 18F-FDG PET/CT metabolic parameters can improve the accuracy of pre-surgical N-staging in patients with lung adenocarcinoma.