Your browser doesn't support javascript.
loading
Prediction model for probability of malignancy in solitary pulmonary nodules on 18F-FDG PET/CT of smokers with pulmonary fibrosis / 中华核医学与分子影像杂志
Article in Zh | WPRIM | ID: wpr-884787
Responsible library: WPRO
ABSTRACT
Objective:To establish and validate a malignant risk prediction model of solitary pulmonary nodules (SPNs) with pulmonary fibrosis in long-term smokers based on 18F-flurodeoxyglucose (FDG) PET/CT. Methods:PET/CT images of 222 SPNs combined with pulmonary fibrosis which were shown in integrated CT scan in 169 patients (all males; age 68(63, 75) years) were analyzed retrospectively. All patients were examined in PET/CT Center of the Affiliated Hospital of Qingdao University from January 2011 to December 2019 and all had definite smoking history. The benign and malignant nodules were judged according to the pathological diagnosis or follow-up imaging data of lung lesions (follow-up≥2 years). The clinical characteristics (age, smoking index), morphological characteristics (longest diameter of lesion, density, location, distribution, relative position of fibrosis, spiculation, lobulation, calcification, vacuole, vascular convergence, pleural indentation, emphysema and severity of bilateral pulmonary fibrosis) and metabolic characteristics (maximum standardized uptake value (SUV max)) of the benign and malignant lesions were analyzed by χ2 test and Mann-Whitney U test. Then multivariate logistic regression analysis was applied to select independent risk factors of malignant nodules, and a risk prediction model was established and verified by the area under the receiver operating characteristic (ROC) curve and k-fold cross validation ( k=10) respectively. Results:Among 169 patients, 222 SPNs were detected (157 malignant nodules, 65 benign nodules). Univariate analysis showed that smoking index, speculation, lobulation, vascular convergence sign, calcification, emphysema, nodule size, relative position of nodule and fibrosis, SUV max and severity of bilateral pulmonary fibrosis were significantly different between the benign and malignant nodules ( z values: 2.514-9.858, χ2 values: 4.353-18.442, all P<0.05). Result of multivariate logistic regression analysis showed that calcification, vascular convergence and SUV max were the independent risk factors of malignant nodules combined with pulmonary fibrosis (odds ratio ( OR): 0.048-2.534, all P<0.05). The risk prediction model was as follow: P=1/(1+ e - x), x=-1.839-3.033×calcification+ 0.930×vascular convergence+ 0.754×SUV max(with calcification/vascular convergence=1, without calcification/vascular convergence=0). The area under ROC curve was 0.932(95% CI: 0.895-0.969), and the sensitivity and specificity of the model were 87.9% and 86.2%, respectively. Results of k-fold cross validation showed that the prediction accuracy of 10 test sets was 0.847±0.075, and was 0.862±0.010 in training sets. Conclusions:Calcification, vascular convergence and SUV max are independent risk factors of malignant SPNs combined with pulmonary fibrosis in long-term asymptomatic smokers. The model based on the above variables presents high diagnostic efficiency in diagnosing malignant SPNs.
Full text: 1 Index: WPRIM Type of study: Prognostic_studies / Risk_factors_studies Language: Zh Journal: Chinese Journal of Nuclear Medicine and Molecular Imaging Year: 2021 Type: Article
Full text: 1 Index: WPRIM Type of study: Prognostic_studies / Risk_factors_studies Language: Zh Journal: Chinese Journal of Nuclear Medicine and Molecular Imaging Year: 2021 Type: Article