The Value of 18F-FDG PET/CT Mathematical Prediction Model in Diagnosis of Solitary Pulmonary Nodules

摘要

Purpose. To establish an 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) mathematical prediction model to improve the diagnosis of solitary pulmonary nodules (SPNs). Materials and Methods. We retrospectively reviewed 177 consecutive patients who underwent 18F-FDG PET/CT for evaluation of SPNs. The mathematical model was established by logistic regression analysis. The diagnostic capabilities of the model were calculated, and the areas under the receiver operating characteristic curve (AUC) were compared with Mayo and VA model. Results. The mathematical model was y=exp⁡x/[1+exp⁡(x)], x = −7.363 + 0.079 × age + 1.900 × lobulation + 1.024 × vascular convergence + 1.530 × pleural retraction + 0.359 × the maximum of standardized uptake value (SUVmax). When the cut-off value was set at 0.56, the sensitivity, specificity, and accuracy of our model were 86.55%, 74.14%, and 81.4%, respectively. The area under the receiver operating characteristic curve (AUC) of our model was 0.903 (95% confidence interval (CI): 0.860 to 0.946). The AUC of our model was greater than that of the Mayo model, the VA model, and PET (P0.05). Conclusion. The mathematical predictive model has high accuracy in estimating the malignant probability of patients with SPNs.