Efficacy of PET/CT in the characterization of solid or partly solid solitary pulmonary nodules

We aimed to evaluate the efficacy of (18)fluorine fluorodeoxyglucose ((18)F FDG) PET/CT for the characterization of solitary pulmonary nodules (SPNs) compared with the use of PET alone or CT alone. Our institutional review board approved this retrospective study with a waiver of informed consent. We selected 100 patients (M:F=56:44, mean age; 58 years) with a pathologically proven solid or partly solid SPN. Three chest radiologists assessed the nodule characteristics independently and retrospectively. Diagnostic efficacies were compared for three different approaches: consideration of CT findings only, PET findings only, and both PET and CT findings. The McNemar test, kappa statistics, and receiver operating characteristics (ROC) curve analysis were performed. Sixty patients had benign and 40 had malignant nodules. Overall sensitivity values for malignant SPNs for CT, PET and PET/CT were 82%, 88%, and 88%, respectively, whereas the specificity values were 66%, 71%, and 77%, respectively. PET/CT was significantly better in terms of specificity than the use of PET alone or CT alone (P<.05). The areas under curve (Az) values for the ROC analyses of PET/CT and PET alone, respectively, were larger than that of CT alone (P<.05). Interobserver agreement was moderate (kappa=0.46-0.56) for CT, good to excellent (kappa=0.78-0.90) for PET, and good for PET/CT (kappa=0.64-0.78). For the characterization of SPNs, integrated PET/CT provides significantly better specificity than CT alone or PET alone and both integrated PET/CT and PET alone allow more confidence than CT alone.