Construction of a risk prediction model for isolated pulmonary nodules 5–15 mm in diameter

Background: Based on current technology, the accuracy of detecting malignancy in solitary pulmonary nodules (SPNs) is limited. This study aimed to establish a malignant risk prediction model for SPNs 5–15 mm in diameter. Methods: We collected clinical characteristics and imaging features from 317 patients with SPNs 5–15 mm in diameter from the 900th Hospital of the Joint Logistic Support Force as a training cohort and 100 patients with SPNs 5–15 mm in diameter as a validation cohort. Univariate logistic regression analysis, least absolute shrinkage and selection operator (LASSO), and binary logistic regression analysis were used to screen for the independent influencing factors of benign and malignant SPN and to establish a prediction model for benign and malignant SPN with a diameter of 5–15 mm. The model in this study was compared with the Mayo model, Veterans Affairs (VA) model, Brock model, and Peking University People’s Hospital (PKUPH) model. Finally, the clinical application value of this model was assessed. Results: Univariate logistic regression analysis showed that smoking history, nodule diameter, nodule location, nodule density, margin, calcification, lobulation sign, spiculation sign, and vascular cluster sign were statistically significant factors. The results of LASSO and binary logistic regression analysis showed that smoking history, nodule diameter, nodule density, margin, lobulation sign, and vascular cluster sign were independent influencing factors of SPNs. The prediction model was successfully constructed and demonstrated a good predictive performance, with an area under the curve (AUC) value of 0.814 [95% confidence interval (CI): 0.768–0.861; P<0.001] in the training cohort and 0.864 (95% CI: 0.794–0.934; P<0.001) in the validation cohort. This model was shown to be highly accurate in predicting malignant SPNs and thus has a high clinical application value. Compared with previously described prediction models, including the Mayo model, VA model, Brock model, and PKUPH model, the proposed model demonstrated a significantly superior predictive ability. Conclusions: The prediction model developed in this study can be used as an early screening method for SPNs 5–15 mm in diameter.