Construction and Verification of A Nomogram Model for Predicting Invasive Risk of Ground Glass Nodules

Abstract

Objective To investigate the importance of a nomogram model based on biomarkers and CT signs in the prediction of the invasive risk of ground glass nodules. Methods A total of 322 patients with ground glass nodule, including 240 and 82 patients in the model and verification groups, respectively, were retrospectively analyzed. Independent risk factors for the invasive risk of ground glass nodules were screened out after using single and multiple Logistic analysis. R software was used to construct the nomogram model, and clinical decision curve analysis (DCA), receiver operating curve (ROC), and calibration curve were used for internal and external verification of the model. Results In this study, the independent risk factors for the invasive risk of ground glass nodules included systemic immune-inflammation index (SII), CYFRA21-1, edge, vascular cluster sign, and nodular consolidation tumor ratio (CTR). The area under the ROC curve of the constructed nomogram model had a value of 0.946, and that of the external validation group reached 0.932, which suggests the good capability of the model in predicting the invasive risk of ground glass nodules. The model was internally verified through drawing of calibration curves of Bootstrap 1000 automatic sampling. The results showed that the consistency index between the model and actual curves reached 0.955, with a small absolute error and good fit. The DCA curve revealed a good clinical practicability. In addition, nodule margin, vascular cluster sign, and CTR were correlated with the grade of pathological subtype of invasive adenocarcinoma. Conclusion A nomogram model based on biomarkers and CT signs has good value and clinical practicability in the prediction of the invasive risk of ground glass nodules.

Keywords

• ground glass nodules
• biomarkers
• nomogram model
• ct signs
• pathological subtype
• invasive adenocarcinoma of lung