Establishment and validation a prediction model for discrimination of invasive adenocarcinomas for patients with peripheral pulmonary subsolid nodules.

Abstract

The optimal management of patients with subsolid pulmonary nodules is of growing clinical concern. This study sought to develop and validate a more precise predictive model to evaluate the pathological invasiveness of patients with lung peripheral subsolid nodules (SSNs). The data of 1,140 patients with peripheral SSNs who underwent surgical resection at Shanghai Renji Hospital from January 2014 to December 2018 were retrospectively analyzed. The patients were randomly assigned to a training and validation cohort (at a ratio of 2 to 1). Clinical parameters and imaging features were collected to estimate the independent predictors of pathological invasiveness of SSNs. A nomogram model was developed and applied to the validation cohort. The predictive performance of the nomogram model was evaluated by a calibration curve analysis, an area under the receiver operating characteristic curve (AUC) analysis, and a decision curve analysis (DCA), which was also compared with other diagnostic models. In the multivariate analysis, the nodule diameter (P<0.001), solid component size (P<0.001), mean CT attenuation (P=0.001), spiculation (P<0.001), and pleura indentation (P=0.011) were identified as independent predictors of the pathological invasiveness of SSNs. A nomogram model based on the results of the multivariate analysis was developed and showed a robust discrimination in the validation cohort, with an AUC of [0.890; 95% confidence interval (CI), 0873-0.907], which was higher than another two reported models. The calibration curve showed optimal agreement between the pathological invasive probability as predicted by the nomogram and the actual probability. We developed and validated a nomogram model to evaluate the risk of the pathological invasiveness for patients with lung SSNs. The AUC of this nomogram model was higher than another two reported models. Our nomogram model may help clinicians to make individualized treatment more precisely.