Radiomic feature analysis of computed tomography images for the diagnosis of invasive pulmonary adenocarcinoma appearing as ground-glass nodules / 中国基层医药

ABSTRACT

Objective:To investigate the value of radiomic features of computed tomography (CT) images in the diagnosis of invasive pulmonary adenocarcinoma appearing as part-solid ground-glass nodules.Methods:The clinical data of 100 part-solid ground-glass nodules from 88 patients with pulmonary adenocarcinoma confirmed by pathological diagnosis who received surgical treatment in Taizhou Tumor Hospital, China between February 2016 and April 2019 were retrospectively analyzed. Among these 100 part-solid ground-glass nodules, 56 from 53 patients were diagnosed as invasive pulmonary adenocarcinoma and 44 from 35 patients as non-invasive pulmonary adenocarcinoma. A set of regular risk factors and visually-assessed qualitative CT imaging features were compared with the radiomic features using logistic regression analysis. Three diagnostic models, i.e., a basis model using the clinical risk factors and qualitative CT features, a radiomics model using significant radiomic features, and a nomogram model combining all significant features, were established and their diagnostic efficacy was compared based on receiver operating characteristic (ROC) curves. Decision curve analysis was performed for the nomogram model to explore its potential clinical benefit.Results:Multiple logistic regression analysis showed that three qualitative CT imaging features (pleural traction ( P = 0.006), solid component size ( P = 0.045) and solid component proportion ( P = 0.020)) and quantitative Rad score ( P = 0.046) were significantly correlated with invasive pulmonary adenocarcinoma. The adjusted ratios were 7.189, 0.075, 194.786 and 2.016, respectively. The diagnostic nomogram model based on these four features showed that the area under the ROC curve (AUC) was 0.903 (95% CI 0.845, 0.975). The diagnostic nomogram model showed a significantly higher performance (AUC = 0.903) in differentiating invasive pulmonary adenocarcinoma from non-invasive pulmonary adenocarcinoma than either the basis model (AUC = 0.853, P = 0.000) or the radiomics model (AUC = 0.769, P < 0.001). Decision curve analysis indicated a potential benefit of using such a nomogram model in clinical diagnosis. Conclusion:Quantitative radiomic features provide additional information regarding clinically-assessed qualitative features for differentiating invasive pulmonary adenocarcinoma from non-invasive pulmonary adenocarcinoma appearing as ground-glass nodules, and a diagnostic nomogram model including all these significant features may be clinically useful in preoperative strategy planning.