Development and validation of a nomogram based on CT texture analysis for discriminating minimally invasive adenocarcinoma from glandular precursor lesions in sub‑centimeter pulmonary ground glass nodules.

In a recent reclassification, adenocarcinoma has been redefined as a glandular precursor lesion (GPL), alongside adenomatous hyperplasia. This updated classification necessitates corresponding adaptations in clinical diagnostic and therapeutic protocols. Consequently, the present study aimed to construct and validate a nomogram utilizing computed tomography (CT) texture features to effectively discriminate between minimally invasive adenocarcinoma (MIA) and GPL within sub-centimeter pulmonary ground glass nodules (GGNs). To achieve this objective, the present study employed rigorous statistical methodologies, including the Mann-Whitney U test and binary logistic regression analysis, to identify distinguishing features and establish predictive models. Subsequently, the diagnostic performance of these models underwent evaluation through receiver operating characteristic (ROC) curves. The area under the curve (AUC) in ROC curves was compared using DeLong's test. Additionally, the nomogram was constructed using R software and its diagnostic performance was validated through calibration curves. Within both the training and validation datasets, the AUCs were observed to be 0.992 [95% confidence interval (CI): 0.980-1.000] and 0.975 (95% CI: 0.935-1.000), respectively. DeLong's test revealed significant disparities in the AUCs between the nomogram and single-parameter models (P<0.001). Furthermore, calibration curves demonstrated concordance between the training and validation datasets. In conclusion, the application of a CT texture-based nomogram model has demonstrated aptitude in differentiating between MIA and GPL within sub-centimeter GGNs. This model streamlines the identification of optimal surgical interventions and enhances the sphere of clinical decision-making and management.