A clinical model for growth rate-driven prediction of histological subtypes in invasive pulmonary adenocarcinomas

To develop and validate a nomogram for predicting high - grade components in primary lung adenocarcinomas using tumor growth rate and clinical radiological characteristics.

This retrospective study included 735 patients who underwent surgical resection for lung adenocarcinoma. Tumor segmentation was performed on two preoperative CT scans to calculate the specific growth rate (SGR), defined as the natural logarithm of 2 divided by the volume doubling time (VDT) of the tumor. A nomogram was constructed using multivariable logistic regression analysis with significant predictors.

The nomogram included tumor volume, consolidation - to - tumor ratio (CTR), and SGR. The model showed excellent discrimination in the training cohort (AUC = 0.884), validation cohort (AUC = 0.910), and test cohort (AUC = 0.969). Decision curve analysis indicated a higher net benefit compared to imaging or growth models alone.

The nomogram, incorporating clinical and radiological characteristics with tumor growth metrics, accurately predicted high - grade components in primary lung adenocarcinomas. This tool may aid in preoperative planning and personalized treatment. Future work should focus on prospective multic - center validation and exploring automated segmentation methods.