Two nomograms based on CT features to predict tumor invasiveness of pulmonary adenocarcinoma and growth in pure GGN: a retrospective analysis

Purpose: The aim of the study is to construct two nomograms for predicting the invasive extent of pulmonary adenocarcinoma and nodule growth in patients with pulmonary pure ground-glass nodules (pGGN).

Method: Consecutive patients with pGGNs (n = 172) were retrospectively studied at one institution, formed the development cohort in predicting IPAs' nomogram. A separate cohort of patients with pGGNs (n = 116) from another institution was used for validation. For the predicting growth nomogram, the primary cohort of patients with pGGNs (n = 80) was from the former institution. We developed the nomogram for predicting IPA using binary logistic regression model, and a Cox multivariable model for the growth nomogram. We assessed nomogram model performance by calibration and discrimination (C-index).

Results: The variables selected in binary logistic regression model (lesion size and shape) had a significant effect on identifying IPA from preinvasive lesion. The C-index of the development and validation cohort were 0.819 (95% CI 0.753-0.874) and 0.811 (95% CI 0.728-0.878), respectively. The risk variables (lesion size, blood vessel types) were selected in the multivariable Cox model. The C-index was 0.880 in the development cohort.

Conclusion: Our nomograms are reliable prognostic methods that can predict the invasiveness of pulmonary adenocarcinomas and the growth of pure GGN in preoperative.