Clinical T categorization in stage IA lung adenocarcinomas: prognostic implications of CT display window settings for solid portion measurement.

Abstract

Our study aimed at evaluating the prognostic implications of lung and mediastinal CT display window settings for solid portion measurements on the eighth-edition lung cancer staging system's clinical T (cT) categorization. We retrospectively analyzed 691 surgically treated patients from 2009 to 2015 for clinical stage IA lung adenocarcinomas. Solid portions were measured at the lung and mediastinal window settings, respectively, and cT categories were determined for each measurement (cT_lung and cT_mediastinum). The prognostic power of the two cT factors for disease-free survival (DFS) was assessed using Cox regression, and concordance indices (C-indices) were compared using the Student t test. Subsequently, the patients were split into training and validation cohorts to calculate optimal cutoffs for the cT categorization of mediastinal window-based solid portions (cT_optimal) and validate its prognostic performance. Both cT_lung ((cT1b: adjusted HR, 3.547; p = 0.017), (cT1c: adjusted HR, 9.439; p < 0.001)) and cT_mediastinum ((cT1b: adjusted HR, 4.635; p < 0.001), (cT1c: adjusted HR, 11.235; p < 0.001)) were significantly associated with DFS for each multivariable Cox model. The C-indices were 0.772 (95% CI, 0.702-0.842) for cT_lung and 0.787 (95% CI, 0.726-0.848) for cT_mediastinum (p = 0.789). The optimal cutoffs for cT categorization of the mediastinal window-based solid portions were 0.9 cm and 1.8 cm. However, there were no significant differences in the C-indices among cT_lung, cT_mediastinum, and cT_optimal (p > 0.05). The prognostic performances of the cT categorizations at the lung and mediastinal windows were not significantly different. The current cT categorization based on the lung window measurement is appropriate as it stands. • Discriminatory power of the eighth-edition clinical T category was not significantly affected by the CT display window settings. • Given the facts that the lung window setting enables more sensitive detection of the solid portions and higher correlation with the pathological invasive components, our findings may support adherence to the usage of the lung window setting for the solid portion measurement per the current recommendations.