Predictive Values of Quantitative Computed Tomography Parameters for Spread Through Air Spaces in Patients with Invasive Lung Adenocarcinoma

Objective: This study aimed to predict spread through air spaces (STAS) in patients with invasive lung adenocarcinoma based on quantitative computed tomography (CT) parameters. Methods: The clinical and CT data of patients with invasive lung adenocarcinoma confirmed by surgical pathology at the Affiliated Suzhou Hospital of Nanjing Medical University between September 2020 and June 2023 were retrospectively collected. The quantitative parameters of invasive lung adenocarcinoma were obtained using the InferRead CT Lung 4.0 artificial intelligence software. The quantitative CT parameters of the STAS-positive and negative groups were compared. The parameters with significantly different values for both groups were evaluated for diagnostic performance using the area under the receiver operating characteristic curve and assessed for collinearity. A quantitative CT parameter prediction model for STAS in patients with invasive lung adenocarcinoma was established. Results: A total of 148 patients with invasive lung adenocarcinoma were included. Forty-six of them had STAS, while 102 did not. Significant differences were observed in the following parameters between the two groups: median CT value, mean CT value, proportion of solid components, minimum CT value, mass, axial maximum diameter, axial minimum diameter, axial mean diameter, axial maximum area, maximum CT value, 3D long diameter, surface area, volume, standard deviation of CT value, distance from pleura, compactness, sphericity, kurtosis, and skewness. The AUC values for these parameters were 0.941, 0.937, 0.936, 0.926, 0.839, 0.817, 0.799, 0.815, 0.796, 0.795, 0.789, 0.784, 0.775, 0.373, 0.351, 0.313, 0.312, 0.215, and 0.149, respectively. Binary logistic regression revealed the proportions of solid components and skewness as independent predictors of STAS (all P< 0.05). Their combined area under the curve was 0.943. Conclusion: Quantitative CT parameters have significant predictive values for STAS in patients with invasive lung adenocarcinoma. The consolidation tumor ratio and skewness are independent risk factors of STAS.