| Autor: |
Saha P; The University of Alabama at Birmingham, Computer Science, Birmingham, Alabama, United States., Bodduluri S; University of Alabama at Birmingham, Pulmonary, Allergy and Critical Care Medicine, Birmingham, Alabama, United States., Nakhmani A; University of Alabama at Birmingham, Electrical and Computer Engineering, Birmingham, Alabama, United States., Chaudhary MFA; The University of Alabama at Birmingham Heersink School of Medicine, Pulmonary, Allergy and Critical Care Medicine, Birmingham, Alabama, United States., Amudala Puchakalaya PR; The University of Alabama at Birmingham Heersink School of Medicine, Pulmonary, Allergy and Critical Care Medicine, Birmingham, Alabama, United States., Sthanam V; University of Alabama at Birmingham, Pulmonary, Allergy and Critical Care Medicine, Birmingham, Alabama, United States., San Jose Estepar R; Brigham and Women's Hospital, Radiology, Boston, Massachusetts, United States., Reinhardt JM; University of Iowa, Radiology and Biomedical Engineering, Iowa City, Iowa, United States., Zhang C; University of Alabama at Birmingham, Computer Science, Birmingham, Alabama, United States., Bhatt SP; University of Alabama at Birmingham, Pulmonary, Allergy and Critical Care Medicine, Birmingham, Alabama, United States; sbhatt@uabmc.edu. |
| Abstrakt: |
Rationale Emphysema progression is heterogeneous. Predicting temporal changes in lung density and detecting rapid progressors may facilitate selection of individuals for targeted therapies. Objective To test whether computed tomography (CT) radiomics can be used to predict changes in lung density and detect rapid progressors. Methods We extracted radiomics features from inspiratory chest CT in 4,575 subjects with and without airflow obstruction at enrollment, who completed a follow-up visit at approximately 5 years. We quantified emphysema using adjusted lung density (ALD) and estimated emphysema progression as the annualized change in ALD (∆ALD/year) between visits. We categorized participants into rapid progressors (>1% ∆ALD/year) and stable disease (≤1% ∆ALD/year). A gradient boosting model was used (1) to predict ALD at 5-years and (2) to identify rapid progressors. Four models using demographics (base clinical model); CT density; radiomics; and combined features (clinical, radiomics, and CT density) were evaluated and tested. Results There were 1,773 (38.8%) rapid progressors. For predicting ALD at 5-years in the 20% held-out data, the base model explained 31% of the variance (adjusted R2 = 0.31) whereas R2 was 0.74 for the CT density model, 0.66 for the radiomics-only model, and 0.77 for the combined features model. For detecting rapid progressors, the base model (AUC = 0.57, 95%CI 0.53-0.61) was outperformed by the radiomics-only model (AUC = 0.73, 95%CI 0.69-0.76, ∆ =0.0003, p < 0.001) and the combined model (AUC = 0.74, 95%CI 0.71-0.77, ∆ = 0.0003, p < 0.001). Conclusions Parenchymal and airway radiomics features derived from inspiratory scans can be used to predict temporal changes in lung density and help identify rapid progressors. |
