Abstract Rationale The SPIROMICS Study of Early COPD Progression (SOURCE) provides a unique opportunity to investigate early mechanisms underlying COPD by linking biological and imaging markers in younger adults. CT-detected mucus plugs mark disease severity in COPD, but their relevance in younger smokers without obstruction is unknown. We applied a deep learning algorithm to SOURCE baseline CTs to assess associations between mucus plugging, lung function, symptoms, and sputum mucin concentrations. Methods SOURCE (ClinicalTrials.gov Identifier: NCT05033990) is an observational cohort of participants aged 30-55 years with a smoking history of ≥ 10 pack-years. Using a two-stage deep learning pipeline, which combined a 3D nnU-Net airway segmentation model trained on n = 279 inspiratory CT scans from the Airway Tree Modeling (ATM’22) dataset with a lightweight three-layer encoder convolutional neural network, we detected airway obstructions consistent with mucus plugs. We analyzed induced sputum samples for total mucin and individual MUC5AC, and MUC5B concentrations. Associations between mucus plug burden (number of plugs) and mucin concentrations, smoking exposure, and chronic bronchitis diagnosed at the first follow-up were evaluated using negative binomial regression analyses. Results Among n = 657 participants with available CT data, mucus plug counts ranged from 0-11. Lung function categories included GOLD 0 (n = 531), PRISm (n = 35), GOLD 1 (n = 36), and GOLD 2 (n = 25) (Figure 1). Greater mucus plug burden was associated with lower post-bronchodilator FEV1% predicted (r = -0.10, p = 0.02), higher total CAT score (p = 0.05), and a trend toward greater phlegm production (p = 0.068). In negative binomial regression analyses, greater mucus plug burden was significantly associated with higher sputum MUC5AC concentration (p 0.001) and MUC5AC/MUC5B ratio (p = 0.02), but not MUC5B (p = 0.12) or total mucin concentrations (p = 0.27) in this cohort. Higher plug burden was also significantly associated with current smoking (p 0.05), greater pack-years (p 0.001), and diagnosis of chronic bronchitis (p = 0.004). Conclusion In the SOURCE early COPD cohort, increased mucus plug burden was associated with lower lung function, higher symptom burden, elevated MUC5AC concentrations, and increased MUC5AC/MUC5B ratio. These findings suggest that in early disease, MUC5AC likely contributes to abnormal mucus formation and airway mucus plugging that drive airflow limitation and may represent a modifiable therapeutic target in early disease to prevent progression. Quantitative mucus plug assessment could serve as an early imaging biomarker of disease activity and progression in COPD. This abstract is funded by: This work was supported by the National Heart, Lung, and Blood Institute (NHLBI) of the National Institutes of Health (NIH) Grants R01 HL139690 (to CJG) and R01 HL150023 (to MKH, CJG, and CRH), and R01 HL144718 (to JLC, MKH, and FJM), which support the SOURCE/SPIROMICS study.
Ram et al. (Fri,) studied this question.