A20-01 One-year Progression in Quantitative CT Parameters Predicts Subsequent Lung Function Decline Over 5 Years: The Subpopulations and Intermediate Outcome Measures in COPD Study (Spiromics)

Background COPD is the fourth leading cause of death worldwide, yet the only approved surrogate endpoints for disease progression are FEV₁ decline and exacerbation rates. SPIROMICS, a multicenter longitudinal study of COPD cases and smoking and non-smoking controls, was designed to identify additional surrogate endpoints of disease progression. Given the high short-term (6-week) reproducibility of quantitative computed tomography (CT) measures, we tested if their 1-year progression predicted subsequent 5-year decline in the FEV₁. Methods SPIROMICS comprised four annual visits and a fifth visit 5-7 years after initial enrollment, all with post-bronchodilator spirometry. At Visits 1 and 2, post-bronchodilator CT scans were acquired at full inspiration and expiration and QC’d following the SPIROMICS CT protocol. Quantitative CT parameters were selected for short-term reproducibility, including novel CT emphysema subtypes (combined bronchitic-apical emphysema [CBaE] and diffuse emphysema), density-based measures (percent emphysema-950HU, volume-adjusted lung density [VALD]), AMFM-derived texture, disease probability measures, pulmonary vascular volume, and segmental airway measures (wall area percent [WA%]). Among participants with acceptable inspiratory scans and 3-4 subsequent valid spirometry tests, we tested associations between progression of CT parameters from Visits 1-2 and subsequent FEV₁ decline in Visits 2-5 using linear mixed-effects models with random intercepts adjusting for demographics, enrollment strata, site, baseline CT measure and time-varying height, weight, smoking status, and pack-years. Results Among 880 participants (43% mild/moderate COPD, 17% severe COPD, 29% smoking control, 11% non-smoking control), one-year increases in CBaE and diffuse emphysema were associated with subsequent accelerated FEV₁ decline over a mean of 4.7 years follow up (SD 0.6) in both unadjusted and adjusted analyses (adjusted: CBaE, −4.7 mL/year/one-year increase SD unit; 95% CI: −8.98, −0.36; diffuse, −5.5 mL/year/one-year increase SD unit; 95% CI: −9.86, −1.19). Percent emphysema-950, VALD, and %WA also predicted FEV₁ decline (all P < 0.05). Comparison of these five parameters demonstrated that only CBaE and diffuse emphysema were independently predictive of FEV₁ decline. Other quantitative CT measures showed limited or no associations. The figure shows one-year change in CBaE, diffuse emphysema and percent emphysema-950 (top vs bottom quartile) and subsequent FEV1 decline in among 375 participants with complete data. Conclusions One-year progression of several quantitative CT parameters, particularly the novel, machine-learned CT emphysema subtypes of CBaE and diffuse emphysema, predict long-term FEV1 decline in COPD. These findings support the potential of quantitative CT parameters to serve as intermediate endpoints in COPD clinical trials.