Abstract Rationale In severe asthma (SA), biologics reduce exacerbations; however, lung function improvement is modest and varies across individuals and phenotypes. Predicting who will experience meaningful improvement remains challenging. One potential contributor is spatial heterogeneity of airway involvement, particularly in small airways, which spirometry may not detect. Parametric Response Mapping (PRM), a voxel-wise analysis of paired inspiratory-expiratory CT scans, provides spatially resolved measures of lung abnormalities. We evaluated whether baseline high-attenuation parenchymal disease on PRM (PRMPD) and measures of small airways disease (PRMfSAD) predict changes in lung function after biologic initiation. Methods Single-center, retrospective cohort of adults with SA initiating biologics. Baseline paired CT scans were classified using PRM into normal parenchyma (PRMNorm), high-attenuation parenchymal disease (PRMPD), and functional small-airways disease (PRMfSAD). Responders and non-responders were defined by post-treatment improvement in FEV₁ % predicted. Bivariate associations used Spearman’s ρ. Linear mixed-effects model estimated the association of baseline PRMPD and PRMfSAD with ΔFEV₁ % predicted, adjusting for age and CT-to-treatment interval. Results 31 adults with SA(mean age 53.4±11.1 years; 55% female; BMI 30.6±6.1 kg/m²). Biologics initiated were mepolizumab (45%), dupilumab (26%), benralizumab (16%), and omalizumab (13%). Mean PRM composition was PRMNorm 58.1±20.9%, PRMfSAD 10.3±11.0%, and PRMPD 15.7±12.7%. Baseline FEV₁ % predicted was 75.1±21.6%. Responders (n = 23) had higher PRMNorm (60.5±23.0% vs 51.3±11.3%; p = 0.042) and lower PRMfSAD (7.7±9.4% vs 17.7±12.6%; p = 0.034) than non-responders (n = 8), with a trend toward lower PRMPD (13.8±10.8% vs 21.0±16.6%; p = 0.058). Representative CTs/PRM maps for a responder and non-responder are shown (Figure 1A). FEV₁% predicted did not differ (73.4±22.5% vs 79.8±19.5%; p = 0.51).Overall, lung function improved after treatment (ΔFEV₁ % predicted 6.97±12.62; ΔFEV₁ absolute value 0.14±0.36 L), and annualized exacerbations declined from 5.74±4.04 to 1.29±2.31. Responders showed greater improvement (ΔFEV₁ % predicted 12.74±8.61 vs -9.63±5.10, p 0.001; ΔFEV₁ absolute value 0.30±0.26 L vs -0.30±0.17 L, p 0.001).Higher PRMPD was associated with smaller FEV₁ improvement (ρ=-0.438, p = 0.014; Figure 1B), while PRMfSAD was not significant (ρ=-0.190, p = 0.307). In the adjusted models controlling for age and CT-to-treatment interval, PRMPD remained significant (β = −0.35; 95% CI − 0.69 to − 0.01; p = 0.043); each 1% increase in PRMPD corresponded to 0.35% less FEV₁ improvement. PRMfSAD was not significant (β = 0.05; 95% CI -0.48 to 0.57; p = 0.856; Figure 1C). Conclusions In SA, baseline PRMPD independently associated with diminished post-biologic lung function improvement, whereas PRMfSAD did not. PRMPD may have the potential to serve as an imaging biomarker to identify patients less likely to achieve lung-function improvement despite overall clinical benefit. This abstract is funded by: Funding/Product support for this study was provided by GSK (NCT06922760). GSK was provided the opportunity to review a preliminary version of this publication for factual accuracy, but the authors are solely responsible for the final content and interpretation.
Aslam et al. (Fri,) studied this question.