Citations

Abstract Rationale Alpha-1 antitrypsin deficiency (AATD) continues to be underdiagnosed. The estimated prevalence of the most common severe deficient genotype, PI*ZZ, is ∼1:3500 or 100,000 Americans but only ∼10,000 such individuals have been recognized. Diagnostic delays persist and these delays are associated with worse disease. High-resolution CT (HRCT) imaging may contain latent phenotypic AATD signatures that are not visibly apparent to clinicians. We hypothesized a deep learning model could identify imaging patterns predictive of AATD from inspiratory HRCT scans. Methods Confirmed genotypes and HRCT scans were pooled from COPDGene, SPIROMICS, and GRADS, yielding 350 MM, 102 MZ, and 69 ZZ participants. Only scans reconstructed with smooth kernels (GE STANDARD, Philips B, Siemens B31f/B35f) and slice thickness < =1.25mm were used. The cohort was selected to have similar distributions of quantitative emphysema, sex, and age by matching five MZ and two MM subjects to each ZZ subject (<10-year age difference, same sex, and <3% difference in %LAA). Tobacco smoke-exposure was more prevalent among the MM/MZ groups given these were predominantly from the tobacco smoke-enriched COPDGene and SPIROMICS cohorts. A DenseNet121 convolutional neural network (CNN) was trained using the MONAI platform to predict genotype by outputting probabilities for MM, MZ, and ZZ. Input to the CNN was a three-channel LAA-910, LAA-950, and LAA-970 map downsampled to 2.0mm3 resolution to remove bias due to differences in scanner and image reconstruction parameters. Five-fold cross-validation was performed. Area under the curve (AUC) was calculated for classification of ZZ vs other genotypes (MM/MZ) and MZ vs MM, with statistical significance assessed at p < 0.05. Results The three groups were similar across age (mean±SD years: MM 57.9±8.8, MZ: 57.5±8.9, ZZ: 55.0±10.7), sex (% female: MM: 48.2%, MZ: 46.7%, ZZ: 47.6), and emphysema (LAA-950%: MM: 13.0±12.2, MZ: 12.4±11.8, ZZ: 13.1±11.9). There was a higher prevalence of ever smokers among MM (100%) and MZ (91.7%), compared to ZZ (47.6%). The model achieved an AUC of 0.806 (p < 0.001) for classification of ZZ vs other genotypes, demonstrating robust discriminatory ability for homozygous AATD. Classification of MZ vs other genotypes did not achieve an AUC significantly greater than 0.5. Conclusions This proof-of-concept study demonstrates that deep learning applied to routine inspiratory HRCT can accurately identify homozygous AATD. These findings support the potential role of imaging-based machine learning as a non-invasive screening tool for undiagnosed ZZ individuals. Future work will focus on cohort enrichment, integration with clinical parameters, and explainability of model-derived imaging features. This abstract is funded by: NIH NHLBI K23HL173570

Abstract Rationale: The D'Nerva Targeted Lung Denervation (TLD) System uses radiofrequency energy delivered bronchoscopically to the main bronchi to disrupt parasympathetic nerve innervation to the lung. AIRFLOW-3 clinical trial was a randomized, double blind, sham controlled study to determine the impact of TLD on COPD exacerbations. Methods: AIRFLOW-3 patients had symptomatic COPD (CAT ≥10) with moderate to very severe airflow obstruction (GOLD stage II-IV: 25% ≤ FEV1 ≤ 80% predicted, FEV1/FVC<0.70) and GOLD E status (≥2 moderate or ≥1 severe exacerbation in the prior 12 months) despite optimal medical therapy. The primary endpoint was a time-to-first event comparison of the probability of subjects having a moderate or severe exacerbation, between the Treatment arm (TLD + optimal medical therapy) and the Sham Control arm (optimal medical therapy). Results: 388 patients were randomized 1:1 at 32 clinical sites across the US, EU, and UK. Patients were 67.8 (SD 7.1) years of age, ex-smokers with a 46.7 (22.0) pack-year smoking history, and 41.0 (11.8) post bronchodilator FEV1% pred. SGRQ-C and CAT scores were 62.5 (17.0) and 23.6 (6.3), respectively. Over 90% of patients were on triple therapy (LAMA+LABA+ICS) at baseline and through 1 year of follow up. 42% of patients had prior-year COPD hospitalization(s). There was no difference between groups in the analysis of the primary endpoint. At 1 year, the sham arm exhibited a statistical decline in lung function (decrease in FEV1 and FVC) and worsening of gas trapping (increase in RV). Lung function and gas trapping were stable in the treatment arm. The treatment arm exhibited a statistically significant improvement in dyspnea at one year compared to sham (mMRC, and TDI). These observations led to the identification of a subgroup of TLD treatment responders corresponding to an airway predominant phenotype (lung hyperinflation in the absence of significant emphysema) that demonstrated clinically and statistically significant improvements in lung function and exacerbation frequency and severity (Table). Conclusions: The AIRFLOW-3 clinical trial failed to meet its prespecified primary endpoint. Post-hoc analysis of this robust dataset led to the identification of a responder profile that achieved significant and sustained improvements in lung function, and exacerbation reduction. A future clinical study is being designed to confirm these findings.

Abstract Rationale Reliable data about the natural history of lung function decline in alpha-1 antitrypsin (AAT)-deficient Pi*MZ heterozygotes is largely missing. Objectives We hypothesized that, in adults with a tobacco smoking history, lung function deteriorates faster in Pi*MZ compared with the Pi*MM genotype. Methods We identified 1,856 Pi*MM and 79 Pi*MZ participants with ⩾20 pack-years tobacco smoking history from the SPIROMICS (Subpopulations and Intermediate Outcomes Measures in COPD Study) cohort by DNA sequencing and followed them over a median of 4.8 years, comparing radiographic and clinical characteristics between the two groups over time using regression models. Results Adjusted for age, sex, race, smoking pack-year history, and smoking status, Pi*MZ participants had a lower baseline percent predicted forced expiratory volume in 1 second (FEV1) (65.4% vs. 75.1%; difference 95% confidence interval [CI], −15.4% to −3.9%), more radiographic emphysema (<−950 Hounsfield units%, 12.9% vs. 7.8%; difference 95% CI, 2.8% to 7.5%), and nonsignificantly lower lung density. In the longitudinal analysis, the FEV1 annual rate of decline was similar in both groups over the course of the study (−34.5 ml/yr vs. −34.6 ml/yr for Pi*MZ and Pi*MM; difference 95% CI, −16.9 to 17.1 ml/yr). There were no significant differences between Pi*MZ and Pi*MM individuals in the annualized change in lung density, emphysema, patient-reported outcomes, exacerbations, or survival. The proportion with faster FEV1 decline (annual loss ⩾40 ml) was similar in Pi*MZ and Pi*MM groups. In both groups, faster FEV1 decline was associated with more air trapping and small airway disease at baseline. In Pi*MZ only, faster decline was associated with higher blood eosinophil counts (310 vs. 220 cells/μl; difference 95% CI, 30 to 140 cells/μl). In the subgroup analysis limited to a small number of currently smoking participants, no significant differences in longitudinal outcomes were found. Conclusions Despite a lower FEV1 and more emphysema at enrollment, the longitudinal analysis did not demonstrate significantly greater lung function decline or lung density loss in Pi*MZ compared with Pi*MM participants with tobacco smoking history. Limited sample size and duration of longitudinal follow-up constrain generalizability of our findings, thus prohibiting the conclusion that longitudinal trajectories did not differ between these groups. However, our results may suggest that earlier life events could be responsible for more extensive lung disease at enrollment in Pi*MZ compared with Pi*MM tobacco-exposed individuals.

Abstract Rationale Dysanapsis refers to a mismatch between airway tree caliber and lung size arising early in life. Dysanapsis assessed by computed tomography (CT) is evident by early adulthood and associated with chronic obstructive pulmonary disease (COPD) risk later in life. Objectives By examining the genetic factors associated with CT-assessed dysanapsis, we aimed to elucidate its molecular underpinnings and physiological significance across the lifespan. Methods We performed a genome-wide association study of CT-assessed dysanapsis in 11,951 adults, including individuals from two population-based and two COPD-enriched studies. We applied colocalization analysis to integrate genome-wide association study and gene expression data from whole blood and lung. Genetic variants associated with dysanapsis were combined into a genetic risk score that was applied to examine association with lung function in children from a population-based birth cohort (n = 1,278) and adults from the UKBiobank (n = 369,157). Measurements and Main Results CT-assessed dysanapsis was associated with genetic variants from 21 independent signals in 19 gene regions, implicating HHIP (hedgehog interacting protein), DSP, and NPNT as potential molecular targets based on colocalization of their expression. A higher dysanapsis genetic risk score was associated with obstructive spirometry among 5-year-old children and among adults in the fifth, sixth, and seventh decades of life. Conclusions CT-assessed dysanapsis is associated with variation in genes previously implicated in lung development, and dysanapsis genetic risk is associated with obstructive lung function from early life through older adulthood. Dysanapsis may represent an endophenotype link between the genetic variations associated with lung function and COPD.

Abstract Rationale Ground-glass opacities (GGOs) in the absence of interstitial lung disease are understudied. Objectives To assess the association of GGOs with white blood cells (WBCs) and progression of quantified chest computed tomography emphysema. Methods We analyzed data of participants in the SPIROMICS study (Subpopulations and Intermediate Outcome Measures in COPD Study). Chest radiologists and pulmonologists labeled regions of the lung as GGOs, and the adaptive multiple feature method (AMFM) trained the computer to assign those labels to image voxels and quantify the volume of the lung with GGOs (%GGOAMFM). We used multivariable linear regression, zero-inflated negative binomial, and proportional hazards regression models to assess the association of %GGOAMFM with WBCs, changes in percentage emphysema, and clinical outcomes. Measurements and Main Results Among 2,714 participants, 1,680 had chronic obstructive pulmonary disease (COPD) and 1,034 had normal spirometry. Among participants with COPD, on the basis of multivariable analysis, current smoking and chronic productive cough were associated with higher %GGOAMFM. Higher %GGOAMFM was cross-sectionally associated with higher WBC and neutrophil concentrations. Higher %GGOAMFM per interquartile range at visit 1 (baseline) was associated with an increase in emphysema at 1-year follow-up visit by 11.7% (relative increase; 95% confidence interval, 7.5–16.1%; P < 0.001). We found no association between %GGOAMFM and 1-year FEV1 decline, but %GGOAMFM was associated with exacerbations and all-cause mortality during a median follow-up of 1,544 days (interquartile interval, 1,118–2,059). Among normal spirometry participants, we found similar results, except that %GGOAMFM was associated with progression to COPD at 1-year follow-up. Conclusions Our findings suggest that GGOAMFM is associated with increased systemic inflammation and emphysema progression.

Chronic obstructive pulmonary disease (COPD) is a disease with marked metabolic disturbance. Previous studies have shown the association between single metabolites and lung function for COPD, but whether a combination of metabolites could predict phenotype is unknown. We developed metabolomic severity scores using plasma metabolomics from the Metabolon platform from two US cohorts of ever-smokers: the Subpopulations and Intermediate Outcome Measures in COPD Study (SPIROMICS) (n = 648; training/testing cohort; 72% non-Hispanic, white; average age 63 years) and the COPDGene Study (n = 1120; validation cohort; 92% non-Hispanic, white; average age 67 years). Separate adaptive LASSO (adaLASSO) models were used to model forced expiratory volume at one second (FEV1) and MESA-adjusted lung density using 762 metabolites common between studies. Metabolite coefficients selected by the adaLASSO procedure were used to create a metabolomic severity score (metSS) for each outcome. A total of 132 metabolites were selected to create a metSS for FEV1. The metSS-only models explained 64.8% and 31.7% of the variability in FEV1 in the training and validation cohorts, respectively. For MESA-adjusted lung density, 129 metabolites were selected, and metSS-only models explained 59.0% of the variability in the training cohort and 17.4% in the validation cohort. Regression models including both clinical covariates and the metSS explained more variability than either the clinical covariate or metSS-only models (53.4% vs. 46.4% and 31.6%) in the validation dataset. The metabolomic pathways for arginine biosynthesis; aminoacyl-tRNA biosynthesis; and glycine, serine, and threonine pathway were enriched by adaLASSO metabolites for FEV1. This is the first demonstration of a respiratory metabolomic severity score, which shows how a metSS can add explanation of variance to clinical predictors of FEV1 and MESA-adjusted lung density. The advantage of a comprehensive metSS is that it explains more disease than individual metabolites and can account for substantial collinearity among classes of metabolites. Future studies should be performed to determine whether metSSs are similar in younger, and more racially and ethnically diverse populations as well as whether a metabolomic severity score can predict disease development in individuals who do not yet have COPD.