Abstract 9139: Inspiratory and Expiratory Main Pulmonary Artery Dimensions on Chest CT Exclude Pulmonary Hypertension in Chronic Obstructive Pulmonary Disease

Abstract

Background: Pulmonary hypertension (PH) is present in 30-70% of chronic obstructive pulmonary disease (COPD) and is a strong predictor of morbidity and mortality. Although previous studies have demonstrated that a chest CT-determined main pulmonary artery diameter (MPAD) of ≥ 29 mm during inspiration is associated with PH, these have utilized small groups of heterogenous pulmonary disease. Consequently, the predictive value of MPAD in a pure COPD population remains unknown. In addition, previous studies have not evaluated the diagnostic utility of expiratory phase MPAD. Since variable inspiratory effort and loading of the pulmonary vasculature may confound efforts to predict PH, expiratory phase MPAD measurement may be a reliable indicator of PH. COPDGene is multicenter prospective study that correlates genotype with chest CT COPD phenotype. Using a large subgroup of the COPDGene cohort, we hypothesized that inspiratory MPAD (iMPAD) and expiratory MPAD (eMPAD) on chest CT accurately predict PH when defined by echocardiographic estimation of RVSP in a cohort of COPD patients.

Methods: COPDGene subjects who completed both high-resolution chest CT as well as echocardiography were included for analysis. Cross-sectional iMPAD and eMPAD were measured at the widest dimension at the level of the aorta using the VIDA Pulmonary Workstation 2 software by 3 blinded readers. Based upon the previous World Health Organization definition of mild PH as 40-50 mmHg, a right ventricular systolic pressure (RVSP) ≥ 45 mmHg was classified as PH.

Results: Multivariate regression analysis of 182 COPD subjects revealed significant correlations between RVSP and iMPAD (r .45, p < .0001) as well as eMPAD (r .41, p < .0001). Optimal diagnostic performance of iMPAD occurred at ≥ 28 mm with an AUC of .69 and a NPV of 88% but a PPV of only 38%. In comparison, optimal performance of eMPAD occurred at ≥ 31 mm with an equal AUC (69%) and nearly equivalent NPV (85%) and PPV (42%).

Conclusions: Both iMPAD and eMPAD perform with excellent NPV but poor PPV for prediction of PH in a COPD cohort. In contrast to previous studies indicating good PPV in heterogenous pulmonary disease populations, the diagnostic utility of pulmonary artery diameter measurement in the COPD population may reside in its NPV for PH.