Chronic obstructive pulmonary disease and heart failure — insights from cardiovascular imaging

Background: Chronic obstructive pulmonary disease (COPD) is a leading cause of death worldwide and is frequently associated with heart failure with preserved ejection fraction (associated with diastolic dysfunction). None of the current therapies reduce mortality or slow disease progression for either of these diseases. The classic clinical COPD subtypes (panacinar, centracinar emphysema and others) do not apply to all patients and do not have targeted therapies.

However, recent advances in COPD phenotyping by machine learning have established novel well-defined COPD subtypes based on genetic/molecular data and computed tomography (CT) imaging of the lungs – apical emphysema, diffuse emphysema and combined pulmonary fibrosis emphysema (CFPE). Pilot studies from previous projects at the host institution have suggested possible associations between these new COPD subtypes and specific cardio-pulmonary alterations, such as cor pulmonale, diastolic dysfunction and pulmonary vascular dysfunction.

These pathologies can be quantitatively assessed by thoracic imaging, specifically echocardiography and cardiovascular magnetic resonance (CMR).

Hypothesis: The general aim of this project is to test the hypothesis that there is an association between these novel COPD subtypes and specific cardio-pulmonary alterations assessed quantitatively by thoracic imaging.

Project description: In a well-characterized patient population with COPD, we will call patients for a new visit in which we will perform echocardiography — for characterization of the left ventricle and grading of diastolic dysfunction —, and CMR — for evaluation of cardiac function, estimation of pulmonary artery pressure, and measurement of right venctricular kinetic energy. We will then test the association between some of these variables and the new proposed COPD subtypes.

Significance: This project is the first to include novel applications of both echocardiography and CMR in a large, well-characterized patient population with COPD. These patients have been divided into COPD subtypes based on robust machine learning methods based on both molecular/genetic characteristics and CT spatial lung maps. If the hypothesis listed above are confirmed, these results would be highly significant for two main reasons - (1) they could pave the way to personalize diagnosis and treatment in COPD based on its subtypes, and (2) clarify the link between COPD and heart failure, from a pathophysiologic perspective.