Abstract 13967: Association Between Artificial Intelligence Identified Radiographic Pulmonary Vascular Injury and Cardiac Structure and Function in Patients With Chronic Obstructive Pulmonary Disease

Background Early recognition of cardiac dysfunction in patients with chronic obstructive pulmonary disease (COPD) may prevent future cardiac impairment and improve prognosis. Artificial intelligence (AI) using neural networks on Computed Tomographic (CT) imaging provides a measurement of loss of small vessel volume, as a surrogate of pulmonary vascular injury, and a compensatory increase in large vessel volume. This study investigated pulmonary vascular alteration in patients with COPD and its association with cardiac structure and function assessed by echocardiography. Methods: We studied 205 prospectively enrolled patients with COPD with an available CT scan and echocardiogram. AI algorithms calculated the small - and large vessel fraction as respectively the vascular volume in small vessels (5-10 mm 2 ) and large vessels (20-90 mm 2 ), indexed to total intrapulmonary vessel volume. Univariable and multivariable linear regressions were conducted, and standardized ß-coefficients were calculated. The multivariable model was adjusted by CT reconstruction kernel (spatial resolution). Results: In the unadjusted model, we found that loss in small vessel fraction and increase in large vessel volume were associated with higher left ventricular (LV) mass, LV diastolic dysfunction, and inferior vena cava dilatation. Small vessel fraction was also associated with right ventricular remodeling, while large vessel fraction was associated with higher pulmonary pressure. After adjustment, all measures remained significantly associated, except for tricuspid annular plane systolic excursion which did not. Measures of LV mass and right atrial pressure displayed the greatest associations with pulmonary vascular damage. Conclusion: We found that pulmonary vascular damage assessed by loss in small vessel fraction and increase in large vessel fraction, as assessed by AI on CT imaging, may negatively affect cardiac structure and function in patients with COPD.