Carrie Pistenmaa, MD, MS
(She/Her/Hers)
Rank
Assistant Professor
Department
Medicine
Pulmonary and Critical Care Medicine
Authors
Carrie L. Pistenmaa, MD, MS*, Pietro Nardelli, PhD, George R. Washko, MD, MsC, Raul San Jose Estepar, PhD
Principal Investigator
George R. Washko, MD, MsC and Raul San Jose Estepar, PhD
Twitter / Website
Categories
Introduction: Pulmonary vascular structure is important in lung development and may relate to lung function decline.
Methods: The COPDGene Study enrolled ever-smokers with follow-up after 5 and 10 years. On baseline CTs after 3D vascular reconstruction arteries and veins were separated by deep learning technique. Small arteries had CSA <5 mm^2 (BV5a). Arterial fractal dimension (FDa) was measured by box-counting method; higher values indicate increased tree complexity. We defined “sparse” vasculature as lower than predicted BV5a and low FDa (30.3% of sample) and “full” vasculature as greater than predicted BV5a and high FDa (29.3% of sample). Linear regression adjusted for age, sex, race, height, weight, pack-years and baseline smoking status, FEV1 and COPD.
Results: The 6,703 with complete measures were 59.7+/-9.1 years old, 47.8% female, 44.8% had COPD, 51.5% currently smoked; 3,668 had repeat PFTs after 8.5 years on average. Due to a significant interaction by smoking (P=0.017) analyses were stratified. In former smokers, FEV1 decline was faster in those with “sparse” vs. “full” vasculature (-7.4 ml/year, P=0.006). In current smokers, results were in the opposite direction but non-significant.
Conclusions: A combined CT vascular risk assessment identified those at highest risk of FEV1 decline, which was modified by smoking status.