Sunday, December 1 | 9:10 a.m.-9:20 a.m. | S1-SSCH01-2 | E451A
Computing thoracic bone and muscle metrics in patients with chronic obstructive pulmonary disease (COPD) improves understanding of the role of certain comorbidities in COPD disease progression, according to findings to be announced in this scientific presentation.
Toward retrospectively characterizing osteoporosis and sarcopenia, researchers applied deep learning-based automated algorithms to inspiratory chest CT images of 1,127 participants from the Genetic Epidemiology of COPD (COPDGene) Iowa cohort at baseline visits.
The predictive algorithms involved generalized additive models (GAMs) that were built by accounting for possible nonlinear effects of demographics and risk variables. This session will review the model formulas of the GAMs, as well as findings in relation to pectoral muscle area (PMA) and spinal volumetric bone mineral density (BMD) metrics.
Study participants were divided according to COPD severity: preserved lung function (n = 520); mild COPD (n = 214), moderate COPD (n = 212); and severe COPD (n = 181).
"Participants with increasing COPD severity had greater negative shifts in BMD and PMA falling in the osteoporosis, sarcopenia, and osteosarcopenia quadrants, which indicates adverse relations of bone and muscle features with COPD," wrote the team that included Eric Hoffman, PhD, from the University of Iowa and Elizabeth Regan, MD, PhD, from National Jewish Health in Colorado.
For the session, the researchers added that spherical distributions of shifts in bone and muscle metrics are suggestive of different pathways of bone and muscle loss in COPD. Stop by Sunday morning to add to your understanding of these aspects of COPD disease progression.