Navigating the Progression of Chronic Obstructive Pulmonary Disease
Year: 2025
Type: editorial
Abstract: Pulmonary DiseaseDetermining where a patient is on their evolving path of disease is the goal of all clinicians.This "quest for longitude," however, is not unique to medicine.In a book of that name, Jonathan Betts describes the navigational challenges facing the seafaring trade in the 17th and 18th centuries due to the inability to determine longitude (1).The losses of life and material were so large that British Parliament established a prize in 1714 for anyone who could provide a solution.This spurred the development of John Harrison's chronometer, which revolutionized waterborne trade by providing the foundation for navigation until global positioning satellite systems became available in the late 20th century.The great challenge facing the respiratory community is the abundance of rich but largely cross-sectional data available for investigation.Although significant investments are being made to continue examinations for multiple cohorts, there is a clear need to better analyze and interpret short-term follow-up data because we cannot wait decades for disease to develop.There are several datarelated "chronometers" designed to tackle this challenge, such as the SuStaIn (Subtype and Stage Inference) algorithm and Bayesianbased disease trajectories (2, 3).Each of these models endeavors to define and impute disease progression using cross-sectional and limited longitudinal data.A new and worthy addition to that effort called "elastic principal graph analyses" (EPGAs) is described by Bell and colleagues (pp.569-576) in this issue of the Journal (4).Using cross-sectional data from almost 9,000 ever-smokers enrolled in the COPDGene (Genetic Epidemiology of COPD) Study ( 5), the present study by Bell and colleagues leveraged 30 clinical, physiologic, and imaging (computed tomography) features to define aspects of chronic obstructive pulmonary disease (COPD) and train their model ( 4).An elastic principal tree was then used to summarize the data and identify branches representing subsets of participants with differing disease characteristics.This semisupervised technique is adept at integrating multiple data types and accommodating missingness.It serves as a tool for dimension reduction, clustering, and visualization (6).Most important, its measure of distance along branches may provide insight into the temporal aspects of disease progression.The result is an integrated display such as what the authors provide in Figure 1.Discrete subtypes, such as chronic bronchitis, airway disease, and emphysema, are represented by branches.Membership in these branches reflects differentiation into a disease subtype with the most severely affected at the tips.
Authors: Carrie Pistenmaa, George R. Washko
Institution: Brigham and Women's Hospital
Language: English
Subfield: Pulmonary and Respiratory Medicine
Field: Medicine
Domain: Health Sciences
Open Access status: hybrid
OpenAlex ID: W4406313168