Lack of Diversity in Pulmonary Fibrosis Genetic Studies Limits Clinical Applicability: Ayodeji Adegunsoye, MD, PhD, MSc

In an interview with The American Journal of Managed Care^® (AJMC^®), Ayodeji Adegunsoye, MD, PhD, MSc, FACP, FCCP, assistant professor of medicine at the University of Chicago, explains that many genetic studies of pulmonary fibrosis have excluded minority populations, limiting the generalizability of their findings.

In the first part of this interview, Adegunsoye explained how MUC5B polymorphisms and rare telomere-related gene variants are changing the way clinicians think about pulmonary fibrosis disease progression. Here, he dives into how the lack of representation in research affects the development and usefulness of tools like polygenic risk scores, which may not apply to underserved racial or ethnic groups. According to Adegunsoye, broader inclusion is necessary to ensure genomic tools are equitable and clinically relevant across all affected populations.

This transcript has been lightly edited; captions were auto-generated.

Transcript

Your research focuses on identifying and optimizing treatment for patients with pulmonary fibrosis across racial groups. How has the exclusion of certain populations from genetic studies seemed to impact outcomes thus far?

I think it's a big factor in how we think about the disease itself. So far to date, there have been large exclusions of certain racial populations [and] certain cultural or ethnic populations in many of the genetic studies to date. And that really limits our understanding of what genetics contributes to the disease pathophysiology itself.

These large exclusions of minority populations limit our ability to translate what we find from 1 population to the entire population, so the generalizability becomes incredibly limited. It also impairs our ability to think about polygenic risk scores. Certain pooled genetic risk factors, which we call risk scores, cannot be applied to other populations because they were not derived in them, and so they wouldn't have external values in them.

I think we need to think about making this test more widescale, more broadly implemented, so that the value we get from them are broadly applicable as well.