This article has been updated to correct Seamus Harrison's title at Genomics Plc.
NEW YORK – Polygenic risk scores (PRS) are moving one step closer to clinical implementation with the recent publication of cross-ancestry scores for 10 chronic conditions and clinical utility data following next year.
The National Human Genome Research Institute (NHGRI)-funded Electronic Medical Records and Genomics (eMERGE) study aims to make PRS applicable to diverse populations, addressing the challenge that PRS currently work best for populations of European descent, which are most represented in genome-wide association studies (GWAS). The study also seeks ways to improve the interpretation and communication of genetic results to both providers and patients.
In a recent article in Nature Medicine, researchers affiliated with the study published PRS for asthma, breast cancer, atrial fibrillation, chronic kidney disease, coronary heart disease, hypercholesterolemia, obesity, prostate cancer, and both type 1 and type 2 diabetes.
These 10 conditions were selected from an initial list of 23, based on the performance of each PRS, their medical actionability, and their potential clinical utility. From there, the researchers developed a pipeline for clinical PRS implementation, consisting of transferring scores to a clinical laboratory as well as validating and verifying their performance. They furthermore developed a regulatory compliance framework for returning PRS clinical reports to providers and for inclusion in an additional genome-informed risk assessment.
"It’s really exciting to see the eMERGE study describing a pipeline for selection, validation, and return of polygenic risk scores across multiple diseases," Seamus Harrison, VP and head of medical of Genomics Plc, a UK-based firm developing precision health tools based on genetics, said in an email. "The study builds upon a growing body of evidence in the area which shows that we can actually operationalize and scale decades worth of genomic discovery in common diseases to support population health. At some stage soon, we'll have to ask ourselves if it is acceptable not to be using these technologies routinely for the prevention of common, chronic diseases."
The new study represents some of the final stages of the eMERGE Genomic Risk Assessment and Management phase, which launched in 2020. Earlier phases of the project, initiated in 2007, consisted of discovery efforts aimed at evaluating the use of electronic medical records (EMRs) in genomic medicine research. Later stages shifted toward implementation, with efforts to develop methods for returning genomic results to clinicians using EMRs and understanding how a set of 109 genes contribute to an individual's risk of developing a disease by using EMR data to help with the interpretation of genetic results.
"Every phase we have builds on the knowledge we gained from the [prior] phase," said Jodell Jackson, operations director of the eMERGE Network, the organization managing the study across its 10 participating institutions.
"This round takes the knowledge that we've built for the last 15 years and … is kind of the pinnacle of all of the work we've put into it," he said.
The initial 23 conditions selected for trial implementation of PRS reports in this study were based on factors such as having high prevalence and heritability, strength of evidence for PRS performance based on existing literature, clinical actionability, and the potential for health disparities. The research team then evaluated the performance of each PRS across four major ancestry groups: African, Asian, European, and Hispanic.
The cross-ancestry portability of PRS is a key component of the eMERGE study, since the predominance of European ancestry populations in GWAS has meant that PRS resulting from these studies generally work best in those populations, contributing to longstanding ethnic healthcare disparities.
Nearly half of the eMERGE study population, drawn from the National Institutes of Health-funded All of Us Research Program, consists of racial and ethnic minority groups. Additionally, Jackson said that "a handful of percent" of the population comes from rural or low-income settings, representing another underserved group of people.
In contrast to many other studies examining PRS portability across diverse ancestries, the eMERGE team adopted the agnostic approach of defining a participant's ancestry based on genotype, rather than by self-reported group identity.
"One of the main novel things reported in [this] paper is how we deal with the ancestry question," said Niall Lennon, director of the genomics platform at the Broad Institute and co-principal investigator of the study.
The researchers built the PRS models using a diverse set of study data for each of the 10 diseases. To optimize the PRS, they used data from the All of Us program to recalibrate the tails of the PRS distribution, which determine actionability and are particularly important for improving the applicability of PRS to diverse populations.
These procedures, Lennon explained, also help account for mixed ancestry. The researchers built their models using the All of Us cohort, which comprises a large group of people with multiple ancestries, all genotyped with the same technology.
"What we do [with that]," Lennon said, "is create a model that doesn't require us to know ahead of time what your ancestry is."
When a new participant is enrolled, the multi-ancestry PRS is expected to be relevant because it is applicable across a broad range of the genetic ancestry spectrum.
Harrison praised the study's approach to calculating ancestry, while commenting that further improvement is still needed. "We all sit somewhere in the spectrum of genetic diversity, and the frequency of some genetic variants differs in different populations," he said, adding that "the scores reported by eMERGE still generally perform better on those with European ancestry, which reflects the underlying training data."
Once a person's PRS has been calculated, it is then integrated into a report which is returned to patients and their physicians. Part of the eMERGE study consists of ensuring that both physicians and patients can understand these reports.
Josh Peterson, professor of biomedical informatics at Vanderbilt University Medical School and co-principal investigator, acknowledged that the report is "pretty extensive."
"It's a lot of different genetic risks," he said, "so we did our best to make it comprehensible. But we do think that people are going to be challenged to understand it right off the bat."
Genetic counselors, he added, have been involved in the study to help participants and their physicians go through and understand the reports.
Studies on the clinical utility of polygenic risk tests, Lennon said, will provide the kind of data that professional societies will need to determine whether such tests should become standard of care, or that insurers will need to decide if they should cover them or not.
The eMERGE study is not the only one to evaluate the clinical utility of PRS reports.
Genomics Plc recently published a study evaluating the feasibility of incorporating cardiovascular PRS into routine clinical care in northeast England. The company found that in this setting, the reports were generally understood and well accepted by patients and physicians, and in several cases, they triggered changes in care plans, suggesting clinical utility.
Lennon said that the translation and implementation of genetic reports will be the focus of the final part of the eMERGE study, to be published in the coming years.
"Do the healthcare professionals and the patients understand the result, and how do they act on the result? Those are the [topics of the] follow-on papers that will come out over the next two years," he said.
One open question concerning the utility of PRS is how informative they are when combined with other clinical factors. Myriad Genetics, for example, determined in 2022 that combining clinical factors with its RiskScore PRS test for breast cancer led to predictions that better matched real-world observations.
Although the eMERGE study did not attempt to combine PRS with other factors into a single comprehensive score, Peterson said that this may be the focus of future research.
"Polygenic risk gets a lot of attention," Lennon said, "but it's just one factor in a big picture of risk."
For now, the focus of the eMERGE study is on returning the final reports to participants and studying their possible clinical utility. "We're in a sprint to return the reports that are being generated," Peterson said.
After that, he said, the team is "very interested" in observing whether physicians order screening tests based on recommendations contained within the eMERGE reports.
"We'll have until March 2025 to do that observation," Peterson said. "Reports on the outcomes will take a little bit of time after that to come out, but we're close to the end of the study, and that's exciting."