George Busby, Allelica CSO & Co-Founder
Allelica, Inc’s risk reports align with disease specific guidelines to communicate an individual’s risk of disease based on their PRS. Risk can be communicated as relative or absolute: for example, for breast cancer we integrate a woman’s PRS value into clinical risk models that include other elements of risk, which combined allow us to provide an estimate of 5 year, 10 year or lifetime absolute risk of disease.
As we’ve written about before, careful considerations need to be made as to how to translate the relative effect of a PRS on a disease into an absolute estimate of risk. These include estimating the underlying incidence of disease in the population to which an individual belongs and how much effect non-PRS based risk factors have on disease.
A framework for estimating absolute risk from PRSs
A new study published in Nature Communications, has described a framework designed to estimate the cumulative incidence of 18 common diseases across different countries, taking into account each individual’s PRS value.
This work builds on earlier strategies that translate PRS into absolute risk by incorporating population-level epidemiological statistics to anchor the effects of PRS with real world data. The concept here is that by understanding what the baseline, population-specific incidence of disease is, as well as population-level disease-specific and all-cause mortality, you can provide a PRS-based estimate of the absolute risk of disease.
The effect of a PRS must be validated and calibrated in different populations
As with any risk factor, PRSs don’t predict guaranteed outcomes or diagnose disease. What, PRS do detect are high risk individuals who are invisible to traditional risk factors and so provide opportunities for physicians to treat those patients, prevent diseases and save their lives. PRSs have also been shown to have greater predictive power compared to other risk factors, in very prevalent diseases such as heart attack and breast cancer.
At Allelica, we strive to ensure health equity when applying PRSs across diverse populations. It’s essential to understand the genetic ancestry of individuals on whom you are applying a PRS. This is because we know that the effect of a PRS needs to be calibrated to different groups; so to be applied effectively, an individual’s ancestry needs to be known.
However, there are other factors, in particular sex and age, which can influence the effect of a PRS. For example, as a general rule, the predictive power of a PRS decreases with age, whilst sex can affect predictions in different ways.
The new study provides a framework for applying any PRS to any population whilst also accounting for potential sex and age differences in a PRSs predictive power.
A Unified Framework for Disease Prediction
The research team developed a unified framework that integrates genetic, epidemiological, and demographic data to estimate how likely individuals are to develop certain diseases over their lifetime. The authors concentrate on 18 diseases, including coronary artery disease, type 2 diabetes, breast cancer, and Alzheimer’s disease, but the framework is extendable to any disease where enough data is available.
The study highlights the transformative potential of PRS in public health. By pinpointing individuals at high genetic risk, healthcare systems can tailor prevention and intervention strategies more effectively, optimizing resource use and improving health outcomes. The research also stresses the need for genetic diversity in studies to ensure these tools are equitable and broadly applicable.
Putting It in Context
This study builds on a growing body of research underscoring the importance of genetic information in health care. A landmark study by Khera et al. (2018) demonstrated the potential of PRS in predicting the risk of common diseases like heart disease and diabetes, emphasizing the added value of genetic data alongside traditional risk factors. Similarly, research by Inouye et al. (2018) showcased how PRS could identify individuals at high risk for coronary artery disease who might benefit from early preventive measures.
The new framework takes these insights a step further by providing a scalable method to apply these genetic insights across different populations globally. This is particularly important because most genetic studies have been predominantly based on European populations, which limits their applicability to other ethnic groups. By emphasizing the inclusion of diverse populations, the new study describes one way in which PRSs can be applied across populations.
However, whilst this framework anchors the effects of PRSs into real world local population data, it does not provide an approach for incorporating other risk factors into these risk estimates. We know that PRS is one of several risk factors for most diseases, but these additional risk factors are not accounted for within this framework.
To enable this sort of integrated risk assessment, Allelica, Inc’s breast cancer and coronary artery disease reports include integration of PRSs for these diseases with established risk models such as the Tyrer-Cuzick for breast cancer and the Pooled Cohort Equations for cardiovascular disease risk.
So however you want to use PRSs in your clinical workflow, we’ve got the solution. Get in touch to find out more.
Allelica, Inc helps physicians to improve the health outcome of their patients using genetics. We do that by developing Polygenic Risk Scores (PRS) that identify more people at high risk who are otherwise invisible to traditional risk assessments
What about you?
Are you using PRSs in your practice? What has your experience been with communicating absolute versus relative risk? Let us know below!
