Understanding Polygenic Risk Scores for Advanced Precision Prevention
Image: Christopher Rayner (Edit Lab Blog - King’s College London)
Your genes are not your destiny, but they do hold clues to your future. As we accumulate gene sequences and pair them with diagnosed disease, we learn those patterns of mutations (or SNPs – Single Nucleotide Polymorphisms) that spell increased risk of disease. There are some mutations, like the one for Huntingdon’s Disease, that confer a certainty of contracting a future condition, and there are others, like the BRCA1 and 2 genes for breast cancer, or APO-E4 for Alzheimer's, where possession of the mutation dramatically increases the risk of this form of dementia. However, more commonly a pattern of mutations that results in a “polygenic risk score” (PRS) indicates an elevated risk of future disease and points to a population that would benefit from more intensive screening, starting earlier and being performed more frequently.
The US company Myriad Genetics scans 94 SNPs and the Estonian company Antagenes scans 77 SNPs, in both cases in order to start mammography earlier and conduct it more frequently. Those with elevated risk can also modify their diets, take more exercise etc to lower risk, but of course knowing that one’s risks are higher can also cause stress, with its countervailing effect.
PRS technology is also being applied to cardiovascular disease to steer precision prevention strategies. The most recent example is an at-home, saliva-based test developed by the London-based Institute of Cancer Research: 40% of the men (ages 55-69) who were found to have the highest risk were subsequently diagnosed with the disease! This may turn out to be a more reliable predictor that the controversial PSA test (where absolute level is less important than increases over time). This all sounds very positive - where are the challenges?
Firstly, the approach is only as strong as the quality and breadth of the genomic information on which it is based. It remains true that too large a share of the data comes from white, Caucasian studies. And so the answers are on less solid ground for other ethnic groups. The second issue is reimbursement – who will pay for these tests? The per-test costs are in the $100-200 range: not too high for individual cases for which the customer woll pay, but a sizeable investment for a total population. Even more fundamentally, health systems don’t generally have established processes to take account of – and act upon - predictive information about the future risks an individual faces.
Most health systems are at full stretch dealing with diseases that have already appeared. An exception may be existing screening programmes for which entry criteria can be modified to prioritise those whose PRS indicates elevated risk. Apart from that, too much will depend on the motivation and persistence of the patient, arranging for a test and consulting their clinicians armed with the result. So precision prediction is now firmly on the horizon, with PRS in the forefront. But there is still a lot of non-scientific work to be done to embed it in the world’s health systems, so we can reap the returns of anticipating and heading off disease rather than waiting for it to appear.
