Original conceptual editorial illustration, not microscopy or study data; colours do not encode risk.
Could comparing embryos by genetic disease risk increase the chance of having a healthy child? It is an understandable question during IVF. But calculating a ranking and demonstrating that decisions based on that ranking improve health are different tasks. A recent study of polygenic embryo screening addresses the first. It does not establish the second. That distinction matters both for families considering testing and for anyone communicating the research.
What do PRS and PES mean?
A polygenic risk score, or PRS, combines information about many DNA variants into a model of susceptibility to a particular condition. Polygenic embryo screening, or PES, uses such information to compare embryos. This paper examined a breast cancer PRS. It was not a comprehensive assessment of a future child’s health, intelligence, development or susceptibility to every disease. Naming the outcome precisely is essential.
A score is not a diagnosis. It does not mean that an embryo is ill, that a future child will develop the condition, or that another child will never develop it. It describes a statistical characteristic within a model and its limitations. Calling an embryo “better” or promising a “healthier child” replaces that limited meaning with a much broader claim. A ranking is also not a judgment about the value of a person.
This is not the same as testing for a single-gene condition
Preimplantation genetic testing for monogenic conditions, usually abbreviated PGT-M, addresses a different question. It concerns particular inherited conditions associated with specific genetic variants. PES concerns complex susceptibility. The researchers used data from PGT-M cycles, but that does not make the two approaches interchangeable. Their presence in the same analysis is not evidence that one can replace the other.
For example, a favourable polygenic score does not cancel the significance of a familial pathogenic variant, including a relevant BRCA variant. Genetic findings need to be interpreted according to what each test measures. Selecting whichever result sounds most reassuring is not an appropriate way to combine them. Counselling should explain these differences before a family is asked to make decisions.
What the researchers actually did
Qi Wang and colleagues analysed family data and embryo data from PGT-M. The family component examined the breast cancer score, while the embryo component explored selection priorities. The work was retrospective. The accepted manuscript in Human Reproduction Open is not a trial in which couples were randomly assigned to selection methods and their children’s health was subsequently compared.
Instead, the researchers simulated choices within available cycles. They compared prioritisation based on embryo morphology with prioritisation based on the lowest PRS percentile, considering successive eligibility stages. This asks how a rule behaves in an existing dataset. It does not show that following that rule during actual treatment improves long-term outcomes.
The measured change concerned the selected embryo’s position in the score distribution. That is a surrogate, not an observed health event. This boundary is central to interpreting the methods and discussion. It belongs in the headline, illustration and short video, not just in a qualification placed at the bottom of an article.
A percentile is not a percentage reduction in disease
A percentile describes a position within a reference distribution. It is not itself the probability of developing breast cancer. Moving to a lower percentile cannot simply be translated into an equally large reduction in disease risk. That translation requires an appropriately validated relationship between the score and the clinical outcome in the relevant population. The study did not demonstrate a reduction in breast cancer among offspring.
Think of two lists that order the same options using different criteria. Choosing a different item when the criterion changes shows that the rules produce different rankings. It does not establish which choice has the better long-term result. This is an explanatory analogy, not a result from the study. The gap between changing a ranking and improving health remains the important unanswered question.
The available embryo pool matters
Ranking requires alternatives. The paper specifically considers a best-prognosis PGT-M context, where several embryos may remain eligible for comparison. This is not the situation of every couple undergoing IVF. A person with very limited options faces different constraints from someone with several embryos that meet the preceding requirements.
Each eligibility step can narrow the pool. Consequently, the result of applying a ranking rule depends on the options left after earlier decisions. It would be misleading to promise the same potential effect to every patient. The simulation also does not justify additional procedures performed solely to enlarge the pool for a polygenic comparison. Such decisions require evidence and an individual clinical discussion, not an extrapolation from an algorithm.
Parental background constrains the comparison
Embryos from the same couple are not unrelated samples drawn from the entire population. Parental genetic backgrounds help shape the distribution of scores available within that family. This does not mean that a child’s future is fixed. A susceptibility score remains uncertain; it is not a statement that a complex condition will certainly be inherited or avoided.
The population context also matters. The research involved a Han Chinese setting and an Asian-population modelling context. A similarly named test should not automatically be assumed to perform identically in every ancestry group. The relevant questions are where a model has been validated, for which outcome, and for which intended use. A familiar acronym is not evidence of universal portability.
What remains unknown?
The paper did not demonstrate healthier children, fewer breast cancer cases among offspring, or an improvement in overall fertility treatment outcomes. This does not prove that a benefit is impossible. It means those benefits were not established by this study. An intuitively attractive rationale should not be presented as if it were already a clinical result.
Patient experience, total costs and the effects on decision-making also deserve attention. Information about uncertain future risks can be difficult to interpret, especially when its implications for action are unclear. Anxiety and unequal access are reasons for careful evaluation, not outcomes that this particular analysis quantified. More information is not automatically more useful information for every family.
Why this explanation does not reproduce every number
Our full-text review identified inconsistencies between some descriptions of participants, tables and figure captions. It is not clear whether these reflect different subsets, features of an early accepted version, editorial issues or another explanation. We therefore do not reproduce the disputed counts or percentages, and we do not build a quantitative graphic from unresolved denominators.
This is not an accusation about the researchers. It is an editorial limit on what can safely be communicated before clarification or a reconciled final version is available. The central interpretation does not depend on those disputed values: a simulation of selection priorities and a change in a surrogate score are not evidence of a clinical benefit in children.
Questions worth bringing to genetic counselling
The following questions are educational suggestions, not a validated counselling protocol. What exactly does the test estimate? Does it concern one specified condition, or is the advertising making a broader promise? In which population was it validated? Is the result a rank, a relative risk or an absolute probability? These distinctions should be explained before a financial decision is made.
It is also reasonable to ask what the result would change. Has the benefit of that change been demonstrated? How will uncertainty be discussed? What happens when different selection criteria point in different directions? A calm conversation should allow room for these questions without implying that declining additional information means neglecting a future child’s health.
The practical takeaway
Understanding when a genetic score changes embryo priorities is a legitimate scientific question. Claiming that the resulting choice produces healthier children requires evidence about health. Combining those statements without explaining the missing step creates unwarranted certainty. Precise language protects families while leaving room for worthwhile research.
This paper informs a particular ranking problem in PGT-M. It does not establish routine PES as a way to ensure healthier children. We can recognise the scientific contribution without making that promise. For a family considering testing, the useful next step is genetic counselling about the purpose, limits and possible consequences of the result. A position in a ranking is not the final answer to a question about health.
Source and transparency
Wang Q et al. A context-specific evaluation of polygenic embryo screening in best-prognosis preimplantation genetic testing cycles. Human Reproduction Open. Accepted manuscript published online 8 September 2026. DOI:10.1093/hropen/hoag078. Open-access full text.
This commentary draws on the methods, results and discussion, not only the abstract. The authors declared no conflicts of interest; funding included CITIC and Hunan provincial institutions. The source is licensed under CC BY4.0. This is independent educational commentary, not an individual treatment recommendation.