Performance of polygenic risk scores in screening, prediction, and risk stratification: secondary analysis of data in the Polygenic Score Catalog
Compose Response

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Author Information
First or given name, e.g. 'Peter'.
Your last, or family, name, e.g. 'MacMoody'.
Your email address, e.g.
Your role and/or occupation, e.g. 'Orthopedic Surgeon'.
Your organization or institution (if applicable), e.g. 'Royal Free Hospital'.
Statement of Competing Interests


  • A rapid response is a moderated but not peer reviewed online response to a published article in a BMJ journal; it will not receive a DOI and will not be indexed unless it is also republished as a Letter, Correspondence or as other content. Find out more about rapid responses.
  • We intend to post all responses which are approved by the Editor, within 14 days (BMJ Journals) or 24 hours (The BMJ), however timeframes cannot be guaranteed. Responses must comply with our requirements and should contribute substantially to the topic, but it is at our absolute discretion whether we publish a response, and we reserve the right to edit or remove responses before and after publication and also republish some or all in other BMJ publications, including third party local editions in other countries and languages
  • Our requirements are stated in our rapid response terms and conditions and must be read. These include ensuring that: i) you do not include any illustrative content including tables and graphs, ii) you do not include any information that includes specifics about any patients,iii) you do not include any original data, unless it has already been published in a peer reviewed journal and you have included a reference, iv) your response is lawful, not defamatory, original and accurate, v) you declare any competing interests, vi) you understand that your name and other personal details set out in our rapid response terms and conditions will be published with any responses we publish and vii) you understand that once a response is published, we may continue to publish your response and/or edit or remove it in the future.
  • By submitting this rapid response you are agreeing to our terms and conditions for rapid responses and understand that your personal data will be processed in accordance with those terms and our privacy notice.

Vertical Tabs

Other responses

  • Published on:
    Performance of polygenic risk scores in screening, prediction, and risk stratification - reply to rapid responses
    • Aroon D. Hingorani, Professor and Honorary Consultant UCL and UCL Hospitals
    • Other Contributors:
      • Nicholas J. Wald, Professor

    Reply to rapid responses
    Contrary to Padrik and colleagues assertion, we did examine polygenic risk scores for coronary artery disease and breast cancer together with conventional risk factors and screening tests (see Table 1 and Figure 7 of the paper) [1]. We showed that, among women aged 50, a polygenic risk score above the 95th or 97.5th centiles had minimal value in screening. Padrik and colleagues question the use of the 97.5th centile, because it detects only 6% of cases. Selecting the 80th centile instead, the detection rate would be higher at 32% (68% of cases missed) but so would the false positive rate (20%), giving a likelihood ratio of only 1.5, which is poor screening performance.

    Both Padrik and colleagues, and Evans, refer to the fourfold gradient of risk across the range of polygenic risk scores for breast cancer and infer clinical utility. As has been shown before [2–4], and in our paper, much higher relative risk differences are needed to achieve worthwhile medical screening. Padrik and colleagues and Evans envisage polygenic risk scores being used in combination with other risk factors, but such an addition confers a very small improvement in screening performance, as previously reported and explained [5].

    Evans’ analysis of women attending for mammography as part of the PROCAS study [6] can be reduced to the TC8-DR-SNP313 risk model incorporating polygenic risk scores identifying 42% (144/340) of interval breast cancers among wom...

    Show More
    Conflict of Interest:
    ADH is a member of the Advisory Group for the Industrial Strategy Challenge Fund Accelerating Detection of Disease Challenge, and a co-opted member of the National Institute for Health and Care Excellence Guideline update group for ‘Cardiovascular disease: risk assessment and reduction, including lipid modification, CG181. ADH is a co-Investigator on a grant from Pfizer to identify potential therapeutic targets for heart failure using human genomics. NJW is a director of Polypill Limited, a company that provides an online cardiovascular disease prevention service accessed on
  • Published on:
    Breast Cancer Polygenic Risk Scores Have Satisfactory Performance and Clear Clinical Utility
    • Peeter Padrik, Oncologist Tartu University Hospital; OÜ Antegenes
    • Other Contributors:
      • Siim Sõber, Geneticist
      • Robert Koesters, Molecular biologist

    The authors analysed performance metrics for 926 polygenic risk scores (PRSs) for 310 diseases (1). To assess the potential clinical impact of PRSs, it is essential to analyse each disease separately within its specific clinical context. The authors examine coronary artery disease and breast cancer as illustrative examples, but not in the full context of current prevention and screening approaches which are far from perfect. PRSs are not the primary screening methods for diseases; instead, they are supplementary tools that enhance disease risk assessment.
    The UK NHS Breast Screening Program offers routine screening to women between the ages of 50 and 70. Accordingly, risk stratification for screening uses age as a risk factor only. Independent UK Panel on Breast Cancer Screening concluded that a relative risk reduction from screening was 20%, for 10,000 women invited to screening, from age 50 for 20 years, it is estimated that 681 cancers will be diagnosed, of which 129 will represent overdiagnosis and 43 deaths from breast cancer will be prevented (2). We can say the vast majority of women are screened without any benefit but with harms. On the other side, 18% of breast cancer cases are diagnosed among women younger than age 50, they never reach screening (3). The authors state that reducing the age cut-off value for mammography for all women without determining their PRS might be more sensible. However, this non-risk-stratified approach makes screening even more no...

    Show More
    Conflict of Interest:
    Dr. Peeter Padrik has ownership in the health-tech company OÜ Antegenes.
  • Published on:
    • Harry Hill, Researcher The University of Sheffield

    The poor epidemiological performance of polygenic risk scores, as outlined in this paper, is not a reason to dismiss the consideration of such tests. We know that it is only by a comparative assessment to other risk assessment approaches (and even in addition to other risk assessment approaches) can a decision be reached as to whether polygenic risk tests are worthwhile use of NHS resources. So what remains to be done is an evaluation of whether the costs associated with genetic testing are justified considering the additional improvement in risk assessment performance. Even evidence from an economic evaluations in this area do not alone constitute definitive proof of the merits of polygenic risk assessment. This is because the economic advantages of added enhancement in risk assessment accuracy are closely linked to the economic benefits from screening programs provided to individuals at varying risk levels, as well as the number of risk groups in the population, factors that are independent of the accuracy of polygenic risk scores.

    Conflict of Interest:
    None declared.
  • Published on:
    Response to: Performance of polygenic risk scores in screening, prediction, and risk stratification: secondary analysis of data in the Polygenic Score Catalog and press release
    • D Gareth Evans, Ptofessor of Medical Genetics and Cancer Epidemeiology University of Manchester

    Hingorani et al {1] purport to show that polygenic risk scores (PRS) are not fit for purpose as they ‘performed poorly in population screening, individual risk prediction, and population risk stratification.’ Using breast cancer as one of two examples they use a PRS to show 10-year odds of becoming affected for individuals aged 50 with a polygenic risk score at the 2.5th, 25th, 75th, and 97.5th centiles were 1:91, 1:56, 1:34, and 1:21 for breast cancer with the upper 2.5% of the distribution, contributing 6% of cases. Many other known risk factors are easily available and can be used alongside a PRS in existing risk models such as CanRisk and Tyrer-Cuzick. Their assertions about not using a PRS are tantamount to saying you should not use any risk factor information to assess risk. A woman with a PRS of 2-fold will be at average risk if she has no family history of breast cancer and early first pregnancy and late menarche. Using other risk factors would substantially reduce the ‘false positives’ alluded to in the article as well as identifying more at true ‘high risk’. We have shown that using all available risk information including standard risk factors age and mammographic density as well as a PRS that 43.5% of breast cancers in women of screening age can be identified in 19.9% of the population at ≥5% 10-year risk in women of screening age (46-73years) [2]. Using the NICE defined 10-year threshold of ≥5% was also able to identify 48.5% of all stage 2 or higher cancers....

    Show More
    Conflict of Interest:
    I have received consultancy fees from Everythinggenetic Ltd and Astrazeneca