Improving risk prediction for cancer

What are the implications of adding new knowledge about genetic information to the health risk equation? What are the challenges for service delivery systems – and society – where effectiveness must be balanced against cost, and possible harms weighed against potential benefits?

The PHG Foundation and researchers from the University of Cambridge investigated these questions as collaborators in the European Commission funded multicentre research initiative, COGS (Collaborative Oncological Gene-environment Study).

Personalising screening through genetics 

Screening programmes make an important contribution to improving public health, but maximising their value often depends on careful targeting of populations - according to the current understanding of who is most at risk. Those at greatest risk benefit from increased surveillance via screening, to improve early detection and treatment of disease, whilst lower risk individuals benefit from avoiding the inconvenience and potential harms of screening.

Using advanced modelling techniques, we looked at how knowledge from nearly 70 genetic variants known to be associated with breast cancer susceptibility can be used to stratify populations into lower, medium and higher risk groups, and whether such information could be used to improve cancer screening programmes. Currently, age is used as a proxy measure of risk, with those at the greatest risk being offered screening.

Including genetic data along with age to predict risk could improve the accuracy of risk prediction and allow screening to be targeted more effectively 

Although the increased risk conferred by each individual genetic variant was low, our findings confirmed (even to sceptics) that the combined information from many variants together provided useful data to refine risk estimates. In the right conditions a more personalised approach to screening, that combines age and polygenetic risk, could increase the number of cancers detected – and prove a cost-effective option.

A new strategy for cancer screening?

As part of our work on this issue we led a series of expert workshops to address how a screening programme that combines age and polygenetic risk should be implemented. As one commented: "this is when it starts to get difficult" , not least because new genetic variants that affect risk continue to be identified, but conclusive evidence of how far more personalised screening can improve health outcomes is absent.

Our 2014 report, Stratified Screening for Cancer sets out the action needed. We are urging decision makers to prepare for the introduction of risk-stratified screening for breast and prostate cancers. Crucially, this must include the gathering of further empirical evidence about whether a risk-based screening approach improves the benefit-harm balance of screening for breast and prostate cancers.

The cost effectiveness of implementation, the success of which is likely to depend on significant adaptation of IT systems, professional education and public awareness campaigns, must also be clarified. In addition, both professionals and public require greater genetic knowledge, so that those offered stratified prevention understand enough to give informed consent, and that health professionals are knowledgeable and able to provide appropriate support. 

Our recommendations and advice are now available to policy makers and stakeholders as they consider how to prepare for a future where genetic and genomic risk stratification becomes an effective and accepted element in routine healthcare provision.

What was COGS?

The European Commission funded Collaborative Oncological Gene-environment Study (COGS) was a multi-centred international project with multiple work packages, which set out to:

  • Define individual risk of breast, ovarian and prostate cancers based on lifestyle factors and genetic variants
  • Improve identification of people at increased risk of these cancers
  • Prevent disease

With support from colleagues at the University of Cambridge, the PHG Foundation led Work Package 7 of COGS - an investigation into the efficacy and cost effectiveness of using genomic and other information in stratified prevention strategies, including the organisational, ethical, legal and social implications that arise.