What can imaging-based screening programmes learn from each other?

Earlier this year, Eleanor Wheeler from the Network Secretariat was delighted to be part of a panel discussion moderated by Mireille Broeders, professor of personalised cancer screening, as part of Cancer Care Day at the 2024 European Congress of Radiology (ECR). She was joined on the panel by Ritse Maarten Mann, an expert in breast cancer screening, and Hein Van Poppel, a leader in prostate cancer screening. The panellists discussed the history of cancer screening in the EU and the advances that many hope will be catalysed following last year’s update of EU screening guidelines, with a focus on three cancer screening programmes that use imaging. This co-authored blog post distils some of the valuable insights that resulted from the discussion.

A (very) brief review of screening in the EU

Until last year, there was only one established imaging screening programme in Europe: mammography, for breast cancer. Now there are two more where implementation is being considered and, in some cases, is already underway: low-dose computed tomography (LDCT) for lung cancer and magnetic resonance imaging (MRI, used alongside laboratory tests) for prostate cancer. In addition, exploring other imaging modalities for breast cancer screening is the subject of ongoing discussion.

In the mid-1980s, the Europe Against Cancer programme aimed to reduce cancer mortality by 15% from expected levels by 2000. And in 2003 the Council of the European Union issued recommendations calling for EU countries to implement national, population-based screening programmes for breast, cervical and colorectal cancer. By 2007, many EU Member States were piloting or starting screening roll-out, and in 2016 many locations in the EU had established breast, cervical and colorectal cancer screening programmes.

In 2021, the European Commission presented Europe’s Beating Cancer Plan; it is made up of multiple pillars, with early detection being a vital component. One of the plan’s aims is to increase access to screening, and we hope to see this happen as a result of the recently updated EU screening recommendations, which raised the number of cancer types recommended for screening from three to six; lung, prostate and gastric cancer (in some sites) are now included. Together, these six cancers account for more than half of the burden of new cancer diagnoses and deaths in the EU.

One way that the EU is continuing to support the introduction of new screening programmes is via the EU4Health projects, including the SOLACE consortium, with which the Network is involved. And experts in the EU and beyond are exploring opportunities to learn from each other, to optimise existing programmes and guide the roll-out of new ones.

A transition from population-based screening to risk-based screening

We are seeing a paradigm shift for screening, from a ‘one-size-fits-all’, population-based approach to a more personalised risk-based approach. This shift is new for breast cancer, but for lung cancer and prostate cancer it is the way that screening programmes are being designed from the start.

We’re also seeing technological advances, such as the greater use of biomarkers, the development of new tests that can detect multiple types of cancer, and the increasing use of AI. These could play a role in reducing overdiagnosis, enabling more personalised approaches to establish eligibility criteria and screening recall. However, the utilisation of these advances relies on health system decision-makers being able to fully assess potential benefits, harms and impact – for example through the results of randomised controlled trials – as well as developing appropriate protocols to build new technologies into new or existing programmes.

Ritse reflected that while it is positive that the early detection of breast cancer is addressed by a well-established screening programme, current EU breast cancer screening programmes lack structured risk assessment and stratified screening. As such, it might be more difficult for established programmes to adapt to new imaging modalities and AI, among other technologies.

The shared challenge of engaging potential screening participants

Since wide-scale national screening programmes for lung and prostate cancer are yet to be comprehensively established across the EU, a key challenge for newly emerging programmes is how to engage potential participants. One suggestion discussed was to explore, as a community of screening advocates, opportunities to generate more shared messaging about the value of screening, and to work together to streamline access.

Building on this idea, could we imagine a scenario where people see cancer screening in the same way as they view regular dental check-ups? Positioning screening as a health check-up can be seen in the UK’s Targeted Lung Health Check programme.

Generating shared messaging about screening could streamline information, potentially make messages clearer, and make more efficient use of resources to engage potential participants. However, we should be mindful that there are different eligibility criteria for some screening programmes, so a degree of personalisation would be required in practice.

The importance of ensuring adherence to screening

There is a shared challenge across screening programmes of achieving good engagement uptake and adherence by the target populations.

While there is broadly a good uptake for breast screening, in recent years there have been concerns about falling engagement. Hein shared that we can learn a lot about different approaches to engagement from prostate screening programmes set up in Czechia, Estonia and Sweden, and pilot studies that will soon start in Spain, Lithuania, Poland and Ireland. These anticipated pilots will be funded by the European Commission, within the PRAISE-U consortium, and more information can be found here.

There are also lessons from lung cancer screening. Programmes are being developed in collaboration with underserved and high-risk communities to try to optimise engagement and adherence at the earliest stages of programme design. In addition, the use of mobile CT scanners for lung cancer can address this by bringing screening to people rather than requiring them to travel.

The role of artificial intelligence

A commonality among breast, lung and prostate cancer screening programmes is that, due to their use of imaging technology, they all rely on radiologists. As such, the impact on radiologist capacity and workforce requirements is an important consideration when developing a national programme.

Incorporating AI could offer benefits for streamlining workflow and reducing capacity pressures on radiologists. For example, in a Swedish breast cancer screening trial, AI was used to provide a results read, with the hope that it may help reduce workload for radiologists without compromising accuracy. There are examples of the AI reader being as good as, if not better than, the human reader. Therefore, there is potential to make even more use of AI, and future screening programmes may integrate AI from the start.

There is a broader growth of AI in cancer diagnosis and treatment. In a recent example from the UK, funding has been allocated to 64 National Health Service trusts to deploy AI tools that will analyse X-rays and CT scans. It is predicted that this will speed up diagnosis and treatment.

There is also interest in exploring the potential to use AI to stratify screening participants – for example, to offer different screening recall intervals based on personalised risk, identified via AI analysis. There is interest, too, in the potential use of AI as part of risk modelling, refining the eligibility criteria of screening. This could decrease both the cost of screening and the risk of false positives, something that the Network explored in its recent report on early detection.

The future of screening looks collaborative

In future, we could see up to six different invitations for cancer screening coming to individuals across the EU. While the evidence-based expansion of screening provision could be hugely beneficial for the population, there are risks, particularly around the potential for disjointed messaging or information, and information burden for potential participants. However, there are also opportunities to explore, including communicating about screening jointly – presenting clear, shared messaging – and streamlining the experience of being invited to screening. It is also important for us to continue discussions about the workforce and technical capacity of health systems to offer screening, as well as opportunities to share resources.

We look forward to continuing discussions and learning from other programmes as we support the accelerated implementation of LDCT screening for lung cancer and advocate for earlier detection and improved care.

Authored by: Eleanor Wheeler, Ritse Maarten Mann, Hein Van Poppel and Mireille Broeders