Overview

Cancer screening can reduce cancer morbidity and mortality by identifying early abnormalities, before they cause signs and symptoms.

Organized, population-based screening delivers higher quality outcomes than opportunistic testing, because it is systematic, quality-assured and offers more equitable access and coordinated pathways to cancer detection and care.

Cancer screening programmes optimize their full potential when they achieve high uptake in the population, particularly among high-risk groups, and are able to deliver fast, accurate cancer detection services.

Given the above, the European Union is encouraging Member States to consider three innovations with potential to transform cancer screening pathways:

a) self-administered testing
b) risk-based stratification
c) artificial intelligence (AI)-aided reading of radiological imaging.

Current knowledge about these screening innovations stems largely from small scale analyses and more robust studies are needed. Nonetheless, implementation experiences so far have generated a number of key insights.

Self-administered testing improves screening access by shifting screening from clinics to people’s homes – thus reducing time, travel and stigma-related barriers. Evidence shows that:

• Self-administered testing is effective for colorectal, cervical and gastric cancer screening;
• Direct mailing of test kits is effective in increasing uptake, particularly when combined with health professional-led outreach and education;
• Vulnerable populations require tailored community engagement to ensure equitable benefit.
  

Risk-based stratification improves screening efficiency by using individual risk profiles (including, for example, family or smoking history) to tailor screening eligibility, screening intervals and follow-up pathways.

• Risk stratification channels screening resources to those most likely to benefit and reduces screening-related harms in lower-risk groups; it is also expected to improve overall value for money.
• Risk prediction models (or risk calculators) can only achieve the above impacts if they include evidence-based risk factors and if they are validated in the populations and settings they are implemented in, to ensure accuracy and minimize bias.
• Given its involvement in preventive services, primary care is likely best placed to manage risk assessment and risk calculation, but needs to be supported by well-developed data infrastructures (such as electronic health records) and training for health professionals (particularly in risk interpretation and communication).
  

AI-assisted image reading shows promise in improving the accuracy and timeliness of cancer screening, compared to traditional computer-assisted detection (CAD) for lung and breast cancer.

• When used as a second reader of radiological imaging, AI can streamline screening workflows, reduce diagnostic delays and alleviate pressures on the health workforce.
• However, there are implementation challenges to the use of AI.
  • The availability of representative real-world datasets for validation and bias mitigation of AI tools and algorithms is limited.
  • Resources and tools to support clinician–patient communication about AI-aided clinical decisions are lacking.
  • Clinician training and co-creation are needed to ensure that AI tools are safe, ethical and acceptable.

Integrating cancer screening innovations into population-based programmes is important and more likely to succeed if policy-makers tackle cross-cutting concerns. Areas worth focusing on include:

• Encouraging screening uptake through:
  • Clear public health education and communication to build trust and improve participation;
  • Tailored approaches that fit the participants’ circumstances and enable equitable access;
  • Compliance with ethical and governance frameworks to protect informed consent and privacy.
• Making implementation of innovations and change feasible in practice by:
  • Training health professionals to help innovations become part of routine practice;
  • Integrating innovations into existing screening and diagnostic workflows;
  • Maintaining human support and involvement to complement technological solutions;
  • Building robust data infrastructures to support validation, testing and implementation of technological innovations.
• Improving the evidence base on the effectiveness, implementation and cost-effectiveness of cancer screening innovations by:
  • Drawing on cancer screening programme monitoring and evaluation data;
  • Developing and supporting large, prospective, multi‑country evaluations.