A Personal Touch to Cancer Screening

There is huge potential for artificial intelligence (AI) to significantly improve women’s health care – that much is obvious. Over the past few years, we have seen exciting developments in the field, and AI is at the forefront of a digital health revolution. As we try to bounce back stronger from COVID-19 and keep prevention at the heart of healthcare, these developments are more important than ever.

In 2021, the Industrial Strategy AI Mission set a goal to use data, AI, and innovation to transform the prevention, early diagnosis, and treatment of chronic diseases by 2030. It predicts that, within 15 years, better use of AI and data could result in over 50,000 more people receiving an early, rather than late-stage, cancer diagnosis each year (1). Looking ahead to the future of cancer screening, we must harness the power of AI to enable timely results that inspire confidence in our patients.

Personalizing cancer screening
 

The key to this mission is a shift toward personalized screening, rather than a one-size-fits-all approach. This will require using AI for risk stratification by creating a molecular profile to determine relative risk for each patient and identify those at highest risk. There are certain risk factors in both breast and cervical screening that are well understood; dense breasts, family history, and an elevated body mass index can indicate higher risk of breast cancer (2), whereas persistent human papillomavirus infection – particularly with types 16 and 18 – is associated with higher risk of cervical cancer (3).

AI and machine learning can help to identify the most diagnostically relevant information and determine which cases to prioritize for further analysis. This information can also determine screening intervals and test types according to each person’s risk.

Easing the burden
 

On receiving screening results, it is vital for healthcare professionals to be able to rapidly review them for speedy diagnosis and treatment initiation, if needed. How can AI ease this burden? For cervical cancer, implementing AI and advanced imaging has the potential to improve the screening process. Creating images that can be rapidly reviewed using an advanced algorithm assesses the cervical cells in the sample and provides the screener with an image gallery of the most diagnostically relevant cells. This helps medical experts more rapidly identify and accurately diagnose abnormalities because they have fewer cells to analyze.

Challenges to overcome
 

Although AI offers great potential in cancer screening, it still faces some challenges before it can be truly transformative. A report by Imperial College London highlighted evidence of racial bias in AI and the risk of exacerbating existing health inequalities (4) – highlighting the fact that, when it comes to personalized screening, datasets need to be inclusive and consider all risk factors in all demographics. The need for access to diverse datasets is clear; understanding how diseases progress in different populations allows us to ensure accurate and unbiased profiling of patients. The more data points available, the bigger the database and the more accurate the results returned by the AI.

We also need national portals to enable healthcare professionals across the UK to access these data, rather than being limited to smaller regions. To ensure widespread AI adoption, our healthcare systems require adequate infrastructure to handle it. For example, there needs to be a network between different pathology departments to better support information sharing and workload. This could allow smaller pathology departments without AI technology to access AI analysis from larger departments or share screening results for a second opinion.

What the future holds
 

Addressing the challenges outlined here will require close collaboration between regulatory bodies, health systems, innovators, and developers. Only by working together can we unlock the full power of AI, which has the potential to significantly improve cancer screening and aid earlier detection for so many patients. I am excited to see what the future holds and the real-world difference that AI will make to patients, healthcare organizations, and our society.