Tim Simpson, Managing Director, Hologic UK & Ireland explores the potential of AI to improve access to targeted and personalized diagnosis and explores what is needed to truly personalize healthcare with AI.
Technological innovation is set to shape healthcare in the future and as we seek to bounce back from COVID-19, we must explore how we can harness the power of artificial intelligence (AI) to transform and personalize women’s health care.
The government aims to put the UK at the forefront of the AI and data revolution in early diagnosis, innovation, prevention and treatment[i]. We need to leverage this focus on AI to understand how it can facilitate the identification of women at high risk for diseases such as cancer, to ensure rapid diagnosis and access to treatment.
Women’s health issues are not new, but the COVID-19 pandemic has accelerated the urgent need to improve women’s health outcomes, address health inequities, and create an improved patient journey through personalized screening.
One size does not fit all
AI has the potential to improve access to targeted and personalized diagnosis.
While AI has been adopted in some cancer screening programs, what we now need to see is a move towards a prioritized, risk-assessed screening system rather than a one-size-fits-all approach.
Personalized screening will require the use of AI for risk stratification to identify high-risk patients. We need to explore how to use AI to create a molecular profile to determine relative risk. For example, we know that women with dense breasts are at higher risk of developing breast cancer and that 40% of European women aged 40-74 have dense breasts.[ii] We need to make sure these women are identified early so they can be prioritized for screening.
The PROCAS 1 and 2 study in Manchester looked at the impact of creating a risk score for breast cancer in women. The studies revealed that there is a demand from patients to understand risk, with 94% of those recruited wanting to know their risk score. The studies also found a positive benefit of risk stratification, as women were then more likely to act on risk information.[iii]
It is encouraging to see investments in AI for personalized diagnosis, as recently shown by a research study at the University of Strathclyde, which is developing new AI technology to calculate the risk of pre-eclampsia in the women. The tool will examine large datasets of women who have participated in previous research projects taking into account a range of factors, including ethnicity, socio-economic status and details of the current pregnancy of wife.[iv]
Diverse datasets are key to truly personalizing healthcare with AI
For AI to truly transform healthcare, datasets must be inclusive, consider all risk factors, and not focus on a specific demographic.
Researchers at Loughborough University recently announced a new study using AI to reduce the risks faced by pregnant black women. Working with the Health Care Safety Investigations Branch, the team will review hundreds of investigations into adverse outcomes during pregnancy and childbirth. They will use machine learning to identify risk factors and, therefore, design ways to improve care for pregnant women and babies.[v]
Access to diverse datasets provides insight into how the disease progresses in different populations and is essential for ensuring accurate and unbiased patient profiling. The more data points you have, the larger the database and the more accurate the AI.
While significant progress has been made in adopting AI to improve healthcare, as we build on innovations, we need to ensure collaboration between industry, clinicians and researchers to unleash all the power of AI to radically improve women’s health and save lives.
[i] GOV.UK [Internet] The Grand Challenge missions [cited 2022 April 22] Available at: https://www.gov.uk/government/publications/industrial-strategy-the-grand-challenges/missions
[ii] Berg, WA & Vourtsis A, authors. Using education to overcome inequity in access to additional screening for women with dense breasts [Internet] DI Europe: 2020. Available at: dense-breast-screening-vourtsis-berg-febmch2020-1.pdf (densebreast-info.org)
[iii] Manchester Cancer Research Center [Internet] PROCAS and BC-PREDICT Predict cancer risk during screening. [cited 2022 July 22] Available at: https://www.mcrc.manchester.ac.uk/impact-case-studies/procas-and-bc-predict/
[iv] University of Strathclyde in Glasgow [Internet] Funding for AI technology used to calculate pre-eclampsia risk [cited April 26 2022] Available at: https://www.strath.ac.uk/whystrathclyde/news/2021/fundingforaitechnologyusedtocalculatepre-eclampsiarisk/
[v] Loughborough University [Internet] New research from Loughborough will use artificial intelligence to help reduce maternal harm in mothers from black ethnic groups [cited July 19 2022] Available at: https://www.lboro.ac.uk/departments/compsci/news/2021/new-research-help-reduce-maternal-harm/