News
AI model could predict breast cancer risk without racial bias, study finds
The deep learning model has been shown to outperform traditional risk models in predicting a woman’s risk of developing breast cancer
A new AI model developed using mammogram image biomarkers could accurately predict both invasive and non-invasive forms of breast cancer without racial bias, new research has shown.
The findings, presented at the annual meeting of the Radiological Society of North America (RSNA), have suggested the model showed no bias across multiple races.
Traditional breast cancer risk assessment models use information obtained from patient questionnaires, such as medical and reproductive history, to calculate a patient’s future risk of developing breast cancer.
Data shows old models are more likely to demonstrate poor performance across different patient races, most likely due to the data used to develop the model.
“In the domain of precision medicine, risk-based screening has been elusive because we have not been able to accurately evaluate a woman’s risk of developing breast cancer,” said study lead author Leslie R. Lamb, a breast radiologist at Massachusetts General Hospital (MGH) in Boston.
“Even the best existing traditional risk models do not perform well on the individual level.
“Traditional models likely have racial biases due to the populations on which they were developed. Several of the commonly used models were developed on predominantly European Caucasian populations.”
According to the American Cancer Society, Black women demonstrate the lowest five-year relative survival rate for breast cancer among all racial and ethnic groups. This translates to a six per cent to eight per cent disparity in five-year survival rates between Black and white women across all breast cancer types.
The new deep learning AI risk assessment model developed using mammographic images has been shown to outperform traditional risk assessment models in future breast cancer development while also mitigating the racial biases seen in traditional models.
In the first study of its kind, Dr Lamb and colleagues assessed the performance of the image-based deep learning risk assessment model in predicting both future invasive breast cancer and non-invasive breast cancer across multiple races.
The study included 129,340 routine screening mammograms performed in 71,479 women between 2009 to 2018 with five-year follow-up data. Patient demographics were obtained from electronic medical records and instances of cancer were identified from the regional tumour registry.
The racial makeup of the study group included white, Black, Asian, self-reported and other races. The mean age of the women was 59 years old.
The deep learning model was shown to outperform traditional risk models in predicting a woman’s risk of developing early-stage, non-invasive breast cancer, also known as ductal carcinoma in situ (DCIS), as well as invasive breast cancer, known as invasive carcinoma.
“The model is able to translate the full diversity of subtle imaging biomarkers in the mammogram, beyond what the naked eye can see, that can predict a woman’s future risk of both DCIS and invasive breast cancer,” explained Lamb.
“The deep learning image-only risk model can provide increased access to more accurate, equitable and less costly risk assessment.”
Constance D. Lehman, senior author and breast radiologist at MGH, said: “This is a particularly exciting domain for AI, as it demonstrates the opportunity to apply ‘AI for good’—to reduce well-known racial disparities in risk assessment.
“We are now poised to translate these findings into improved clinical care for our patients.”
Supermarket loyalty card data suggests more than a quarter of women buying menstrual products also buy pain relief at the same time.
The findings also suggest women in lower-income areas are significantly less likely to do so, pointing to disparities in access to over-the-counter pain relief across England.
The study was led by Dr Victoria Sivill of the University of Bristol and colleagues.
The authors said: “The study highlights the need for greater awareness and policy interventions to address the high prevalence of menstrual pain as well as socioeconomic dimensions of menstrual pain.
“Public health initiatives should incorporate menstrual pain relief as part of broader efforts to improve health equity.”
Researchers analysed anonymised loyalty card data from a major UK health and beauty retailer, covering 211m transactions by 3.4m people between 2006 and 2015.
The study examined how often shoppers bought menstrual products and pain relief in the same transaction, and compared this with a customer’s usual rate of buying pain relief.
It found that 26.7 per cent of customers who bought menstrual products also bought pain relief at the same time.
These customers were nearly four times more likely to buy pain relief when buying menstrual products than on other shopping trips.
As a check on the approach, researchers found the most common interval between consecutive menstrual product purchases was exactly 28 days, consistent with the average menstrual cycle.
Menstrual pain is common and can affect daily activities, including school and work attendance.
Regional income was the strongest predictor of menstrual pain purchases.
Customers in the lowest-income areas were 32 per cent less likely to buy pain relief at the same time as menstrual products than those in the highest-income areas.
The authors said lower rates of pain relief purchases in deprived areas are likely to reflect an inability to afford over-the-counter medication rather than lower rates of menstrual pain itself.
Co-author Dr James Goulding said: “It is wonderful that smart data research in the UK is able to bring issues which may have once been overlooked in scientific settings, such as the sheer scale and impact of menstrual pain, to light. This is well overdue.
Co-author Dr Anya Skatova said: “Like many women, I was aware of how common menstrual pain is, but the scale of painkiller purchases alongside menstrual products was still striking.
“Using shopping data, we can see just how widespread the need for pain relief really is. This kind of evidence helps make menstrual pain visible at a population level and provides a strong foundation for systemic change in how it is recognised, treated, and prioritised in public health.”
Women are still being failed by menopause care despite a surge in online advice, with experts saying symptoms are too often minimised or dismissed.
The researchers exposed a gap between the surge of menopause information available online and the quality of medical care women receive.
A team of anthropologists and psychologists explored the physical and emotional toll of menopause, including its impact on work and personal lives, through interviews with 60 women aged 45 to 61 between March and June 2021.
The findings were published in a new book, We Need to Talk About Menopause.
The interviews showed how women continue to “needlessly suffer” as they sort through misinformation from influencers, celebrities and so-called experts.
The authors said: “You would think in an age where humans are developing commercial space flight, self-driving cars, and AI personal assistants who can project movies onto the palm of your hand, we would understand more about menopause, something a little over half of the population is guaranteed to experience in their lives.”
The interviews revealed wide variation in women’s experiences of doctors.
Some felt they could have an open dialogue, while others said they were “shut down”, including being told they were too young for menopause.
One woman said she bled heavily for a year before she was taken seriously.
The authors said menopause is still poorly understood, with disagreement over whether it should be seen as a medical condition or a natural part of ageing.
There are more than 100 recognised symptoms, although some women experience none.
Among those interviewed, 78 per cent reported weight gain and redistribution, particularly around the belly area, which was resistant to diet and exercise.
Fifty-eight per cent experienced mood disturbances including anxiety, depression, irritability and unprecedented levels of rage.
Many women said they were blindsided by symptoms they had never known existed.
One participant said she only realised rage was a menopause symptom after seeing it mentioned in a television commercial.
Women also described severe memory problems and brain fog that colleagues mistook for incompetence, leading successful professionals to question their abilities at the peak of their careers.
Brain fog can include problems with concentration, memory and clear thinking.
According to Statistics Canada, 70 per cent of women turn to the internet as their primary source of menopause information.
The authors said this information vacuum has spawned a £14.7bn global “meno-tech” industry, with influencers, celebrities and telehealth companies offering products ranging from £98 creams to unnecessary blood tests.
They said: “The growth of interest in menopause has also been accompanied by a wave of unsubstantiated information.
“Many websites market expensive creams, supplements, and weight-loss schemes that promise to keep women youthful and attractive, with little evidence to support their claims.
“Reliable, accessible information about menopause and perimenopause is still lacking. Despite increased attention to the importance of physician training and the search for menopause specialists, the medical profession as a whole continues to provide limited support in this area.”
The authors also highlighted the effect of menopause in the workplace.
The House of Commons Women and Equalities Committee has warned that Britain is “haemorrhaging talent” because of menopause, with 14 million workdays lost each year, according to the Office for National Statistics.
Despite this, studies have shown 80 per cent of UK employers have yet to implement proper support measures for menopausal women.
Some women said simple workplace adjustments made a significant difference.
The Women and Equalities Committee (WEC) has warned against “financial half measures” on women’s health as the government published its response to the report.
Ministers launched the renewed Women’s Health Strategy in April after the committee’s March report concluded it was not convinced that the menstrual and gynaecological needs of young women and girls had been sufficiently prioritised in wider healthcare reforms.
It followed the committee’s 2024 “medical misogyny” report, which found women with painful reproductive health conditions such as endometriosis, adenomyosis and heavy menstrual bleeding were frequently finding their symptoms “normalised” and their “pain dismissed” when seeking help.
In both reports, MPs called on the government to recognise the benefits of increased investment in early diagnosis and treatment of women’s reproductive health conditions and provide additional funding needed to transform the support available to millions of women.
In its response, published on 26 May as a command paper, the Department of Health and Social Care outlined action on reducing gynae waiting times, ensuring procedures are conducted with women’s full consent and adequate pain relief, and improving access to contraception for menstrual healthcare in line with the committee’s recommendations.
It said: “The government agrees with the committee’s overarching findings and recommendations for improving women’s health outcomes and experiences.
“We acknowledge the impact that menstrual health conditions can have on women’s lives, relationships, and participation in education and the workforce.
“We recognise that more needs to be done to support women with menstrual health conditions, particularly around listening to women, improving information and education, and enhancing patient experience.”
However, there was no commitment to increase school nurse provision, no measurable actions and targets on countering online misinformation, no new commitments to end inappropriate censorship of women’s online health content, and no further initiatives on tackling racial discrimination or understanding the menstrual wellbeing needs of young disabled and Deaf women.
The response comes after analysis by The Times suggested the government is allocating 60 per cent more funding to its men’s health strategy than to its renewed strategy for women’s health.
Sarah Owen, chair of the Women and Equalities Committee and Labour MP, said: “WEC’s 2024 ‘medical misogyny’ report warned 18 months ago of women in unnecessary pain and undiagnosed for years and called on the Government to recognise the benefits of increased investment in early diagnosis and treatment.
“Our follow up report this March cautioned girls’ and women’s health are not being sufficiently prioritised in system-wide NHS reforms, while initiatives which have proven to be successful in reducing waiting lists and improving women’s healthcare access, such as women’s health hubs, risked being scaled back or discontinued.
“While it’s welcome to see a focus on tackling ‘medical misogyny’ in April’s renewed Women’s Health Strategy and an emphasis on women’s voices being heard, this must be backed by adequate funding, not financial half measures, particularly when compared to men’s health.
“Significant questions remain following today’s response publication over the adequacy of investment being provided, including for workforce training, menstrual health education in schools, research and additional ring-fenced funding for women’s health hubs to deliver services within the emerging neighbourhood health framework.
“There are both opportunities and risks when it comes to increasing use of technology in women’s healthcare.
“As the Committee’s report set out, social media companies should be held to account for inappropriate and disgraceful ‘shadow banning’ censorship of important women’s health content and there should be a rigorous approach to tackling the risks from ineffective, unsafe and exploitative for-profit FemTech apps.
“The Government should take the problem of ‘shadow banning’ more seriously.
“A strategy which does not fully address the concerns set out in WEC’s report, alongside measurable actions and timescales, will only scratch the surface of the issues facing women’s health.
“WEC will keep a close eye on progress and continue to push for long overdue tangible change for women and girls.”
Opinion4 weeks agoWhat Maternal Mental Health Month reveals about where postpartum support actually breaks down
News3 weeks agoNIH Grant terminations disproportionately impact minority scientists, research finds
Menopause4 days agoPerimenopause misinformation ‘putting women at risk’
Wellness3 weeks agoWUKA brings Period-Positive Pool Party to London Aquatics Centre to keep girls swimming through puberty
Insight2 weeks agoPCOS renamed after decade-long campaign to end ‘cyst’ misconception
Events4 weeks agoWHIS 2026 unveils agenda and first speakers for the leading women’s health summit
Menopause3 weeks agoCBT shows promise for menopause insomnia and hot flashes
Insight4 weeks agoOnline abuse and deepfakes ‘pushing women out of public life’