AI is increasingly being leveraged to address challenges in breast cancer care, from a “second set of eyes” during screening to enabling more personalised treatments. Innovators say it can not only deliver better outcomes, but also make medicine more accurate and accessible.

Breast cancer is the leading cancer in women globally, affecting around two million women, or one in eight, every year.

Early detection is critical for improving survival rates, with over 90 per cent of women surviving for five years or more when the cancer is diagnosed in the early stages.

However, research has shown that between 20 and 40 per cent of cancers may be missed through routine 2D mammography alone, particularly among women with dense breast tissue, which makes tumours harder to detect without specialist diagnostic tools. 

AI is increasingly being leveraged to address these challenges, from enhanced imaging interpretation and risk prediction to personalised diagnostics and treatment planning. 

On World Cancer Day (February 4), the NHS announced a ‘world-leading’ trial to test how AI tools can be used to catch breast cancer cases earlier, with 700,000 women taking part at more than 30 breast clinics across England. 

The EDITH trial (Early Detection using Information Technology in Health) is backed by £11 million of government support. It aims to double radiologist capacity, speed up results, and detect cancers earlier via algorithm‑based comparisons with historical imaging. 

The technology will assist radiologists in identifying changes in breast tissue that show possible signs of cancer, enabling just one specialist to complete the same mammogram screening that would previously have required two to carry out the process safely and efficiently. 

In a previous trial, involving 10,000 women receiving care through the NHS, an AI tool known as Mia identified small signs of breast cancer in 11 women which had been missed by doctors.

Announcing the EDITH pilot in February, Lucy Chappell, chief scientific adviser at the Department of Health and Social Care and CEO of the National Institute for Health Research, said the “landmark trial” could lead to a “significant step forward in the early detection of breast cancer”, offering women faster, more accurate diagnoses.

“A second set of eyes”

AI algorithms are now being widely used as “second readers” to analyse scans and flag abnormalities that might otherwise go unnoticed. Clinical trials in Europe and the US have demonstrated that AI-supported screening can increase detection rates while maintaining or reducing false positives.

A nationwide study in Germany, thought to be the first to use AI screening from the outset in a real-world setting, found that radiologists with access to AI technology had a 17.6 per cent higher rate of detection compared to the control group.

“The addition of AI has improved our ability to detect cancers years earlier, especially in women with dense breast tissue, where traditional mammography has limitations,” Dr Sean Raj, chief innovation officer at SimonMed Imaging, tells Femtech World. 

Mammogram+, developed by SimonMed, integrates 3D mammography with FDA-cleared AI to generate up to 400 high-resolution images per scan, compared to just four in a standard 2D mammogram. 

“The AI system in Mammogram+ acts as a second set of highly trained eyes, assisting radiologists in detecting potential malignancies,” Dr Raj explains. 

“It works to analyse each image quickly and provide key data from a completely objective point of view, allowing our radiologists to build on perspectives they may not have had before.”

While the AI is enhancing its capability, human intelligence remains central to the programme’s efficacy. 

“Our radiologists bring extensive expertise and clinical judgment, while AI provides a complementary, unbiased perspective,” Dr Raj adds.

“By working together, we achieve the best of both worlds, human experience combined with advanced computational analysis.”

While long-term outcome data is still being collected, SimonMed has observed promising early results from the implementation of Mammogram+, including increased confidence for radiologists interpreting images and patients receiving their results, he says.

Patients are also demonstrating higher levels of engagement in their care, which he puts down to the programme’s focus on delivering ‘clear, actionable results’.

Patients receive an easy-to-read report the same day, explaining breast density, personal risk score, and outlining a tailored action plan. 

“We believe that when patients are provided with clear information about their breast health and a personal action plan outlining exactly what to do, they are more likely to take the next step,” he adds. 

More personalised and effective treatments 

Once a patient has been diagnosed with breast cancer, outcomes largely depend on having quick access to the most effective treatment.

But despite decades of advancements in the field, there are still a lot of unknowns when it comes to the question of which treatment a tumour will respond to.

“Part of the reason breast cancer is so difficult to treat is that no two tumours are alike. Even within a single tumour, cells can behave very differently, explained Wolfgang Hackl, translational oncology scientist and founder and CEO of Swiss startup OncoGenomX, in a recent editorial for Femtech World.

Hackl believes the decision lies in precision oncology, and with OncoGenomX, has developed a clinical decision support platform to personalise precision cancer treatment to enable individualised therapeutic decisions. 

According to the company’s research, 56 per cent of women living with metastatic breast cancer are at risk of receiving suboptimal treatment compositions, leading to poorer clinical outcomes and avoidable cancer care costs. 

OncoGenomX’s PredictionStar™ tool uses AI and machine learning algorithms to identify whether a certain therapy works in a particular tumour and finds the most effective treatment for each patient and their tumour. It has been found to reduce the over-treatment rate by at least 40 per cent, incurring a five-fold lower risk of cancer regrowth.

“PredictionStar takes away the uncertainty around breast cancer treatment decisions and increases the likelihood of opting for the most effective treatment combination without delay and without exposing the patient to the side effects of an ineffective treatment,” says Hackl. 

AI and its role in the gender health gap

The potential of AI in healthcare extends far beyond breast cancer, with the potential to address other areas of medicine where women are underdiagnosed, undertreated and have poorer outcomes, such as cardiovascular care.

SimonMed has just launched a new AI-powered software that detects breast arterial calcifications (BAC), offering “enhanced care and a two-for-one mammogram”. 

“Using the same screenings from the mammogram, we can detect the presence of BAC, providing dual insights into two of the biggest health risks for women, breast cancer and heart health, all without extra radiation or procedure time for the patient,” Dr Raj explains.

“For years, women have faced a gender health gap, affecting women’s access and affordability, misdiagnosis, and medical research bias.

“We’re excited by the possibility that AI can help us close that gap and improve health outcomes for women by having more accurate, faster tools for the detection and treatment and providing patients and their providers with more data and insights to make more informed decisions, sooner.”

He adds: “With the help of AI in women’s healthcare, like breast cancer detection, we will be able to improve health outcomes for women, making healthcare more accessible, affordable, and accurate.”

Urging caution – governance and collaboration 

However, several experts and peer-reviewed papers have urged caution around the widespread use of AI, with risks including overdiagnosis, false positives, bias, and workflow disruption, highlighting the importance of regulatory oversight and ethical safeguards. 

A review published earlier this year raised several concerns, including inconsistent performance, poor generalisability, lack of robust evidence frameworks, trust issues among clinicians, and gaps in legal and ethical governance. 

The researchers concluded that more robust strategies are needed before AI can be scaled for routine use in breast cancer screening, making recommendations for a “comprehensive AI governance framework”.

“While AI has the potential to improve diagnostic accuracy and efficiency, its broader implications include promoting equitable health care delivery, strengthening patient trust, and supporting the ethical development of AI technologies,” the authors write.

“Policy makers, clinicians, and AI developers must work collaboratively to establish adaptable and transparent systems that prioritise patient safety and societal benefits.

“Future research should focus on real-world case studies, longitudinal assessments, and cross-disciplinary collaborations to effectively refine and implement these governance strategies.”