Artificial Intelligence Detects Microscopic Breast Cancer Tumor Overlooked by Medical Professionals
A groundbreaking clinical study has demonstrated how artificial intelligence technology successfully identified an extremely small breast cancer tumor that had escaped detection by healthcare professionals, potentially saving a patient from more intensive treatment later.
Yvonne Cook, a patient from Aberdeen, became part of a revolutionary trial that showed AI could enhance breast cancer detection rates by more than 10 percent. The sophisticated technology flagged an anomaly in her screening that led to the discovery of an aggressive tumor too minute for human observers to detect.
The patient expressed gratitude for the timing of the AI intervention, noting that without this technological assistance, her cancer likely would have remained undetected until her next scheduled mammogram three years later, or until the tumor had grown large enough to be physically felt.
Enhanced Detection Capabilities
The research, published in Nature Cancer, involved collaboration between the University of Aberdeen, NHS Grampian, and Kheiron Medical Technologies. The study examined 10,889 women undergoing routine breast screening, with the AI system called Mia providing support to healthcare workers.
Results showed that while 106 cancers were identified through standard screening procedures, an additional 11 cases were detected with AI assistance, seven of which were classified as invasive cancers. The technology also demonstrated the ability to reduce notification times for patients from two weeks to just three days.
Current breast cancer screening protocols in the UK invite women aged 50 to 70 for mammograms every three years, with scans typically reviewed by two radiologists. However, some cancers remain difficult to detect, leading to missed diagnoses and unnecessary recalls for additional testing.
Operational Benefits
The study revealed that incorporating AI as either a substitute for one human reader or as an additional safeguard reader could reduce healthcare staff workload by up to 31 percent while maintaining or improving detection accuracy. This approach also decreased the number of women unnecessarily recalled for further testing.
Dr. Clarisse de Vries, the study’s lead author and lecturer in data science at the University of Glasgow, emphasized that the research identified optimal methods for detecting breast cancer more quickly and accurately while reducing unnecessary patient anxiety from false alarms.
Professor Gerald Lip from the University of Aberdeen highlighted the critical nature of the findings, stating that without AI intervention, these particular cancers would not have been caught at such an early stage.
Patient Experience
Cook’s experience began with a routine mammogram in May 2023, followed by a callback for additional imaging after the AI system flagged a potential concern. Subsequent examination confirmed the presence of a small Grade 2 tumor that was invisible to human detection methods.
The patient described feeling extraordinarily fortunate to participate in the research program and to have her cancer identified at such an early stage, potentially avoiding more aggressive treatment protocols that would have been necessary with later detection.
Broader Research Implications
Additional research conducted by experts from Imperial College London, Google, and several universities examined data from over 175,000 women across multiple study phases. The comprehensive analysis confirmed AI’s ability to increase cancer detection rates, identify more invasive cancers, reduce false positive results, and significantly decrease scan reading times.
One portion of the study found that AI could complete scan readings in an average of 17.7 minutes, compared to two days required for initial human assessment. When used in arbitration cases where medical professionals disagreed on diagnoses, AI performed comparably to human experts while reducing overall screening workloads.
Dr. Hutan Ashrafian from Imperial College London’s Institute of Global Health Innovation noted that this research represents the closest AI has come to potentially reducing breast cancer mortality within the NHS, suggesting significant opportunities for healthcare system implementation.
Lord Darzi, who authored an influential NHS report in 2024, emphasized AI’s transformative potential for disease prevention, detection, and treatment, particularly highlighting how the technology can support clinicians in identifying cancers earlier while reducing errors and improving patient care quality.