AI-powered blood test first to spot earliest sign of breast cancer

The team says this rapid, non-invasive technique reveals subtle changes in blood flow that occur during the initial stages of the disease, known as Stage 1a, that cannot be detected by existing tests.

Researchers at the University of Edinburgh say their new method could improve early detection and monitoring of the disease and pave the way for screening tests for a variety of cancers.

Standard tests for breast cancer include a physical exam, an X-ray or ultrasound scan, or an analysis of a sample of breast tissue (called a biopsy). Existing early detection strategies rely on screening based on age or belonging to a high-risk group.

Using this new method, researchers are able to detect breast cancer at its earliest stages by optimizing a laser analysis technique called Raman spectroscopy and combining it with machine learning, a form of artificial intelligence.

Similar methods have been tried to screen for other types of cancer, but the earliest they were able to detect the disease was in the second stage, the team said.

The new technology works by first shining a laser beam into blood plasma taken from the patient. A device called a spectrometer is then used to analyze the properties of the light as it interacts with the blood to reveal tiny changes in the chemical makeup of cells and tissues, which are early indicators of disease.

Machine learning algorithms are then used to interpret the results, identify similar features and help classify the samples.

In a pilot study involving 12 breast cancer patients and 12 healthy control samples, the technology was 98% effective in identifying stage 1a breast cancer.

The test can also differentiate between the four main subtypes of breast cancer with an accuracy of more than 90%, which could allow patients to receive more effective, personalized treatment, the research team said.

The team says using it as a screening test would help identify more people with early-stage breast cancer and improve the chances of successful treatment. Their goal is to expand the work to include more participants and include testing for early forms of other cancer types.

The research was published in Journal of Biophotonics. Blood samples used in the study were provided by the Northern Ireland Biobank and the Breast Cancer Tissue Bank. Researchers from the University of Aberdeen, the Rheinvar University of Applied Sciences and the North Rhine-Westphalia Graduate School of Applied Research are also involved.

Dr Andy Downs from the University of Edinburgh’s School of Engineering, who led the study, said: “Most cancer deaths occur after late diagnosis, after symptoms have become apparent, so future screening tests for multiple cancer types could be used at an earlier stage Diagnosis is the key to long-term survival, we just have to apply it to other cancer types and build a library and then we can use it.