Breast cancer is more treatable the earlier it’s found. Yet current screening methods often pick up the disease only after a tumor has already formed and grown large enough to be seen on a scan or felt as a lump. What if doctors could identify the illness weeks or months before that point, when it’s just beginning to alter the body’s chemistry at a microscopic level?
Researchers at the University of Edinburgh believe they’ve taken a step in that direction. They’ve been experimenting with a new test that examines blood plasma using a specialized laser technique, then applies AI to interpret the results. This pilot study focused on the earliest stage of breast cancer — stage 1a — which is so subtle that standard tests usually can’t detect it. The results were promising. According to the team, their method identified stage 1a breast cancer in the test samples with a very high degree of accuracy. In fact, the data from a small group of patients showed that the approach was about 98% effective at distinguishing the blood samples of people with early cancer from those of healthy individuals.
How does this test work? The key lies in a form of light analysis called Raman spectroscopy. By shining a laser into the blood sample and reading how the light scatters, scientists can see a molecular “fingerprint.” Cancer causes subtle shifts in proteins, lipids, and other molecules that circulate in the blood. At the earliest stage, these changes are faint and easy to miss. But when combined with AI-driven analysis, even these faint signatures can be teased out. The result: a non-invasive, relatively quick test that may catch cancer before conventional methods have any clue it’s there.
Not only did the test detect cancer at a stage where it often goes unnoticed, it also identified which subtype of breast cancer was present. The four major subtypes — Luminal A, Luminal B, HER2-enriched, and Triple Negative — respond differently to treatments. Distinguishing them early could give doctors and patients a head start in planning the most suitable treatment approach. In this pilot, the test classified the subtype with more than 90% accuracy. That’s significant. Instead of a one-size-fits-all approach, doctors might be able to customize therapy from the beginning, potentially improving outcomes and reducing unnecessary interventions.
Of course, a pilot study with a small number of samples is just a start. The next step is to validate these results in larger trials. Is this approach consistent across hundreds or thousands of samples? Can it be scaled up and incorporated into routine screenings? And what about other types of cancer — could the same technique catch them early too?
The researchers are hopeful. Dr. Andy Downes, of the University of Edinburgh’s School of Engineering, who led the study, explains why this could matter: “Most deaths from cancer occur following a late-stage diagnosis after symptoms become apparent, so a future screening test for multiple cancer types could find these at a stage where they can be far more easily treated. Early diagnosis is key to long-term survival, and we finally have the technology required. We just need to apply it to other cancer types and build up a database, before this can be used as a multi-cancer test.”
If these plans pan out, the impact could be far-reaching. Imagine going in for a routine blood test that flags possible issues long before you’d ever know something was wrong. Such a shift could ease the anxiety many patients face when waiting for imaging results, reduce the invasiveness of certain diagnostic procedures, and let doctors intervene when treatments are most likely to succeed. While there’s work ahead, this research offers a glimpse of what early detection could look like: less guesswork, more clarity, and potentially a better future for patients caught in cancer’s early net.
Citation: Tipatet, K.S., Hanna, K., Davison-Gates, L., Kerst, M. and Downes, A. (2024), Subtype-Specific Detection in Stage Ia Breast Cancer: Integrating Raman Spectroscopy, Machine Learning, and Liquid Biopsy for Personalised Diagnostics. J. Biophotonics e202400427. https://doi.org/10.1002/jbio.202400427