If breast cancer is detected at stage I, there is a 99% chance the patient will survive for the next five years at least. However, if the same cancer is diagnosed at stage two, the five-year survival rate drops to 86%.

Currently, most breast cancer detection methods are able to spot the disease at stage II, but this isn’t enough. A new blood test could change all that. 

Breast cancer, the most common form of cancer among women in most countries, claimed an estimated 670,000 lives globally in 2022. The best way to reduce this burden is to detect it earlier.

Now, researchers from the University of Edinburgh have developed a revolutionary blood test capable of detecting breast cancer at stage 1a, a very early stage that is difficult to identify with existing techniques.

“What makes this test special is that it is the first of its kind to utilize Raman Spectroscopy and artificial intelligence (machine learning) to classify the four major breast cancer subtypes at stage 1a,” the researchers note.

A blood test powered by machine learning

Currently, breast cancer detection relies on biopsies, ultrasound, and X-rays—methods that can be invasive or less effective in patients with dense breast tissues. This new test aims to address these limitations with a quicker, less invasive approach.

The researchers claim that their blood test is one of the fastest and highly accurate methods to spot breast cancer. A doctor just needs a blood plasma sample from the patient. 

The sample is first analyzed using laser light. When the light interacts with plasma molecules, a device called a spectrometer examines the changes in the light’s properties to determine the chemical composition of the blood sample. This technique is called Raman spectroscopy.

Next, the researchers employ a machine learning algorithm to understand the results of Raman spectroscopy (RS). The algorithms reveal whether blood cells and tissues have undergone any unwanted molecular changes.

To check the accuracy of their method, the researchers tested the blood samples of 12 cancer patients and 12 healthy individuals. The results showed that the AI and RS-powered blood test identified stage 1a breast cancer with 98 percent effectiveness in the subjects.

Moreover, there are four subtypes of stage 1a blood cancer; Luminal A (HR+HER2−), Luminal B (HR+HER2+), HER2-enriched (HR−HER2+), and TNBC (HR−HER2−). During the experiment, the blood test could identify a patient’s subtype with over 90% accuracy.

“We identified key spectral features linked to cancer biomarkers, which were corroborated by multiple peer-reviewed studies, affirming the potential of RS and machine learning in early-stage breast cancer subtyping,” the researchers said.

Not just limited to breast cancer

The lack of early-stage cancer detection techniques isn’t just killing breast cancer patients. This is a widespread problem among patients with different types of cancer, affecting thousands of lives across the globe. 

Even if a patient is unable to beat the disease, early-stage detection of cancer can play a crucial role in increasing their lifespan. The researchers suggest that in the future, their AI-driven blood test can help in early-stage detection of various other types of cancer. 

“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,” Andy Downes, one of the researchers and a senior lecturer at the University of Edinburgh, said.

The study is published in the Journal of Biophotonics.