AI enables identification and localization of 12 cancers with a simple blood test 🩸

A blood test capable of early detection of a dozen cancers with unmatched accuracy is entering clinical trials in the UK. Its secret? A unique combination of genetic analysis and artificial intelligence.


Developed by the University of Southampton and the startup Xgenera, the miONCO-Dx test could revolutionize screening by identifying tumors long before symptoms appear. With just 10 drops of blood, it targets cancers often diagnosed too late, such as pancreatic or ovarian cancer.

How does this technology work?

The test analyzes microRNAs, genetic fragments released by cancer cells into the blood. The AI compares these markers to a database to identify their origin with 99% accuracy, according to preliminary trials on 20,000 patients.

Unlike traditional methods (biopsies, scans), this approach is non-invasive and saves medical resources. It could also reduce diagnosis times, which are crucial for aggressive cancers.

In addition to detecting the presence of a tumor, the algorithm pinpoints its location in the body. This specificity eliminates the need for additional tests and speeds up treatment.

A potential impact on patient survival

Cancers detected at an early stage have a much higher survival rate. For example, 90% of colorectal cancer cases are curable if identified early, compared to just 10% at advanced stages.

The NHS (National Health Service, UK healthcare system) is currently testing the device on 8,000 patients, with public funding of £2.4 million. If results are confirmed, the test could be rolled out as early as 2025.

This innovation is part of a global effort to improve screening. In the UK, the Bowelbabe lab, created in tribute to activist Deborah James, also supports these advancements.