Key Takeaways:
- New AI test can identify 3 major cancers from a drop of dried blood
- Looks for metabolism byproducts/biomarkers in blood using machine learning
- Could allow for earlier cancer diagnosis when more treatable
Can Cancer Be Detected By AI?
A single drop of dried blood could be all that’s needed for a new AI test that can quickly and effectively identify three main types of cancer in one day.
In early trials, the tool was able to tell apart people without cancer from those with confirmed pancreatic, stomach, or colorectal cancer, and the analysis took only a few minutes, according to Live Science.
The researchers say the test can determine if someone has cancer 82% to 100% of the time by identifying specific molecules in their blood.
The new tool analyzes metabolites—byproducts of metabolism—in blood samples using machine learning, which is part of artificial intelligence (AI).
These metabolites, found in the liquid part of blood called serum, act as “biomarkers” that may indicate if there’s cancer in the body.
Screening for these blood biomarkers could be a useful way to diagnose cancer early, when survival chances are higher with no obvious symptoms.
Currently, there are no stand-alone blood tests reliable enough to diagnose pancreatic, colorectal, or stomach cancers on their own, even though these are some of the deadliest tumors in the world.
Instead, doctors typically use imaging or surgery to find cancerous tissue.
In theory, less than a drop of blood would be needed for the proposed test to diagnose these diseases.
The Chinese scientists who created the test detailed the results in a study published Monday in the journal Nature Sustainability.