Researchers at the University of Oxford have devised a new biomarker using artificial intelligence (AI) that can reportedly identify people at significant risk of experiencing a fatal heart attack roughly five years before it happens.

The biomarker — or fingerprint — is called the fat radiomic profile (FRP) and it recognizes biological signals in the perivascular space that lines the blood vessels that supply blood to the heart. Those signals include blood vessel scarring, inflammation and other alterations — all of which are suggestive of a future heart attack.

Typically, people who experience chest pain undergo a coronary CT angiogram (CCTA), which is an imaging test to assess the health of the arteries. In the event that no significant narrowing of the arteries is detected, the patient is sent home. However, just because a scan does not show narrowing, according to Professor Charalambos Antoniades, professor of cardiovascular medicine at the University of Oxford, does not necessarily mean that the patient is safe from suffering a future heart attack. As such, the researchers believe that the CCTA alone may not be enough, thus making the FRP more valuable.

To demonstrate the FRP's effectiveness, the team examined the fat biopsies of roughly 165 heart surgery patients, studying the expression of genes with changes, in particular. The researchers noted increases in inflammation and scarring and measured those instances against the CCTA scans to locate perivascular fat changes occurring in the blood vessels.

Additionally, according to the research, the team examined the CCTA scans of over 100 patients who would go on to have a heart attack or cardiovascular death within five years of a CCTA. Using machine learning, the alterations in the perivascular space were used to predict such risks.

Meanwhile, the FRPs of 1,575 individuals were tested again in another trial where the fingerprint proved effective, achieving higher heart attack prediction rates than systems currently in place.

Eventually, the research team intends to roll out the tool to health care professionals all over the U.K. to be used in concert with CCTA tests.

The research appears in the European Heart Journal.