AI technology could be used to detect potentially deadly blood clots before they strike, according to research by scientists at the University of Tokyo, who say they have created a non-invasive way to observe clotting activity in blood as it occurs.
The new AI tool uses a powerful microscope to monitor platelets in real time, said Dr Kazutoshi Hirose, lead author of the study.
Platelets, which are tiny blood cells that help form clots to stop bleeding, play a crucial role in heart disease, particularly coronary artery disease, he said.
If the platelet count is too high, it can increase the risk of blood clots.
To prevent dangerous clots, patients with coronary artery disease are usually treated with anti-platelet drugs.
“However, it’s still challenging to accurately evaluate how well these drugs are working in each individual, which makes monitoring platelet activity an important goal for both doctors and researchers,” Hirose said.
How the tool works
Traditional methods of monitoring platelet activity often rely on indirect measurements or invasive procedures, reports Fox News.
“With the AI tool, a high-powered microscope works like a super-high-speed camera that takes sharp pictures of blood cells in flow,” said Yuqi Zhou, assistant Professor of Chemistry at the University of Tokyo.
“Just as traffic cameras capture every car on the road, our microscope captures thousands of images of blood cells in motion every second,” he said. “We then use artificial intelligence to analyse those images.”
The AI can distinguish between a single platelet (“like one car”), a clump of platelets (“like a traffic jam”), or even a white blood cell tagging along (“like a police car caught in the jam”), Zhou described.
“Right now, doctors often give blood-thinning drugs and hope they are working. With this method, they could actually see if the treatment is working.”
This technology stands out because it allows doctors to directly observe platelets in the bloodstream and measure how they interact and form clumps in real time, according to Professor Keisuke Goda, also from the university.
“Our studies have shown it to be highly effective in patients with coronary artery disease, the most common cause of heart attacks and one of the leading causes of death in the US and the world,” he added.
When the researchers tested the technique on more than 200 patients, they found that people with more serious heart problems had more clumping in their blood.
A blood sample taken from a patient’s arm and tested with the tool resulted in nearly the same results as blood taken directly from the heart’s arteries, they noted in their findings, published in Nature Communications.
Treatment implications
Harvey Castro, a Texas-based emergency physician and AI expert, described the implications as significant for patient care.
“Today, we rely on indirect lab markers or invasive cath-lab blood draws to judge clot risk,” Castro, who was not involved in the study, told Fox News Digital.
“This technology turns an ordinary venous draw into live telemetry of platelet behaviour, giving answers in seconds rather than hours.”
These advancements have the potential to change the standard of care for patients with coronary heart disease, according to the researchers.
“Typically, if doctors want to understand what’s happening in the arteries, especially the coronary arteries, they need to do invasive procedures, like inserting a catheter through the wrist or groin to collect blood,” said Hirose.
“What we found is that just taking a regular blood sample from a vein in the arm can still provide meaningful information about platelet activity in the arteries.”
Potential limitations
Castro cautioned that the technology is not yet ready for widespread use.
“The microscope is cutting-edge research gear, not yet a bedside device,” he said. “We need miniaturisation and cost drops before every ER can use it.”
Looking ahead, after further research, Castro envisions that this innovation could change point-of-care decision-making for patients.
“Five years from now, I can imagine a point-of-care analyser that lets me adjust anti-platelet drugs and how I titrate oxygen quickly and safely for the patient in front of me,” he said.
Study details
Direct evaluation of antiplatelet therapy in coronary artery disease by comprehensive image-based profiling of circulating platelets
Kazutoshi Hirose, Satoshi Kodera, Masako Nishikawa et al.
Published in Nature Communications on 15 May 2025
Abstract
Coronary artery disease (CAD) is a leading cause of death globally. Antiplatelet therapy remains crucial in preventing and treating CAD-associated thrombotic complications, but it concurrently amplifies the risk of bleeding. Unfortunately, traditional platelet function measurement methods cannot directly evaluate its efficacy and safety. Here we demonstrate comprehensive image-based profiling of circulating platelets to directly observe thrombotic conditions and assess antiplatelet therapy in CAD patients. Deep learning-based analysis of whole blood samples from 207 CAD patients revealed elevated concentrations of circulating platelet aggregates, especially in acute versus chronic coronary syndrome patients. It also indicated a regimen-dependent reduction in these concentrations upon treatment with antiplatelet drugs, thereby verifying the direct efficacy of the therapy. Notably, consistent concentrations of these aggregates were found in both venous and arterial blood, suggesting venous blood as a reliable therapy efficacy indicator, despite CAD’s arterial nature. These findings support personalised and improved antiplatelet therapy in CAD management.
See more from MedicalBrief archives:
Blood type tied to risk of early stroke – US study
AI algorithms in diagnosis could harm patients – Dutch study