A cutting-edge AI-powered video system could soon provide a quick, non-invasive method for screening high blood pressure and diabetes, offering an alternative to traditional blood tests and blood pressure measurements. According to a preliminary study set to be presented at the American Heart Association’s 2024 Scientific Sessions, this breakthrough technology could potentially allow individuals to monitor their health at home, with the possibility of early detection and treatment for conditions that often go undiagnosed due to avoidance of medical exams.
The system uses high-speed video to capture 150 images per second of the face and hands, detecting subtle changes in blood flow associated with high blood pressure and diabetes. An AI algorithm processes these images to identify telltale signs of these conditions.
Key findings from the study indicate that the video system, combined with AI analysis, accurately detects high blood pressure and diabetes:
High blood pressure detection: The AI system was 94% accurate in identifying stage 1 hypertension (defined by the American Heart Association as 130/80 mm Hg or higher), when compared to continuous blood pressure monitors.
Short-duration video accuracy: A 30-second video was 86% accurate in detecting elevated blood pressure, while a 5-second video was 81% accurate.
Diabetes detection: The video system was 75% accurate in identifying individuals with diabetes, based on comparison with hemoglobin A1c blood tests, which measure average blood sugar levels over 1-2 months.
Study author Ryoko Uchida, a researcher at the University of Tokyo, noted that the detection of blood flow irregularities could help identify diabetes, particularly since peripheral neuropathy—a common complication of diabetes—impacts blood circulation in the hands and feet. She emphasized that this non-invasive method could be a game-changer for patients who are reluctant to undergo traditional testing.
Despite the promising results, the system is still in its early stages. Further research is needed, particularly to improve its accuracy in detecting diabetes, and address potential issues with arrhythmias (irregular heartbeats) that could affect blood pressure readings. The researchers envision a future where affordable sensors, potentially integrated into smartphones or even mirrors, could allow users to monitor their health in real-time.
In addition, Uchida hopes to seek FDA approval for an at-home device to detect diabetes once the system’s accuracy is refined. The potential for a non-invasive, easy-to-use device represents a significant advancement in the fight against diabetes and cardiovascular disease.
While the initial results are promising, experts like Dr. Eugene Yang from the University of Washington caution that validation protocols for such technologies are still lacking. Until these systems undergo appropriate testing and receive proper approval, doctors should continue using validated devices for blood pressure and glucose measurements.
The study, conducted between 2022 and 2024 at the University of Tokyo, involved 215 adults, primarily Japanese and other Asian participants. The research was focused on individuals with varying blood pressure levels and diabetes diagnoses. However, the researchers acknowledge that these results may not be generalizable to other demographic groups.
In summary, while still in development, this AI-powered video system holds great potential for transforming the way we detect and monitor high blood pressure and diabetes, offering a faster, more accessible way to manage these major risk factors for cardiovascular disease.
Related topics:
Diabetes Prevalence Reaches 15.8% Among U.S. Adults, New Data Shows
New Insights into Beta Cell Defects and Their Role in Type 2 Diabetes
