Health AI models predict metabolic risks
HealthTech AI company NuraLogix has announced that it has developed AI models that can predict a person’s risk of Type 2 Diabetes, Hypercholesterolemia, Hypertriglyceridemia and Hypertension. This is important, says the company, because it will someday allow people to screen themselves using any device with a video camera such as a smartphone or tablet.
Researchers at the company trained their machine learning-based models on the facial blood flow patterns of tens of thousands of subjects who had been diagnosed with the above conditions. All of the models predicted with a minimum Area Under the Curve (AUC) > 0.80. Compare this with the classic Framingham model for predicting cardiovascular disease which had an AUC in the range of 0.70, says the company.
“This technology is a disruptive tool for population health,” says Dr. Keith Thompson, Primary Care Physician of over 30 years, and Chief Medical Officer at NuraLogix, “We live in an era where one half of the globe is without access to health care, and in both developed and underdeveloped countries, there are shortages of Human Health resources.”
“Traditional medical encounters are time consuming and costly for most people,” says Thompson. “Tools like this allow us to democratize non-communicable disease (NCD) screening globally to regions in need by moving away from traditional physician face to face encounters and laboratory services, by allowing screening through the ubiquitous technology of smart phones. Given the vast majority of people with metabolic risk factors are unaware they have a condition and may not have access to screening in its current forms, the technology from NuraLogix can literally change the global course of human health history.”
Dr. Naresh Vempala, Director of Research at NuraLogix adds, “Type 2 Diabetes, Hypertriglyceridemia, Hypercholesterolemia, and Hypertension have commonalities which are often associated with metabolic syndrome. However, each of the above conditions have distinct characteristics of their own. What makes our touchless technology seminal is that our classifier models can extract hidden and unique information about each of these health risks through facial blood flow patterns and predict them with a high degree of accuracy. This is unprecedented in health AI.”
The company’s technology can measure 30+ health and wellness parameters using Transdermal Optical Imaging (TOI), a patented technique developed by the company in which a conventional video camera is used to extract facial blood flow information from the human face. This is demonstrated in the company’s app Anura and Anura Lite.
