A group of Indian scientists have developed an artificial intelligence (AI) system based on the properties of individual heartbeats captured on an ECG (electrocardiogram) that can accurately predict diabetes and pre-diabetes.
The researchers from Nagpur's Lata Medical Research Foundation used clinical data from 1,262 people. Each participant had a conventional 12-lead ECG heart trace lasting 10 seconds. In addition, 100 distinct structural and functional features for each lead were integrated for each of the 10,461 single heartbeats recorded to create DiaBeats, a predictive algorithm.
The DiaBeats algorithm swiftly recognised diabetes and prediabetes based on the shape and size of individual heartbeats, with an overall accuracy of 97 percent and precision of 97 percent, regardless of relevant parameters such as age, gender, and co-existing metabolic illnesses.
Important ECG findings matched the known molecular mechanisms underlying the cardiac alterations seen in diabetes and pre-diabetes.
The approach might be used to test for the disease in low-resource settings if validated in larger research, according to the scientists.
"In theory, our work offers a reasonably inexpensive, non-invasive, and accurate alternative (to current diagnostic procedures) that can be utilised as a gatekeeper to detect diabetes and pre-diabetes early in their course." "However," they added, "adoption of our technique into everyday practise will require strong validation on external, independent datasets."
Corbevax as booster dose ready for adults from tomorrow
Vegetarian women are more prone to fracture hips in later life
