Know Labs says study confirms feasibility of non-invasive glucose sensor

Know Labs on Friday unveiled data from a feasibility study demonstrating that its non-invasive Bio-RFID sensor was able to deliver "stable, repeatable results" at predicting blood glucose concentrations when using Dexcom's G6 continuous glucose monitor as a reference device. The findings were presented at the American Association of Clinical Endocrinology (AACE) annual meeting.

Know Labs last year received the go-ahead to conduct a study comparing its sensor to established glucose monitoring technologies. CEO Ron Erickson said the latest study "demonstrates our progress toward getting the first FDA-cleared non-invasive glucose monitoring device in the hands of the nearly 40 million people living with diabetes in the US." The company has so far collected 1.4 billion data points that are now being used to refine its glucose value prediction algorithm, he added.

Harnessing electromagnetic energy

Know Labs conducted a series of internal studies between December 2022 and February 2023 to validate the technical feasibility of its sensor at quantifying blood glucose concentration (BGC) in five healthy participants using the Dexcom G6 as a reference device. The Bio-RFID sensor, which uses electromagnetic energy to non-invasively capture molecular signatures, can be integrated into wearable, mobile or bench-top form factors. The study's aim was to determine hardware and software infrastructure stability, as well as to collect additional data to identify the accuracy of the sensor at quantifying BGC in vivo non-invasively using radio frequency.

For the study, participants placed their forearms on the company's sensor and consumed liquid glucose to simulate a glucose spike. Their BGC was monitored for three hours while readings from both the Bio-RFID sensor and the Dexcom G6 were logged, collecting data on a continuous basis. Know Labs then used the data from its sensor to train a neural network model to predict readings from the Dexcom G6 reference device.

Across five participants, 46 tests and 92 samples, the study collected 4.7 million data points per sample or roughly 430 million data points for all samples. The study resulted in a mean absolute relative difference (MARD) of 20.6%, while it also performed with 46% of predictions within the FDA's recommended limit for accuracy for new blood glucose monitors.

The latest findings come after Know Labs recently presented results of a proof-of-principle study conducted with Mayo Clinic at the American Physiological Society (APS) Summit.

Don't want to miss our top stories? Sign up for our free daily newsletter here.