Mid-infrared quantum cascade laser spectroscopy is used to noninvasively predict blood glucose concentrations of three healthy human subjects in vivo. We utilize a hollow-core fiber based optical setup for light delivery and collection along with a broadly tunable quantum cascade laser to obtain spectra from human subjects and use standard chemo-metric techniques (namely partial least squares regression) for prediction analysis. Throughout a glucose concentration range of 80-160 mg/dL, we achieve clinically accurate predictions 84% of the time, on average. This work opens a new path to a noninvasive in vivo glucose sensor that would benefit the lives of hundreds of millions of diabetics worldwide.
All Science Journal Classification (ASJC) codes
- Atomic and Molecular Physics, and Optics