A bold breakthrough is on the horizon: researchers claim to measure blood glucose noninvasively using imaging, potentially ending the ordeal of finger pricks for many people with diabetes. This work, developed at MIT, could offer an alternative to traditional monitoring devices that rely on skin-penetrating sensors or regular fingerstick testing.
Today, numerous individuals with diabetes rely on continuous glucose monitors (CGMs) to track their glucose levels. While CGMs avoid daily fingersticks, they still involve a small interstitial needle and have a wear window limited to about 15 days. Some patients continue to rely on fingerstick tests, drawing blood from the finger several times daily to check glucose.
The MIT team employs Raman spectroscopy, a technique that reveals tissue chemistry by directing near-infrared or visible light at it. They built a compact, shoebox-sized device around this method to gauge blood glucose without needles. In tests, the method produced readings comparable to those from a commercial CGM in a healthy volunteer.
Though the current prototype is not yet a wearable sensor, the researchers are pursuing a wearable version that is currently under evaluation in a small clinical study.
“Finger sticks have long been the standard for measuring blood sugar, but the daily pricks are unwelcome,” commented Jeon Woong Kang, an MIT research scientist and the study’s senior author. “If a noninvasive glucose monitor can achieve high accuracy, virtually all people with diabetes could benefit from this technology.”
Arianna Bresci, an MIT postdoctoral researcher, led the study, which was published in Analytical Chemistry. Other collaborators include Peter So, head of MIT’s Laser Biomedical Research Center and a professor of biological and mechanical engineering, along with Youngkyu Kim and Miyeon Jue from Apollon Inc., a biotechnology company based in South Korea.
What makes the noninvasive approach work
The MIT team aimed to build a comfortable wearable glucose monitor. Their key breakthrough came from refining Raman spectroscopy to detect glucose signals that are typically overwhelmed by other molecular signals in tissue. By illuminating the skin with near-infrared light at one angle and collecting the Raman signal from a different angle, they can filter out interfering information.
Initial measurements used equipment about the size of a desktop printer. Since then, the system has been miniaturized. In the latest work, the researchers narrowed attention to three bands in the Raman spectrum—out of roughly a thousand bands—one band corresponding to glucose and two accounting for background signals. This simplification lowers both the data burden and the hardware requirements.
In a four-hour session, an arm rested on the device as near-infrared light passed through a small glass window into the skin to take readings. Each reading required just over 30 seconds, with measurements taken every five minutes.
During testing, the participant consumed two 75-gram glucose drinks to induce changes in glucose levels. The results showed accuracy comparable to two commercial CGMs worn simultaneously by the same person.
Looking ahead, the team has developed an even smaller prototype roughly the size of a cellphone and plans a larger hospital-based study next year to include participants with diabetes. The ultimate goal is to shrink the device further to the size of a watch and ensure reliable readings across diverse skin tones.
If successful, this noninvasive approach could transform diabetes management by delivering accurate glucose information without finger pricks, improving monitoring consistency and reducing complications tied to under- or over-testing. Still, questions remain about real-world performance, long-term wearability, and how well the method works across varied populations. Would you welcome a wristwatch-like glucose monitor in daily life, or would you reserve judgment until larger-scale trials confirm its reliability?