FUDiabetes

LifeLeaf by LifePlus non-invasive CGM watch just announced at CES

Anyone heard of this?

Here’s an article with more details, but still nothing on how it actually works, accuracy, or data:


They claim to be able to detect blood glucose from blood vessels under the skin using existing commercially-available optical sensors, but my understanding of this technology was that getting useful glucose readings from them is impossible. The company claims to have “found a way to utilize light from existing sensors to better separate glucose in the blood. The device then takes the isolated data and employs machine learning and artificial intelligence to deliver tracking metrics.”
Sounds like marketing bullshit to me. I would guess they plan to market this to non-diabetics based on all of the other biometric tracking sensors with the CGM data provided as a “bonus”, since it won’t be accurate enough to be useful to anyone who really needs it. It looks like a “wellness” product that doesn’t need FDA approval and not a medical device, but the testimonials section has several from Type-I diabetics with one even claiming that it isn’t subject to the delay of other CGMs and “You just have to check your watch and treat!”. Seems pretty sketchy to be making medical treatment claims via testimonial proxy…

1 Like

Sounds exactly like the Glucowatch, which was a trainwreck of epic proportions. The marketing pitch is pretty much the same too.

I thought about the Glucowatch, but this is a bit different. The Glucowatch used stick-on pads to get glucose readings from interstitial fluid. It provided actual, although not very accurate and very delayed BG readings. This thing claims to use non-contact commercially-available optical sensors to measure glucose in the blood, which I understand to be impossible. The Glucowatch actually worked, but with substantial skin irritation issues and not well enough to be worth bothering with. This sounds like it will not work for BG at all, but will be sold as a “wellness” product to people who will use it mostly for all of the other sensors and won’t care that the CGM data is useless. I am, however, VERY concerned that they seem to be covertly marketing it as an clinically accurate CGM for T1D via their testimonials.

1 Like

this is using light to calculate the glucose value. It might not work (it MOST LIKELY does not work), but it’s unlikely to leave scabs on your wrist, etc.

1 Like

I am not sure it’s actually impossible to get glucose values from light readings in the future. I’m just skeptical they’re anywhere close right now.

I suppose it could be possible, but I understand it to be impossible based on a presentation I saw at a major medical device conference a few years ago (BIOMEDevice 2017) by a researcher in this area who owns a company that sells optical sensors for wearables (Steven LeBoeuf of Valencell, Inc.), including medical grade ones (this was a major focus of the presentation). He would love for the CGM market to open up to him, but I distinctly remember this line from his presentation:
“There are only three things that are certain in life- death, taxes, and the utter absurdity of truly non-invasive medical-grade glucose measurement.”
He went on to explain why this was the case, but I don’t remember those details. But I was convinced and will be very skeptical of any person or company who claims otherwise until I see some data. I’ve still got the PDF of his presentation in case anyone is interested, but I don’t know how to attach a 22-page PDF here.

He also cited this in his presentation, which is a great read for more background in the area- http://www.mendosa.com/The%20Pursuit%20of%20Noninvsive%20Glucose,%20Fourth%20Edition.pdf

1 Like

It could be Raman spectroscopy through which they are able to make the measurement. If it is, it is possible that it will work. However their website has 0 technical details, and I can’t find the pending patent they are crowing about to see what technology they are using. So until more details are present, I think calling bs is fine.

Edit: Found a paper which discusses the technology. http://ijsrset.com/paper/5199.pdf
International Journal of Scientific Research in Science, Engineering and Technology (www.ijsrset.com) Hussain Khanetal. Int. J.S.Res. Sci. Engg.Technol.September-October-2018;4(10) : 367-376.
SensorTechnology is: Photo-Plethysmography (PPG) combined with Machine Learning

1 Like

If you are interested in looking at all the possible methods, this is a recent review paper on the problem and is free:

https://www.researchgate.net/publication/338428896_Noninvasive_Glucose_Measurement_Using_Machine_Learning_and_Neural_Network_Methods_and_Correlation_with_Heart_Rate_Variability

1 Like

Are any commercially available wearable sensors capable of Raman spectroscopy? I thought that needed a decent-sized piece of lab equipment.

Not that I know of, it would be a huge breakthrough and gives a possible mechanism by which a non-invasive measurement could be close to real time that I fully believe would be accurate. Without that breakthrough, I won’t be throwing the G6 away any time soon. I have a hard time believing that PPG will give them adequate results unless their machine learning algo is truly spectacular and their scientists are looking at something new in the results.

PPG sensors are what the guy that gave that talk sells, so that is his primary expertise and most of his talk was how PPG sensors are better than bioimpedance & ECG.

1 Like

Thanks- that’s useful. Hindawi throws up a bit of red flag and lines like “The following optical noninvasive methods have been analyzed to more or less successfully detect the diabetes level” do not inspire confidence, but this is a very comprehensive compilation. I just wish there was more synthesis of the data and commentary on what is promising (and what definitely is not) than what is possible and what might happen in the future.

I’ve also run into some people who were attempting to use nanoparticles of some kind to somehow to implement a non-invasive CGM system, but they couldn’t really tell me any details. Not that I’d have understood or remembered them anyways. I really hope someone has something groundbreaking right around the corner, but I’m more jaded, cynical, and suspicious than optimistic nowadays.

1 Like

I agree with you, and I have studied the area extensively. Nobody seems to talk about how close the actual measurements and trends are to real life. I will stay skeptical just like you. The LifeLeaf product seems like an over-hyped silicon valley deal that will get lots of funding then fade because they can’t deliver results.

Like the G6 and unlike fingersticks which use capillary blood.

I believe it doesn’t matter how you measure it, if you are measuring interstitial fluid you are a few minutes back from capillary blood and that is a few minutes back from what we actually feel and know; the arterial blood flooding our brains.

In Star Trek we simply put a probe into our brains and find out the fundamental truth. In the real world it doesn’t matter how you measure it, what matters is how accurately you predict the future, since you are measuring the past.

So that may seem a weird point to make, but the only limit of science is what you can do with it. Knowing my interstitial blood glucose is 10 minutes too late, telling me what it will be in 10 minutes time is my wet dream.

sure, maybe now it’s too bulky/expensive/impractical for peoples’ purposes. But… miniaturization is a much more tractable problem then “this defies all known laws of physics.”

Somewhat related to the heart rate variability (HRV) paper you posted, I have tried using my HR monitor to see if it would help me determine lows when running.

Low BG definitely causes an increase in my heart rate when running. So in theory I could use it. But in practice, there are so many other things that influence HR when running. A bunch of things. So I can’t really use it.

So I wonder if the same problem would apply in using it for a CGM. How could they possibly factor out so many things.

1 Like

Yeah, I am not sure they can. The technology sounds wonderful, but when people aren’t willing to share the technical details (even at a cursory level) I get nervous that it won’t work as well as advertised. In the literature paper talking about their solution it sounds like they are using the PPG sensor to monitor the capillary beds for changes and predict using a machine learning algorithm. Doesn’t completely defy the realm of possibility, but seems a little out there to me. I would love if they would publish some results or go head to head with a G6 or something.

1 Like