Using Machine Learning to Predict Stem Cell Knee Arthritis Outcomes

By /

stem cell knee outcomes

One of the most frustrating things about stem cells and regenerative medicine is the difficulty in predicting which patients will respond with “home run” results and which won’t respond at all. You would think from looking at the miracle-cure stem cell websites popping up around the Internet that stem cells are magic pixie dust able to cure whatever ails a human. However, that’s not the case at all, in that all stem-cell-based treatments have a success and failure rate. To fix that vexing problem, we’ve been working for years on predictive models that help Regenexx providers decide who is more likely to respond. Our newest model now takes this to the next level by measuring the levels of 25 cytokines and growth factors from stem cell knee patients and inputting this into a predictive model based on machine learning.

What Is Machine Learning?

The human brain is great at deciphering patterns. Take, for example, that your brain can easily recognize thousands of faces. This would seem like a trivial task, but until recently, despite that fact that a computer can process information millions of times faster than your single neurons in your brain, it was awful at recognizing faces. This is because your brain and today’s computers are built quite differently. A computer processes visual information in serial (one piece at a time), and your brain processes the whole picture all at once.

neural network visionTo interpret a picture (like the one of the Coke can), your retina has a two-dimensional array of sensors tied to individual neurons that then connect to other neurons. As the stimulus passes through each layer of millions of neurons, each has a specific weight for various parts of the information that it “remembers.” In the end, that information converges on a collection of neurons that are stimulated in a certain pattern that your brain knows is a Coke can. Since the time it takes the signal to pass through six or seven layers is very short, your brain can recognize the object (or a face) quickly.

A computer using traditional software is awful at recognizing Coke cans. It needs to inspect each and every pixel in the image and compare that to some known picture of the object. Processing that information can take more time than it takes your brain. In addition, throw the computer a variation by letting a shadow fall across the can and it now thinks the can is something else.

Back in the ’60s, in the golden age of early computing when the machines occupied whole rooms, computer scientists believed they were on the verge of creating a human brain in a box. Their hopes were quickly dashed as they realized that tasks like recognizing a Coke can with the technology available back then could take a computer several days. Some smart scientists reasoned that if we designed software with layers of artificial neurons, we could sort of model how the brain processed information. This software approach was called a neural network or machine learning. However, this field labored in obscurity until a seed change happened about five years ago. At this point the big money in companies like Google, Microsoft, Facebook, and others began to realize that they needed machine learning if their products were to ever see their full potential. For example, you may have seen the fanfare behind Google announcing that its image software had been taught to recognize a cat! While this seems trivial, for a software program based in machine learning, it was a big deal.

Request a Regenexx Appointment

Predicting Stem Cell Knee Arthritis Outcomes

We’ve tracked the outcomes of every treated patient in a registry for many years. In fact, we’re the only clinic using stem cells that has done this to date. One of the things that we’ve been able to use all of that data for is to help Regenexx providers determine who is a good or poor candidate for stem cell therapy. However, some things that we thought would be associated with outcome, like the severity of arthritis on an MRI, turned out for some joints not to be related. I’ll never forget spending two weeks pulling out all of the minute details of knee MRI images on patients for whom we knew the final outcome of the stem cell therapy, only to be horribly disappointed that none of it correlated with that outcome!

After the knee MRI fiasco, I knew we needed more data to plug into our predictive model. One commonsense way to do this occurred to me while in our garden. Every industrial farmer on earth knows that the soil conditions must be right if your seeds are to reach their full potential. Could we do the same thing for knees by measuring the status of the normal fluid in the joint, which in effect is the same as the soil in which our stem cell seeds were to be planted? Thus began our knee microenvironment study, measuring 25 different cytokines and growth factors in the knee before and after every Colorado stem cell knee procedure for knee arthritis.

A Glimpse at the Future

Collecting 50 data points on every patient of obscure chemicals inside the knee quickly produces a data set where patterns can be really hard to recognize. Should TGF-beta go up or down? How about anticatabolic molecules like TIMP, A2M, or IRAP? When we first got the data back from the pilot phase of this IRB-approved study, it was difficult to see any specific patterns. Given that no one has ever attempted to use a large panel of the levels of these chemicals in the knee to predict the outcome of treatment, there was no guidebook. That’s when it hit me that we needed a neural network, which can detect patterns in chaotic data that a mere human might miss.

This week, we got back our first result on a predictive model using only the first 14 patients for whom we had an outcome. Our biostats expert trained the system about 100 times on randomly chosen patients with various marker levels, and we ended up with a system that was 67% accurate in predicting outcome. Given that we now have 43 patients in the pipeline awaiting a final outcome measure, by the early summer we should have a system that gets better and better at predicting outcome based on the quality of the soil and know which of these growth factors is the most responsible for that outcome. By the end of the year, we could have an interesting tool that may help Regenexx providers more accurately determine who is a good and a poor candidate for a stem cell knee procedure.

The upshot? Our knee microenvironment predictive model is a great example of the personalized healthcare revolution coming our way. From a pragmatic research standpoint, it should help Regenexx providers determine who they should dissuade from a stem cell knee procedure. From a scientific standpoint, it’s exciting to see how software designed to recognize cats on the Internet may help patients get better care!

Join us for a free Regenexx webinar

This blog post provides general information to help the reader better understand regenerative medicine, musculoskeletal health, and related subjects. All content provided in this blog, website, or any linked materials, including text, graphics, images, patient profiles, outcomes, and information, are not intended and should not be considered or used as a substitute for medical advice, diagnosis, or treatment. Please always consult with a professional and certified healthcare provider to discuss if a treatment is right for you.

Get Blog Updates by Email

By submitting the form, you are agreeing that you read and consent to our Privacy Policy. We may also contact you via email, phone, and other electronic means to communicate information about our products and services. We do not sell, or share your information to third party vendors.

Category: Knee, Latest News

Leave a Reply

Your email address will not be published. Required fields are marked *

34 thoughts on “Using Machine Learning to Predict Stem Cell Knee Arthritis Outcomes

  1. stef

    For a country that glorifies ego and money you look more an more like an alien… I do not know why you put up with all this work for all of us but merci.

    1. Regenexx Team

      Eloquently stated! Thank you!

  2. Connie Fuller

    I am getting a stem cell procedure this week. I am having to pay for it because medicare will not. why wont it because when the Doctor decides who is a candate and who is that should help the outcome and plus what does it
    cost to have a knee replacement . I sure wish medicare would pay for this . I,m not going to want for my knees to get worst see if medicare will pay for it .

  3. Leland J Hoepfinger

    Dr C- -once again you are to be commended for going the extra mile beyond the “average”-at best- stem cell clinic! It will be very interesting to me to hear of your results as one whose stem cell procedure for my arthritic knee was a guarded success and certainly not a home run—-I must have “bad soup”???–It would be great to know how my cytokines and growth factors profile compares as I consider whether to go for a second stem cell treatment on the treated knee as well as possible future results of my deteriorating “other” knee–Would love to hear what my profile was????–Keep up the good work! Thanks Lee H

  4. Christine GradyMiller

    I am strongly considering this procedure. Would PRP be something to start out with and would you consider someone like myself to be a test subject?. I am very healthy other than the DJD of the knees. But they have already said, at age 51, nee replacement will be imminent. Of course, that is all they ever want to consider, always the money maker for them.I have been using the viscosupplements ( Hyalgan) with good results but the effectiveness is lessening. Would you consider me for a trial run????

  5. Lewis

    Love to know what my profile would be. I had both knees done at your clinic in San Rafael. I yielded 1.1 billion stem cells in the extraction. I am feeling better 3 months out!

  6. Venkatachalam

    Congrats Dr. Centeno on this pioneering work.
    Your mission is to discover the truth, cone what may.
    However my initial reaction to neural networks and artificial intelligence is that it doesn’ t work. So I am not sure how you can add a neural layer to the artificial intelligence?

  7. Chris Centeno Post author

    Connie, it will be a good number of years before Medicare covers these procedure. They’re just too new at this point. Having said that, I have little doubt, given the amount of research going on this area, that coverage is likely at some point.

  8. Chris Centeno Post author

    Leland, just send me an e-mail and we can take a look at that data.

  9. Chris Centeno Post author

    Christine, we often have open studies-several in the knee at this point. The most recent one specific to knee arthritis requires you to be local to Colorado as it looks at how this panel of growth factors and cytokines changes over 6 weeks (with samples taken before, 2 weeks, 4 weeks, and 6 weeks after the therapy) with a specific intervention. Our clinical research guru is Dr. Dodson and she can be reached at: [email protected]

  10. Chris Centeno Post author

    Lewis, 1.1 billion cells is a fantastic yield, congrats! We will likely be extending this study out to select network sites over the next year.

  11. Chris Centeno Post author

    Dr. V (AK)-every search you run on the internet, every photo that is tagged with your name on it, every text you dictate into your phone to have it transcribed, and much of the data mined by fortune 500 companies uses a neural network (NN). Check out this review of just recent news: NNs are now being used in self driving cars and the teaching computers how to beat humans at strategy games like Go.

    The big money has entered this space. As an example, right now we’re using a stats package called “R” (The R Project for Statistical Computing), but we’re moving the model to Google TensorFlow soon (this is an open source neural network that Google uses in many of it’s projects-see IBM has been creating a hardware processor capable of mimicking the thought process in the human brain, see . So NNs will be a bigger and bigger part of our daily lives over the next decade.

  12. Dr Susan Ellis

    I have recently had the SD procedure on my right knee despite being told I was in the fair to poor outcome category. At six weeks out I am cautiously optimistic about the results because of lessened pain. I am most interested in the predictive model because my left knee will need attention soon. Please keep me in the loop.

    1. Regenexx Team

      Dr. Susan Ellis,
      We will definitely keep you in the loop. Glad to hear that you are having less pain, and hoping the final result blows the “fair to poor” out of the water! Rather than being used to guide clinical recommendations, our goal for the Knee Microenvironment research study discussed in the blog is to produce a predictive model that will fine tune Candidacy. It would require certain regulatory approvals if this data were ever to be routinely used to guide clinical decisions.

  13. Sandra A O'DONNELL

    Dr. Centeno- you are so amazing and I do hope that the NEW studies will help to help all of us who have struggled, but will not give up and will continue with knee issues. Hope to get to Colorado in July, if you will be there! I will stay in touch with office and hope I have enough cells to use. Having known you for what seems like “forever” I can’t think about where I would be today! Thanks so much!. SANDY O’DONNELL

    1. Chris Centeno Post author

      Thanks for the kind words Sandy!

  14. Jerry Allee

    Is there anyone in Ohio that is currently using this technique?

  15. Eugene Lamski

    Dr. Centeno-I had the procedure done on 10/07/2015 at your San Rafael clinic. I wore the braces for over one month and started taking the Regenexx stem cell formula soon thereafter. I am 81 years of age and exercise three times a week at the local gym. As of this date I can’t confirm or deny any distinct improvement in either knee. My walking has improved slightly, however it is still quite painful to descend stairs. My stem cell count was also 1.1 billion. I”m hopeful with time I will see improvement and will continue to take the Regenexx formula. I had a followup conversation at about 3 months but nothing since. I am available for any additional examination or testing.

  16. Mary DeVasher

    is there anyone in Tennessee currently using this tehnique?

    1. Regenexx Team

      Here’s the link to find the Regenexx Provider closest to you: There is not currently one in Tennessee, however there are several in the general area.

  17. Sandy

    I have a right knee medial meniscus complex tear in the white/white zone. Also some degenerative arthritis. MRI results cite ganglion cysts too. All ligaments are fine. I also have per anserine bursitis in that knee.
    Can you’re procedure help my multiple issues or would I be a candidate for Grand Cayman instead of SD in US?
    Prognosis for improvement? I am female, 65, healthy, but not active.

    1. Regenexx Team

      We treat all of those issues, and by the time we see a knee there is rarely just one. Treatments are customized to the patient and to the issues: However to know what would likely be needed in a particular case a Candidacy evaluation is needed. Please fill out the candidate Form here:

  18. Dianne Riehl

    I had Regennex done in July after having perfectly normal cartilage deatroyed by cortisone in 2 years time. My knee with moderate arthritis is 100% healed. It’s a miracle. My knee with severe arthritis is about 35% healed. I had my procedure done in VT. I can’t say enough good things about the brilliant Dr. Fenton. Why did one knee progress faster than the other? Will I continue to see amazing results with my bad knee? Should I get more PRP to help? Thanks so much. I recommend Regennex to everyone.

    1. Regenexx Team

      Cortisone is infamous for that scenario…if we could only get Doctors to stop using it! Regarding the difference in current results in your knees, there could be many different factors at play. Was the cortisone oral, or by injection?

  19. Dianne Riehl

    Thanks for your reply. The cortisone was by injection. I was being treated for Patellafemoal disorder. I had 4 cortisone shots in the bad knee and went to bone on bone in 2 years time. The better knee only had 2 injections of cortisone so my cartilage was not totally gone. Can you imagine my horror when my doctor told me I had 18 year old cartilage. Then 2 years later he told me I was bone on bone. Will Regennex still help that knee without cartilage? Thanks so much.

    1. Regenexx Team

      Yes. Cartilage is rarely the problem when it comes to pain. But there are so many other issues in a knee badly damaged with Cortisone that it started at a very different place than the other knee. I would contact Dr. Fenton and update him on your progress. Having examined and treated you, he would be the one to determine whether in this particular case something is needed to give it that push to get over the finish line, or, whether he feels that knee will catch up to the other.

    2. Chris Centeno Post author

      Dianne, yes, we have helped many patients with no cartilage in one or more areas of the joint. The big difference for these patients seems to be that they need update treatments every 2-4 years.

  20. Teresa

    My dad has bone on bone in one of his knees. He is 91 yrs. old and in excellent health, Could he benefit from the stem cell procedure?

    1. Regenexx Team

      Likely, as we’ve seen no evidence that age impacts the outcome of our knee stem cell procedures. Please fill out the Regenexx Candidate form, and that will set up a phone Candidacy consult so we can bet a better idea in his specific case.

  21. Dianne Riehl

    Dr Centeno, Thanks for your comment. What does “update treatment every 2-4 years mean?” Are you referring to more stem cells? Or other procedures?

    Thanks for your dedicated work.

    1. Regenexx Team

      Yes, in situations like you describe, stem cell updates over time can be needed.

  22. FrankC

    Dr Centeno,

    I had both knees injected 6 weeks ago with Dr. Movva here in Dallas. Results so far in order of appearance:

    1. Persistent ache in both knees quit.
    2. Glitches (not a medical term, but engineers will understand) in the range of motion eventually ceased in both knees. I can now confidently climb stairs and ladders with both legs in sequence.
    3. I would get a severe pain if I over-pronated accidentally by stumbling into a depression in the yard. I now wince needlessly when that happens. I’m sure I’ll get over it. After all, it’s just neural re-programming isn’t it. :o)

    I’m 68. 2-4 years is a good deal for an update. Most software doesn’t last that long. I’ll start saving now.

    Thanks for your efforts. Keep up the good work.

    1. Regenexx Team

      Frank C,
      Loved your update! Sounds like both you and Dr. Movva have done a great job!

Is Regenexx Right For You?

Request a free Regenexx Info Packet


Learn about the #1 Stem Cell & Platelet Procedures for treating arthritis, common joint injuries & spine pain.

Join a Webinar


Get fresh updates and insights from Regenexx delivered straight to your inbox.

Subscribe to the Blog


9035 Wadsworth Pkwy #1000
Westminster, CO 80021


Copyright © Regenexx 2021. All rights reserved. | Privacy Policy

*DISCLAIMER: Like all medical procedures, Regenexx® Procedures have a success and failure rate. Patient reviews and testimonials on this site should not be interpreted as a statement on the effectiveness of our treatments for anyone else.

Providers listed on the Regenexx website are for informational purposes only and are not a recommendation from Regenexx for a specific provider or a guarantee of the outcome of any treatment you receive.