Can a pre-treatment MRI tell us what will happen after stem cell therapy? The paper’s first author explains
© 2020 The Japanese Society for Regenerative Medicine. Published by Elsevier B.V. Reproduced under CC BY-NC-ND 4.0.
At present, a pre-treatment MRI alone cannot accurately predict what will change after stem cell therapy.
The study did find that pre-treatment MRI findings — cartilage damage, osteophytes and bone marrow lesions — were associated with some of the clinical scores measured afterwards. The direction of that association ran opposite to what the research team had expected.
Note: this was a retrospective observational study without a control group. It does not establish or guarantee efficacy or safety.
When people consider regenerative medicine for knee osteoarthritis, the first question is usually whether it works at all.
But two people can receive the same treatment and have very different experiences afterwards. So is there anything we can know in advance?
That is exactly the question behind this study, published in 2020 in Regenerative Therapy, the English-language journal of the Japanese Society for Regenerative Medicine.
Key points
Study overview
| Title | Associations of clinical outcomes and MRI findings in intra-articular administration of autologous adipose-derived stem cells for knee osteoarthritis |
|---|---|
| Journal | Regenerative Therapy (2020), Japanese Society for Regenerative Medicine |
| Design | Retrospective observational study, no control group |
| Where conducted | A medical institution in Japan (not Ginza YR Clinic) |
| Participants | Clinical scores: 34 patients, 57 knees. Pre-treatment MRI available: 19 patients, 34 knees |
| Inclusion criteria | Kellgren-Lawrence grade 3 to 4, unresponsive to other conservative treatments |
| Mean age | 67.5 years (10 men, 24 women) |
| Treatment | Intra-articular injection of cultured autologous adipose-derived stem cells |
| Cells injected | Approximately 82.6 million on average |
| Assessment points | Before injection, then at 1, 3 and 6 months |
| Role of our director | First author |
| DOI | 10.1016/j.reth.2020.04.003 |
Retrospective observational study means the medical records of people who had already been treated were reviewed after the fact. Because there was no untreated control group, the influence of natural progression and placebo effects cannot be ruled out. In addition, 319 patients and 486 joints were treated at that institution during the same period; this study analysed only the 34 patients and 57 knees for whom written consent was obtained, so some selection bias may be present.
The Kellgren-Lawrence classification grades osteoarthritis severity from 0 to 4 based on X-ray findings. This study covered grades 3 and 4, which correspond to advanced and end-stage disease. Early and mild cases (grades 0 to 2) were not included.
Finding 1: pain changed mainly in the first month
The study tracked change using two measures.
VAS is a 0 to 10 scale on which patients rate their own pain. A lower number means less pain.
KOOS is an international questionnaire that scores the knee across five domains: pain, symptoms, activities of daily living, sports and recreation, and quality of life. A higher number means a better state.
The figures reported over six months were as follows.
| Measure | Before injection | At 6 months |
|---|---|---|
| VAS (pain) | 6.1 on average | 3.5 on average |
| KOOS (total) | 54.4 on average | 64.6 on average |
Beyond how much people improved, when the improvement appeared is also worth noting. On the study’s averages, most of the improvement in pain and symptoms appeared within the first month after injection. For VAS, KOOS-pain and KOOS-symptoms, the average improvement at one month already accounted for roughly 70 to 80 percent of the average improvement over six months. By contrast, scores for daily activities and for sports and recreation improved more slowly, emerging gradually over time. Overall, pain and symptoms improved relatively early, while daily activities and physical function improved step by step.
More telling than the numbers themselves is when the change appeared.
VAS, KOOS-pain and KOOS-symptoms reached roughly 70 to 80 percent of their six-month average change within the first month.
This does not mean that 70 to 80 percent of patients improved, nor that pain fell by 70 to 80 percent. It means that most of the average change recorded over six months had already occurred by the end of the first month.
KOOS-activities of daily living and KOOS-sports and recreation, by contrast, changed gradually.
Pain shifts first. Function follows later.
Even when pain eases, stairs, long walks and sport do not necessarily become easier at the same time.
Finding 2: MRI findings were associated with outcomes, but in the opposite direction
This is the heart of the study, and the part that needs the most careful reading.
The 34 knees of 19 patients who had a pre-treatment MRI were scored using MOAKS.
Patients were not divided by how much they improved, but by whether each score reached its MCID.
MCID (minimal clinically important difference) is the smallest change a patient can actually perceive as meaningful. The study used each score’s MCID to sort knees into an improved group and a poorly improved group. For VAS, the MCID was set at 2.0.
The results were as follows.
Findings in the meniscus, ligaments and tendons, synovium and periarticular structures showed no significant association with any score.
The paper describes these data as implying that the stem cell therapy provided some beneficial effect even for patients whose disease had already advanced.
This does not mean “the worse the knee, the better it works”
There are three reasons.
MOAKS is an international scoring system for semi-quantitative assessment of knee structures on MRI. It divides the knee into 14 subregions.
Bone marrow lesions (BMLs) are changes inside the bone beneath the cartilage. They can only be seen on MRI and do not appear on X-rays.
Finding 3: no association with cell count, culture period or age
The study also examined factors other than MRI findings.
| Factor | Association with outcome |
|---|---|
| Patient age | No significant association |
| Cell culture period | No significant association |
| Number of cells injected | No significant association |
It is worth noting that this study was not designed as a trial comparing different cell doses, and the number of cells injected was not randomly assigned.
So “no association was found” is not the same as “cell count has no effect on what happens after treatment”.
Finding 4: symptoms and imaging did not move together
This is the part of the study that deserves the most attention.
Of all 57 knees, only four patients also had an MRI six months after treatment. The number is small, but the result is instructive.
| Case | Symptoms | MRI findings |
|---|---|---|
| Patient 1 (female, 74) | Improved | No clear change apart from bone marrow lesions |
| Patient 2 (male, 65) | Some scores improved | No clear change apart from bone marrow lesions |
| Patient 3 (male, 55) | Did not improve | Bone marrow lesions decreased and the medial meniscus increased in size |
| Patient 4 (male, 54) | Improved | No structural improvement observed |
The results of these four patients show that even when a knee feels better, a corresponding change will not necessarily be visible on MRI.
The researchers believe the reduction in pain may be related to reduced inflammation inside the knee joint. The paper also notes that several studies have reported improvement in pain and symptoms after treatment, but there is not yet sufficient evidence that cartilage or the meniscus grows back. A knee that feels more comfortable reflects an improvement in symptoms. Whether cartilage or the meniscus has actually been repaired still needs further examination and research to confirm.
Safety
No serious adverse events or complications were reported during the six-month observation period.
| Event reported during the study | Cases |
|---|---|
| Joint effusion | 6 |
| Pain | 5 |
| Local heat | 1 |
| Stiffness | 1 |
| Infection | 0 |
All of these reactions were transient and resolved completely within a few days to one week in every patient.
Given the limited size of the study and the length of the observation period, however, no judgement about long-term safety can be drawn from this result.
The physician behind this study
Dr. Junya Higuchi, director of Ginza YR Clinic, is the first author of this paper.
He trained in orthopaedic surgery at the University of Tokyo, building clinical and surgical experience across joint, spine, trauma and sports medicine, and earned his doctorate in medicine at the same university’s graduate school through research on mesenchymal stem cells.
This work was also presented at the Japanese Society for Regenerative Medicine in the same year, 2020.
What this study tells us
First, change has an order to it. Pain and symptoms shift first; daily activities and sport follow later.
Second, symptom relief is not cartilage regeneration. The paper infers that the clinical changes are more likely to stem from reduced inflammation than from tissue repair. Going into treatment expecting the knee to be restored is likely to lead to a gap between expectation and outcome.
Third, accurate prediction is not yet possible. Pre-treatment MRI findings were associated with later change, but imaging alone is not enough to predict the outcome for an individual patient. That is the paper’s own conclusion.
The value of this paper lies not in showing treatment results, but in stating clearly what remains unknown.
This article is provided for general medical information and does not guarantee any particular treatment outcome. Symptoms and progress vary between individuals. Please consult a physician for details.
So is pre-treatment assessment still worthwhile?
Yes — not to predict the outcome, but to confirm the following.
Knee pain does not always come from osteoarthritis. Rheumatoid arthritis, septic arthritis and referred pain from the lumbar spine can all present as pain in the knee.
At our clinic, the physician reviews the course of symptoms, physical examination findings, imaging and medical history together before judging on an individual basis whether regenerative medicine is appropriate. If it is not, we say so.
Frequently asked questions
If you would like your knee assessed
At Ginza YR Clinic we review symptoms, physical examination findings, imaging and medical history together, then explain the options available to you. If regenerative medicine is not appropriate in your case, we will tell you so directly.
► Autologous adipose-derived stem cell therapy
Details of the treatment, fees, risks and side effects are set out on the page above.
Enquiries and appointments
WebBook online
By appointment only. Self-pay treatment, not covered by insurance.
About our practice
Attending physician
Dr. Junya Higuchi, Director, Ginza YR Clinic
PhD in Medicine, University of Tokyo. Board-certified orthopaedic surgeon, Japanese Orthopaedic Association. Certified physician, Japanese Society for Regenerative Medicine. Certified sports doctor, Japan Sport Association.
Regenerative medicine provision plan registration numbers
Autologous adipose-derived stem cell therapy (joint disease): PB3250224
Autologous platelet-rich plasma (PRP) therapy (chronic arthritis): PB3250225
Specified cell-processed product manufacturing: FC3250098
Registration numbers for other treatments appear on their respective pages.
A registration number indicates that notification under the Act on the Safety of Regenerative Medicine has been completed and the provision plan accepted. It does not mean the government has approved or guaranteed the efficacy or safety of the treatment.
Medical information notice
This article is provided for general medical information and does not guarantee diagnosis, efficacy or safety. Symptoms and their course vary between individuals.
References
- Higuchi J, Yamagami R, Matsumoto T, Terao T, Inoue K, Tsuji S, Maenohara Y, Matsuzaki T, Chijimatsu R, Omata Y, Yano F, Tanaka S, Saito T. Associations of clinical outcomes and MRI findings in intra-articular administration of autologous adipose-derived stem cells for knee osteoarthritis. Regen Ther. 2020;14:332-340. https://doi.org/10.1016/j.reth.2020.04.003
Last updated: 2 September 2026
“`