Why this matters
Low-dose radiotherapy for osteoarthritis has traditionally been viewed primarily as a treatment for pain and inflammation. Whether it can actually alter the structural and long-term clinical course of osteoarthritis is a much more consequential question.
This randomized trial now brings together three levels of evidence from the same patient population: patient-reported symptoms, blinded structural imaging, and clinical outcomes extending beyond a decade.
The consistency across those domains is the central story. The data do not prove disease modification, but they provide an unusually coherent signal that low-dose RT may be doing more than temporarily relieving symptoms.
Study design
Between 2012 and 2014, 292 patients with symptomatic Kellgren-Lawrence grade 0-2 knee osteoarthritis were randomized 1:1 to:
- symptomatic slow-acting drugs for osteoarthritis alone
- the same medical treatment plus orthovoltage low-dose RT, 4.5 Gy in 10 fractions
Pain, physical function, quality of life, and structural progression were evaluated during the original randomized study.
MRI structural progression was assessed using the Whole-Organ Magnetic Resonance Imaging Score through 36 months.
The later study extension examined government-certified disability and total knee arthroplasty after a median follow-up of 11.9 years.
Importantly, the long-term outcomes were exploratory extensions rather than endpoints specified in the original trial design.
Key results
Low-dose RT was associated with more durable improvement in pain and physical function than medical therapy alone.
Structural imaging showed the same direction of effect.
Change in overall WORMS score was:
- at 12 months: -0.3 with LDRT vs +0.9 with control, p=0.02
- at 36 months: +0.5 vs +2.3, p=0.01
The presentation also showed markedly less long-term radiographic progression among patients with baseline KL0-1 disease, with progression in approximately 3% after LDRT vs 36% with control in an exploratory analysis.
Long-term clinical outcomes were similarly favorable.
Government-certified disability occurred less frequently after LDRT:
- approximately 14% with LDRT vs 25% with control
- adjusted HR 0.33
- 95% CI 0.18-0.59
- p<0.001
For total knee arthroplasty across the overall analyzed population, the result favored LDRT but did not reach statistical significance:
- HR 0.51
- p=0.089
The strongest clinical signal was seen in patients with baseline KL2 disease.
In this subgroup:
- disability: 31.0% with LDRT vs 61.1% with control
- HR 0.37
- p=0.005
Total knee arthroplasty:
- 14.3% with LDRT vs 36.1% with control
- HR 0.32
- p=0.021
No acute radiation reactions or cancers within the treated area were reported during follow-up, although the study was not designed to quantify rare late effects or lifetime radiation-associated cancer risk.
Interpretation
The most important aspect of this study is the agreement between several very different outcome measures.
Pain improved.
Function improved.
Quality of life improved.
Blinded MRI showed less structural deterioration.
Long-term radiographic progression was lower.
Government-certified disability was reduced.
And among patients with KL2 disease, fewer patients ultimately underwent knee replacement.
Any one of these findings could be challenged in isolation. Patient-reported outcomes are vulnerable to expectation effects in an open-label study. Imaging has missing data. Long-term endpoints were added later. Subgroup analyses can produce unstable results.
But the investigators' strongest argument is that these independent measures all point in the same direction over more than a decade.
That makes a potential disease-modifying effect biologically and clinically more plausible than symptom relief alone.
Still, "potential" is essential.
This trial does not definitively establish that low-dose RT modifies osteoarthritis progression. The long-term disability and arthroplasty analyses were exploratory, the study was conducted at a single center, and the most striking arthroplasty result comes from the KL2 subgroup rather than the overall randomized population.
The overall total knee arthroplasty endpoint did not reach statistical significance.
So the result should not be framed as proof that radiation prevents knee replacement.
What it does provide is a remarkably consistent randomized signal strong enough to justify the next experiment: a larger multicenter, sham-controlled trial with prospectively defined structural and long-term clinical endpoints.
Limitations
The study was open-label and conducted at a single center, making symptomatic outcomes vulnerable to expectation effects and potentially limiting generalizability.
MRI follow-up had attrition, and only a subset contributed to long-term structural analyses.
The disability and knee arthroplasty outcomes came from a long-term study extension and were exploratory rather than pre-specified endpoints of the original randomized trial.
The most striking reduction in knee replacement was observed in the KL2 subgroup, which should be interpreted as a subgroup finding.
The overall total knee arthroplasty difference was not statistically significant.
The study population was relatively young for osteoarthritis, and contemporary treatment pathways may differ across healthcare systems.
Finally, although no cancers in the treated region were reported, the study was not powered or designed to determine rare late radiation effects or lifetime carcinogenic risk.