TWO ORIGINAL PAPERS ON VIRTUAL KNEE REPLACEMENT: DEVELOPMENT AND APPLICATION TO IMAGING BIOMARKER QUALIFICATION

The term virtual is now everywhere: Reality (VR) | Machine (VM) | Assistant (VA) | Network (VPN) | Currency (VC) and many more. But is “Virtual Knee Replacement” (vKR) a cheaper way of surgery without a surgeon, or a new therapy for late-stage osteoarthritis?

In their article, just published in Osteoarthritis and Cartilage, Kent Kwoh, Bob Boudreau, Felix Eckstein, Frank W. Roemer, M.D., Michael Hannon, Ali Guermazi and David Hunter (https://pubmed.ncbi.nlm.nih.gov/40780453/ )propose a so-called “multi-component” clinical endpoint of vKR as an outcome in clinical trials of (knee) osteoarthritis that is based on three components of patient-reported health status.

Why does this matter: Regulatory agencies currently favour surgical KR as an objective and “hard” clinical endpoint for approving disease-modifying osteoarthritis drugs (DMOADs), demonstrating clinical benefit. Only if a drug effectively reduces surgical KR rates compared with placebo participants, the structural benefit is considered clinically meaningful and the drug deemed effective.

However, the challenge is that only few patients progress to actually requiring surgical KR during the course of a trial. With 3 years of follow-up, sample sizes of 3-18,000 are needed. The cost of a pivotal trial (2 being required) reaches 0.5–1.5 billion USD, depending on drug costs, imaging sub-studies, etc.

How can vKR help? Many patients who qualify for surgical KR never receive it, because of:

• Comorbidities or contraindications that make surgery too risky
• Frailty in older, and concern about revision surgery in younger patients
• Fear of surgery, anesthesia, pain, or complications
• Belief that “I can get through this without surgery”
• Financial concerns or lack of adequate insurance
• Job-related worries
• Missing social support during rehabilitation
• Long waiting lists and surgery delayed
• Limited access to specialized surgical centers (e.g., rural areas)
•  Lack of trust in the health care system, etc.

In clinical trials, clinical status is commonly collected through patient-reported outcomes (PRO) from questionnaires. The current works shows that surgical KR can be predicted reliably, independent of whether surgery is ultimately performed or not. Combining vKR with surgical KR as an outcome in DMOAD trials can therefore substantially enhance the “signal”, reduce sample size, shorten observation times, and lower otherwise prohibitive study costs.

Yet, change in tissue structure vs. pain: Are these rivals on the run or partners in crime?
As we all know, there exists a well-described discordance between structure and symptoms in osteoarthritis.” How many times has the community tortured their keyboards and audiences with this phrase? But can we expect a 1:1 relationship between tissue change and pain? And how do we want to manage OA in the future. Are we willing to therapeutically target structural tissue pathology and natural progression in osteoarthritis, to eventually alleviate pain? Or resign ourselves to “joint anaesthesia,” with known side effects?

Our current application study of the PRO-based vKR endpoint in Osteoarthritis and Cartilage Open by Felix Eckstein, Wolfgang Wirth, Ali Guermazi, Frank W. Roemer, Michael Nevitt, Christoph Ladel, Leena Sharma, David Hunter and Kent Kwoh (https://pubmed.ncbi.nlm.nih.gov/40893583/  ) showed that participants with a high risk of reaching the vKR symptom status displayed substantially greater cartilage loss over 2 years prior to the event than those with a low risk (Fig., right),  revealing a longitudinal link between structural decline and symptom exacerbation.

And there is more to say in favour of those in the pro-structure camp:

Surgery works: In >80% of patients pain disappears after knee replacement (KR) Residual discomfort remains in <20%, but less than pre-op levels. And without a link between structure and symptoms – would the US spend ~$10 billion/year on KR? And we never actually compare structure with “pain” – but with “pain reporting”. Pain is subjective, noisy, culturally shaped, memory-biased, and variable. Try to recall your knee pain last month. That’s what patients must do in trials.

Between-knee, within-person study designs show that radiographic joint space narrowing (≈ cartilage loss) is related to (side differences in) knee pain and MRI has confirmed: Cartilage lesions and synovitis (but not bone marrow lesions [BMLs]) also predict side differences in pain. HTO surgery has been shown to greatly reduce medial cartilage loss and improves knee pain in parallel (NEWs here on July 29 2026))