The silent phase of cartilage loss

How do you know whether your knee is really getting worse, or just having a bad week? For many people, the answer arrives slowly — a dull ache after a long walk, stiffness that lingers into the morning, a sense that something has shifted even when scans come back unremarkable. That last detail matters more than it might seem.

Cartilage itself carries no pain nerve supply. The discomfort of early osteoarthritis is generated by the surrounding structures — the synovium, subchondral bone, joint capsule — not by the cartilage tissue losing its integrity. That biological fact creates a significant clinical gap: meaningful deterioration at the tissue level can be well underway before the signals that bring someone to a clinic actually begin.

Conventional MRI and X-ray are structural tests. They are very good at showing what has already happened — thinning, surface defects, joint space narrowing — but those changes are typically established by the time they become visible. A scan reported as normal in the early stages of degeneration is not the same as a healthy joint; it may simply mean the damage has not yet crossed the threshold that standard imaging can see.

This is the silent phase. The tissue is changing biochemically — collagen fibres losing their organised arrangement, proteoglycans beginning to deplete — but the architecture has not yet visibly collapsed. Evidence suggests this phase may, in some cases, still be reversible, which is precisely what makes it the most valuable window for intervention. T2 mapping MRI was developed to see into that window.

What T2 mapping actually measures inside your cartilage

The test works without any injection or contrast agent — the signal comes entirely from the tissue itself. During a T2 mapping sequence, a radiofrequency pulse briefly disturbs hydrogen protons in the water molecules within cartilage; T2 relaxation time measures how long those protons take to settle back to their resting state, expressed in milliseconds.

In healthy articular cartilage, the collagen fibre network is densely interwoven — think of tightly woven fabric, where the threads hold everything in position and leave little room for movement. That arrangement restricts water molecule motion, producing a low T2 reading, typically between 25 and 40 milliseconds.

When the collagen architecture begins to deteriorate, the effect resembles that fabric starting to fray: the threads loosen, gaps open up, and water moves more freely through the mesh. Free water accumulates, and the T2 value rises — usually above roughly 45 milliseconds — before any visible thinning or surface defect has appeared on a standard MRI scan.

Research adds an important nuance here. The rise in T2 is primarily driven by a specific signal component called R2a — the anisotropic part, which reflects how organised the collagen fibres are, rather than simply how much fluid is present. A 2025 study confirmed that R2a correlates directly with collagen network disruption in the deep cartilage zone, not with bulk water accumulation. T2 mapping is therefore tracking collagen structural integrity at a biochemical level — the number climbing is a sign the internal scaffolding is losing its ordered arrangement, long before that disorganisation becomes visible on conventional imaging.

How onMRI™ converts a standard MRI into reproducible data

Reading a T2 map is straightforward in principle; standardising that reading across different scanners, different radiologists, and different time points is considerably harder. A small difference in how cartilage is outlined on screen — or a different threshold for what counts as elevated signal — can produce meaningfully different numbers from the same underlying tissue. That inconsistency has historically limited the usefulness of quantitative cartilage imaging in routine clinical practice.

onMRI™ addresses this at the analysis layer. Developed at MSK Doctors and the London Cartilage Clinic by Professor Paul Lee, with Innovate UK Knowledge Transfer Partnership support and patent-pending status, it applies deep-learning algorithms to MRI data to produce standardised, reproducible outputs: cartilage and meniscus segmentation, automated T2 relaxation time mapping, three-dimensional joint reconstruction, and precise measurement of focal defect size and depth. The platform is also compatible with low-field Open MRI systems — a point returned to in more detail later.

For patients, the main practical benefit is consistency over time. When the same automated process is applied at each scan, a comparison six months later reflects what is genuinely happening inside the cartilage, rather than a difference in who read the image or how they traced the boundary. Research into deep-learning T2 acquisition suggests scan time can be reduced by around 40% without meaningful loss of quantitative accuracy, which helps make the process feasible in a standard clinical appointment rather than a specialist research setting.

It is worth being clear about the current evidence base: onMRI™ outputs to date are underpinned by a 150-scan cohort study covering cartilage, meniscal injury, and post-operative cases. Larger-scale validation linking imaging biomarkers to clinical outcomes and motion data is ongoing.

The evidence that T2 changes appear before symptoms do

The question of timing matters most: does T2 change before a patient feels anything, or only once pain has already set in?

A 2021 study by Alsayyad et al., examining the role of T2 mapping in routine MRI protocols, found that the technique detects early — and potentially reversible — cartilage damage before the onset of symptoms and before any morphological change is visible on standard imaging. That pairing — pre-symptomatic and pre-structural — is what gives the measurement its clinical significance.

The sensitivity improvement is substantial and concrete. For early cartilage lesions, standard MRI detects somewhere between 45 and 70 per cent of cases. Adding T2 mapping raises that figure to approximately 88–96 per cent. In 2013, Kijowski et al. published a study in Radiology — now cited more than 250 times — confirming that T2 maps added to a routine MR imaging protocol significantly improved sensitivity for detecting areas of cartilage damage that conventional sequences missed.

Sensitivity figures tell us T2 maps find more lesions — but they do not tell us whether those findings precede pain in real patients. Zhong et al. (2016) addressed that question directly: in people with no symptoms at all, variation in T2 signal predicted who would go on to develop symptomatic knee osteoarthritis. That makes T2 mapping prospective rather than merely concurrent — a signal that arrives ahead of the clinical picture, not alongside it.

More recent evidence reinforces this at scale. The 2025 Rotterdam Study, drawing on 1,332 knees, confirmed significantly elevated T2 values across all cartilage segments in patients with MRI-confirmed osteoarthritis, with the strongest associations in the lateral weight-bearing and posterior femoral regions — sites where load concentrates and early degeneration tends to take hold.

Why earlier detection changes your treatment options

Timing shapes what treatment is possible. When T2 mapping identifies elevated signal — collagen disorganising, proteoglycans depleting — but the cartilage surface remains structurally intact, the tissue is still present and still salvageable. That distinction has direct consequences for what can be offered.

A patient whose changes are caught at this earlier stage may be a candidate for therapies that require living cartilage to work with. Platelet-rich plasma and other biological injections deliver concentrated growth factors into the joint environment to support tissue health. Cartilage scaffold procedures — used where a focal defect exists — provide a structural matrix for the body's own cells to repopulate, effectively filling the gap before it widens. Weight optimisation and targeted activity modification also have meaningful impact at this stage, when offloading a damaged area can slow rather than merely delay decline. None of these approaches are universally suitable; defect size, depth, location, and the full clinical picture all determine what is appropriate for a given patient. But they are options that are foreclosed once structural collapse has occurred: once cartilage is lost, the decision tree compresses toward symptomatic management or joint replacement.

onMRI™'s precise defect sizing and depth measurements are where the scan result connects to that decision. A small, focal area of elevated T2 in an otherwise intact joint points toward one pathway; signal changes spread across multiple compartments point toward another. That stratification — difficult to make reliably from conventional imaging alone — is what makes earlier quantitative assessment clinically meaningful rather than academically interesting.

One important caveat: T2 values do not reveal why cartilage is degenerating or which osteoarthritis subtype is present. The imaging informs the decision — it does not replace the consultant's clinical assessment and full patient history.

Having a T2 map done at MSK Doctors

Whether T2 mapping adds anything useful depends on the individual clinical picture — which is why, at MSK Doctors, the decision is made in consultation rather than by booking a scan in advance. The Sleaford clinic in Lincolnshire is where this assessment takes place and where onMRI™ T2 analysis is run. Patients whose symptom pattern doesn't match what a conventional MRI shows, or whose cartilage changes are being tracked over time, are among those for whom quantitative mapping is most likely to alter the clinical picture. It is not offered as a routine add-on.

The Sleaford site uses an Open MRI scanner, which may be relevant for patients who have previously found standard closed-bore scanning difficult — whether due to claustrophobia, build, or anxiety about confined spaces. For patients based in London, the same analytical capability is available through the London Cartilage Clinic, the group's London arm.

Appointments can be arranged without a GP referral. The group is CQC-registered and rated 'Good' across all five inspection domains. Whatever the imaging shows, T2 results are interpreted by the consultant team alongside clinical history and examination findings — the quantitative biomarker is a starting point for a conversation, not a conclusion in itself.

Initial assessments can be booked directly at mskdoctors.com.

1. [1] T2 mapping as a biomarker for knee OA: Rotterdam Study (2025). (2025). https://doi.org/10.1016/j.joca.2025.09.009 https://doi.org/10.1016/j.joca.2025.09.009
2. [2] Asynchronous changes of articular cartilage and osteochondral junction in early knee OA (2025). (2025). https://doi.org/10.21037/qims-24-1492 https://doi.org/10.21037/qims-24-1492
3. [3] Rapid and robust quantitative cartilage assessment: deep learning-enhanced accelerated T2 mapping (2025). (2025). https://doi.org/10.1007/s00256-025-05034-w https://doi.org/10.1007/s00256-025-05034-w
4. [4] Orientation-Independent T2 Mapping Enhances MRI-Based Cartilage Characterization (2025). (2025). https://doi.org/10.1007/s10439-025-03774-3 https://doi.org/10.1007/s10439-025-03774-3