Why MRI sees more than an X-ray in early knee OA

A normal-looking knee X-ray can be genuinely misleading in the early stages of osteoarthritis. X-ray images bone well — joint-space narrowing, osteophytes, and gross structural collapse are all visible — but cartilage, the synovial lining, the subchondral bone marrow, and the menisci are effectively invisible on plain film. If the changes driving your pain sit in those tissues, an X-ray will not find them.

MRI changes that picture substantially. A single scan can simultaneously assess cartilage thickness and surface integrity, fluid and inflammation within the joint lining, bone marrow lesions in the subchondral zone, and the condition of both menisci. This matters in early-to-moderate OA because significant tissue damage can be present and symptomatic long before joint-space narrowing appears on X-ray.

The gap between the two investigations is wide enough that a consensus MRI-based definition of tibiofemoral OA now exists that does not require visible cartilage thinning at all — bone marrow lesions and synovitis on MRI can meet the criteria independently. Beyond standard structural sequences, compositional techniques such as T2 mapping and dGEMRIC can detect disruption to the cartilage's proteoglycan and collagen matrix before any morphological thinning is apparent, offering an earlier window into OA activity.

The practical effect is a shift in the clinical question. Rather than asking how much joint space remains, MRI asks which tissue is driving the problem — and that distinction directly shapes what treatment, if any, is likely to help. The sections below explain what each key finding means and how it informs injection decisions.

The four findings that matter most on a knee MRI

Four tissue findings recur consistently in early-to-moderate knee OA MRI reports. Each reflects a different pathological process, and each has different implications for treatment.

Cartilage changes

MRI shows cartilage surface integrity and thickness that X-ray cannot. Focal thinning, surface irregularity, and partial- or full-thickness defects all appear on standard sequences. In early disease the surface may still look intact even when the internal matrix is already breaking down — a limitation of standard sequences rather than a reassurance, and one reason the compositional techniques introduced in the previous section matter at the diagnostic stage.

Bone marrow lesions (BMLs)

These appear as bright patches in the subchondral bone on fluid-sensitive MRI sequences. Radiology reports often label them 'bone marrow oedema', but this is a misnomer that has persisted from earlier imaging literature. BMLs represent a combination of subchondral micro-fractures, localised inflammatory foci, and fibrosis — structural and reactive changes, not passive fluid accumulation. Their size or signal brightness does not reliably predict how much pain a patient will experience, which makes clinical correlation essential.

Synovitis and effusion

Formal definition requires gadolinium contrast — synovitis is characterised by above-normal post-contrast enhancement thicker than normal synovium. In routine practice, fluid-sensitive sequences can approximate this finding without contrast, identifying joint effusion and thickened synovial tissue that may be driving an inflammatory flare.

Meniscal pathology

Degenerative meniscal tears are common in OA and are not reliably distinguished from incidental findings without clinical context. Evidence from the Framingham study has shown that meniscal damage is a component of OA progression, not merely a bystander — meaning its presence on MRI can inform whether mechanical symptoms are contributing alongside inflammation.

Scoring the whole picture

MOAKS — the MRI Osteoarthritis Knee Score — is the most widely accepted framework for assessing all four domains systematically across each joint compartment. Rather than treating findings in isolation, MOAKS helps identify which tissue is contributing most, a distinction that shapes what intervention, if any, is targeted next.

Pain and imaging don't always match — and that matters

Reading an MRI report can feel alarming. Terms such as 'bone marrow oedema', 'chondral thinning', and 'degenerative meniscal change' appear in black and white, and the instinct is to equate them directly with the level of pain being experienced. The relationship is not that straightforward.

Asymptomatic MRI findings are common in the general population. Studies in people without knee complaints have identified cartilage irregularities, meniscal signal change, and subchondral abnormalities as incidental observations — structurally real but clinically silent. The reverse is equally true: some patients with significant, disabling knee pain show MRI changes that a radiologist might describe as mild or moderate. Neither scenario means the imaging is inaccurate; it means that imaging captures tissue structure, not the subjective experience of pain.

This is a clinically important principle rather than a caveat. A 55-year-old with incidental cartilage signal change on a scan ordered for an unrelated reason needs a very different conversation from a 55-year-old with an identical MRI appearance and six months of pain that limits daily walking. The findings are the same; the clinical significance is not. History, examination, and patient-reported outcomes carry equal weight alongside the scan.

Where MRI adds most value is when it helps explain a symptom pattern or identifies a specific structural driver that changes the treatment decision. That translation — from finding to clinical action — is where the next section picks up.

How MRI findings map to injection type

Matching the injection to the structural target identified on MRI is the central clinical logic — and it is what makes pre-injection imaging more than a formality.

When synovitis or effusion dominates

Corticosteroids remain a short-term first-line option where MRI shows synovitis or significant joint effusion as the primary finding. They act quickly on the inflammatory process and typically provide relief over weeks to around three months. The picture is complicated, however, by a 2025 study published in Radiology, which used WORMS whole-organ MRI scoring in Osteoarthritis Initiative participants and found that corticosteroid injections — including a single injection — were associated with significantly greater structural OA progression on MRI over two years compared with controls. The same study found hyaluronic acid was associated with decreased MRI-measured progression over the same period. This tension between short-term symptom relief and possible longer-term structural consequences is an active area of clinical discussion, not a settled question.

Where hyaluronic acid fits

HA targets the joint environment rather than a specific lesion. It does not directly address BMLs or cartilage defects, but evidence supports a symptom benefit and, as above, some imaging data suggests a potentially protective structural association — a consideration when weighing options alongside corticosteroids.

When BMLs are the dominant finding

Standard intra-articular injections — whether corticosteroid or HA — are considered inadequate when BMLs are the primary target. Neither agent can penetrate the subchondral zone sufficiently to reach the pathological tissue. Here, the delivery route matters as much as the injection type: PRP administered via a combined intra-articular and intraosseous (subchondral) approach has been reported in one series to result in complete MRI resolution of BMLs in 75% of cases at follow-up. Intra-articular PRP alone, by contrast, showed significant MRI improvement in patellofemoral cartilage volume and synovitis in a double-blind RCT of 46 knees at eight months (Raeissadat 2020), but subchondral bone marrow abnormality did not significantly respond — reinforcing that route specificity is not interchangeable.

Head-to-head trials directly comparing injection types in MRI-stratified patient groups remain limited. The matching logic described here reflects current clinical reasoning and evolving observational and RCT evidence, rather than fixed protocol.

The corticosteroid debate: short-term relief versus structural risk

The 2025 Radiology finding carries real clinical weight — but its study design matters for how it should be applied. The Osteoarthritis Initiative data is observational: participants who received corticosteroid injections were not randomly assigned to them, which means the group likely included patients with more severe baseline synovitis or more active disease than those who received hyaluronic acid or no injection. That indication bias — sicker joints getting corticosteroids — is difficult to fully disentangle from the MRI progression signal, even with statistical adjustment. The finding does not establish that corticosteroids cause structural damage; it establishes an association that warrants careful attention and replication in larger, stratified, randomised samples.

That said, the direction of the signal across the 70 OAI participants tracked over two years is consistent enough to shift clinical conversations. Injection frequency compounds the question: a single corticosteroid injection in a patient with an acute synovitis flare and no dominant BMLs carries a different risk calculus than repeated injections in a knee already showing structural progression on MRI.

For a patient with dominant synovitis, limited subchondral involvement, and a specific need for rapid pain relief — around a rehabilitation phase or planned return to activity — a single corticosteroid injection may remain a reasonable choice. What has changed is the framing around that decision. Rather than a reflexive first step, it becomes an option with a structural trade-off to weigh explicitly: how prominent is the synovitis? What does the rest of the scan show? Is there MRI evidence of existing progression? Those are questions an MRI can now help answer — and they belong in the consultation before the injection, not after it.

What to ask when your MRI report comes back

Armed with an MRI report, most patients find the follow-up consultation moves quickly. These questions are worth raising before any injection decision is made.

Which tissue is the primary driver? Cartilage, subchondral bone, synovium, and meniscus each point toward a different treatment pathway. Knowing which finding dominates on your scan changes what is appropriate to inject — and where.

Is the injection type matched to the finding — not just to the knee as a whole? An injection well suited to synovitis may be poorly suited to bone marrow lesions. The match between pathology and agent matters as much as the choice of agent itself.

Has the 2025 structural-progression data been factored into the recommendation? The observational evidence linking corticosteroid injections to greater MRI-measured OA progression over two years is not yet definitive, but it belongs in the conversation — particularly if repeated injections are being considered.

If bone marrow lesions are prominent, can an intra-articular injection actually reach the target? Standard intra-articular injections do not adequately penetrate the subchondral zone. The delivery route is a clinical question, not a detail to assume.

Was the MRI reported using a structured, whole-joint scoring framework? Standard radiology reports vary considerably in how systematically they cover each tissue compartment. A semi-quantitative framework such as MOAKS allows each domain — cartilage, bone marrow, synovium, meniscus — to be assessed in turn, making it easier to identify which finding should anchor the treatment plan. A specialist MSK review can place a standard radiology report into that clinical context where the original report does not.