Yes — mild MRI findings and significant pain can coexist

If your scan came back 'mild' or 'only minor changes' and yet you are still struggling to walk downstairs, sleep comfortably, or get through a working day, you are not imagining it — and you are not alone.

A large 2025 study published in the Journal of the American Academy of Orthopaedic Surgeons examined nearly 4,800 patients from the Osteoarthritis Initiative dataset and tested how well MRI scores could predict knee pain. Even the most comprehensive MRI grading system — a detailed scoring tool called MOAKS — explained only 28% of the variation in patient-reported pain. Standard radiographic grading did worse still, accounting for just 15%. Crucially, predictive accuracy actually fell as symptoms became more severe: the patients in most pain were the ones the imaging was least able to explain.

That finding has a practical implication. A 'mild' MRI report describes the cartilage surface — it does not describe a patient's pain experience, and it should never be used to dismiss one. The reason lies in where knee pain actually originates, and in most cases it is not the cartilage surface itself.

Cartilage has no pain fibres — so where does the pain come from?

Articular cartilage — the smooth, glassy tissue lining the ends of bones inside the knee — contains no blood vessels and no nerve fibres. This means it cannot generate a pain signal. By itself, a cartilage lesion is neurologically silent.

The tissue that does feel pain is the subchondral bone immediately beneath: a dense, well-innervated layer packed with nociceptive nerve endings. When load or mechanical stress transmits through damaged or thinning cartilage, it reaches this bone layer — and that is where the signal originates.

A 2025 study from the Bunkyo Health Study, involving 1,145 participants, put this into numbers. Cartilage damage alone — without any concurrent changes in the subchondral bone — carried no statistically significant association with knee pain (odds ratio 1.10; confidence interval crossing 1.0). Pain risk rose substantially only when bone marrow lesions co-existed with subchondral bone attrition (OR 2.22) or with subchondral bone cysts (OR 1.79).

Think of it like the rubber sole on a shoe: worn rubber is visible from the outside, but the bruising and tenderness happen in the foot beneath. Standard MRI grades the cartilage surface — it does not reliably capture early subchondral stress, which may already be generating significant symptoms long before obvious morphological thinning appears on the scan.

Synovitis and peri-articular inflammation also drive pain independently

The synovium — the thin membrane lining the inside of the knee — is densely supplied with nerve endings and highly reactive to both mechanical and chemical irritation. When cartilage is damaged, even at a mild grade, it can release inflammatory mediators that trigger the synovium into an active pain state of its own.

FNIH Biomarkers Consortium data quantify this: synovitis accounts for roughly 18–22% of the association between bone marrow lesions and knee pain, and carries its own independent pain signal (β=0.40) — contributing to symptoms beyond what subchondral stress alone can explain.

Tissues outside the joint capsule add another layer. A 2025 British Journal of Orthopaedics study of 204 early-OA knees found that irritation in the tendons running around the front and inner side of the knee had a substantially larger effect on reported pain (effect sizes of η²=0.113 and η²=0.033 respectively) than the actual extent of cartilage loss (η²=0.023). These peri-articular changes do not appear in a report focused on cartilage grading alone.

A single cartilage grade is therefore a narrow window into what is often a multi-tissue problem. The practical implication is not alarming — it is useful: a clinical assessment that examines the joint lining, surrounding tendons, and soft tissues alongside imaging gives far more actionable information than a scan report reviewed in isolation.

Why two people with the same scan can have very different pain levels

Two people can sit in the same consulting room with near-identical MRI reports and describe completely different lives. One runs at weekends with only occasional stiffness; the other cannot complete a supermarket shop. The difference often lies not in the cartilage but in how the nervous system is processing pain signals.

Central sensitisation is a well-characterised neurological phenomenon — not a psychological one — in which the brain and spinal cord become progressively more responsive to pain input, amplifying signals out of proportion to the underlying tissue state. The validated CAP-Knee questionnaire, developed specifically for knee pain, identifies eight measurable domains of this amplification: neuropathic-like pain, sleep disturbance, fatigue, anxiety, catastrophising, cognitive impact, depression, and widespread pain distribution. High scores on these domains are reliably associated with worse pain intensity, regardless of cartilage grade.

The bidirectional nature of this disconnect is worth noting. Asymptomatic collegiate basketball players and adolescent footballers have been documented with MRI cartilage changes and no pain whatsoever — confirming that structural findings are neither necessary nor sufficient to produce symptoms. At the other end, a 2024 trial of PRP injections recorded a dramatic fall in pain scores (from a VAS of 7 down to 2.76 at twelve months) with no statistically significant change in MRI cartilage thickness. Symptoms improved substantially; the scan barely moved.

This matters clinically because treating the structural image alone — without accounting for how the nervous system is processing the pain — is unlikely to be enough for many patients. A thorough assessment needs to map the full picture: the joint, the surrounding tissues, and the pain-processing system that links them.

What a thorough assessment actually looks at

A useful diagnostic workup for discordant knee pain — significant symptoms with mild imaging — goes considerably further than reading the cartilage grade off the radiologist's summary line.

The structural layer comes first: reviewing not just cartilage grading but subchondral bone changes (bone marrow lesions, cysts, attrition), synovial status, and peri-articular tendon and bursa condition — each capable of independently driving pain, as the preceding sections show. Standard MRI, however, has well-documented limits in this context. It is highly specific but poorly sensitive for early or superficial cartilage lesions; arthroscopy remains the structural gold standard. Quantitative MRI sequences — particularly T2 mapping — go a step further by detecting biochemical disruption to the cartilage's collagen matrix before any thinning is visible on conventional sequences, meaning a 'mild' standard scan may underrepresent the biological picture.

The second layer is biomechanical. How a patient loads the knee — walking pattern, hip and ankle compensation, timing and activation of the quadriceps — generates forces on the joint that no scan can capture. Objective clinical movement assessment, and where warranted a markerless motion analysis tracking joint loading angles and muscle-activation timing, reveals contributors invisible to imaging.

The third layer draws on clinical history to map central pain-processing patterns.

Taken together, this produces a diagnosis that explains the pain rather than simply grading the cartilage. Treatment selection follows from that fuller picture — not from the MRI report alone.

Treatment options once the real pain source is identified

Treatment selection only makes sense once the dominant pain source has been identified. Targeting a cartilage surface when pain is driven by subchondral bone inflammation, synovitis, or a sensitised nervous system rarely delivers lasting relief.

For the joint environment, options that reduce synovitis or alter the biological milieu — including biologic injections such as platelet-rich plasma — can produce substantial pain reduction without any structural MRI change, as the 2024 PRP trial described earlier demonstrated. The relief comes through biological and anti-inflammatory pathways, not cartilage regeneration.

Where a focal cartilage defect is genuinely part of the pain picture, less invasive repair options have expanded. ChondroFiller, a collagen-scaffold injectable, is designed for focal defects without the recovery burden of open surgery. NanoACi is a needle-delivered cartilage repair approach that removes the need for arthroscopy in the delivery pathway for suitable patients; it aims to support a chondrogenic repair environment rather than guarantee structural regrowth, and suitability depends on defect pattern, subchondral bone status, and overall joint condition.

A 2026 study makes the structural-versus-symptom distinction concrete: adding microfracture to high tibial osteotomy produced significantly better cartilage repair on second-look arthroscopy (62.5% vs 33.3%), yet two-year patient-reported outcome scores were no different between groups. Better cartilage on the scope; no better life for the patient.

The right intervention, then, is determined by which pain mechanisms are dominant — subchondral, synovial, peri-articular, or central — not by MRI grade alone. MSK Doctors consultants assess all of these layers without referral; bookings are available at mskdoctors.com, and London-based patients can access the same specialist pathway through the London Cartilage Clinic. That kind of diagnosis — one that maps the full picture rather than grades the cartilage — is the prerequisite for any treatment option that is likely to help.

1. [1] The Discordance Between Pain and Imaging in Knee Osteoarthritis. (2025). https://doi.org/10.5435/JAAOS-D-24-00509 https://doi.org/10.5435/JAAOS-D-24-00509
2. [2] MRI-based cartilage changes and clinical effectiveness of autologous intra-articular platelet-rich plasma injections in symptomatic patients with moderate osteoarthritis. (2024). https://doi.org/10.1186/s43055-024-01203-4 https://doi.org/10.1186/s43055-024-01203-4
3. [3] Bone marrow lesion coexisted with subchondral bone attrition and/or subchondral bone cyst is associated with knee pain regardless of cartilage lesion. (2025). https://doi.org/10.1186/s13075-025-03644-2 https://doi.org/10.1186/s13075-025-03644-2
4. [4] The Central Aspects of Pain in the Knee (CAP-Knee) questionnaire; a mixed-methods study. (2021). https://doi.org/10.1016/j.joca.2021.02.562 https://doi.org/10.1016/j.joca.2021.02.562
5. [5] Evaluation for Cartilage Lesions on MRI Continues to Improve: Artificial Intelligence Applications May Result in Higher Sensitivity and Specificity. (2024). https://doi.org/10.1016/j.arthro.2024.03.009 https://doi.org/10.1016/j.arthro.2024.03.009
6. [6] Cartilage lesions are not the main factor influencing pain and functional impairment in early knee osteoarthritis. (2025). https://doi.org/10.1302/2633-1462.67.BJO-2025-0010.R1 https://doi.org/10.1302/2633-1462.67.BJO-2025-0010.R1
7. [7] Structural Gains Without Symptomatic Benefit After Microfracture During High Tibial Osteotomy. (2026). https://doi.org/10.1177/23259671261455795 https://doi.org/10.1177/23259671261455795
8. [8] Analysis of multiple MRI-based quantitative structural measurements of knee osteoarthritis in a case control study — association with pain and structural progression. (2025). https://doi.org/10.1016/j.ocarto.2025.100638 https://doi.org/10.1016/j.ocarto.2025.100638
9. [9] Synovitis mediates the association between bone marrow lesions and knee pain in osteoarthritis: data from the FNIH Osteoarthritis Biomarkers Consortium. (2022). https://doi.org/10.1016/j.joca.2022.02.507 https://doi.org/10.1016/j.joca.2022.02.507
10. [10] Quantifying Knee-Adjacent Subcutaneous Fat in OAI — Associations with Cartilage MRI T2, Thickness and Pain. (2025). https://doi.org/10.1016/j.joca.2025.01.001 https://doi.org/10.1016/j.joca.2025.01.001
11. [11] The Large Focal Isolated Chondral Lesion. (2021). https://doi.org/10.1055/s-0041-1735278 https://doi.org/10.1055/s-0041-1735278