What the preservation window actually means

The question most patients arrive with is some version of: 'Is it too late for anything other than a knee replacement?' It is a reasonable fear, and the honest answer is that it depends far less on age or pain level than on one structural question — how contained is the cartilage damage?

The preservation window describes the period when the joint retains enough structural integrity for restorative treatment: repair or regeneration of cartilage rather than wholesale joint replacement. That window is not defined by how many years someone has had knee pain, nor by how badly it hurts on any given day. It is defined by what the damage actually looks like inside the joint.

The clinical hinge point is a single distinction: is cartilage loss focal — a well-circumscribed defect in one compartment, surrounded by intact native cartilage — or is it diffuse, meaning widespread thinning across multiple compartments with structural changes throughout the joint? Surgeons and radiologists use this focal-versus-diffuse framework, not symptom intensity, to determine whether restorative options remain viable.

Two important corollaries follow. First, pain is an unreliable guide: mild discomfort can coexist with surprisingly advanced structural loss, while some patients with early but significant cartilage damage report little pain at all. Second, a standard X-ray can look near-normal even when MRI reveals meaningful cartilage, meniscal, or bone marrow changes — meaning plain films alone cannot confirm whether the window is open or closed. MRI is the tool that operationalises the answer.

MRI features that put you inside the window

Scan reports can feel opaque, but a handful of specific features — when read together — give a radiologist or surgeon a confident picture of whether damage is contained or beginning to spread.

The clearest signal of a preservable defect is its geometry. On MRI, a focal lesion looks like a pothole in an otherwise sound road: the edges are sharp and well-defined, the damage is confined to one compartment (medial tibiofemoral, lateral tibiofemoral, or patellofemoral), and the cartilage surrounding it remains intact. That surrounding healthy tissue matters enormously — it provides the biological scaffolding that restorative treatments rely on. When remaining cartilage forms an island rather than a narrow fringe, the structural environment supports repair.

Bone marrow lesions (BMLs) — the bright signal changes seen in bone — are expected directly beneath an active cartilage defect, and their presence there is not itself alarming. What changes the picture is when BMLs appear across multiple sub-regions or in compartments away from the primary defect; that pattern suggests more widespread stress on the joint and nudges the assessment toward diffuse disease.

Osteophytes — bony spurs at the joint margins — are another useful indicator. Their absence or near-absence reflects limited subchondral remodelling and is consistent with a joint that has not yet mounted a widespread adaptive response to load redistribution.

Finally, the meniscus: a locally torn or thinned meniscus at the defect site is compatible with preservation candidacy, but diffuse degeneration or extrusion of the meniscus body signals joint-wide structural failure rather than a contained problem.

No single feature is read in isolation. Specialists weigh all of them together — along with lesion size, which serves as a guide for procedure selection rather than a hard threshold for candidacy.

MRI signs the window is closing — or already closed

Several MRI features, individually unremarkable, collectively draw a clear line between a closing and a closed preservation window.

Widespread cartilage thinning is the first and most definitive departure from the focal pattern. Rather than a discrete defect surrounded by healthy tissue, the scan shows irregular loss across more than one compartment — the signal that the joint as a whole, not a single zone, is under sustained mechanical stress.

Multi-compartmental bone marrow lesions (BMLs) reinforce that reading. BMLs confined beneath a single defect are expected and not inherently alarming; BMLs appearing across multiple subregions or away from the primary lesion indicate that stress is being distributed widely enough to drive bone-level change throughout the joint.

Subchondral sclerosis and large cysts tell a different part of the same story: structural remodelling has penetrated below the cartilage layer into the underlying bone. Large osteophytes signal the same trajectory — they are a mechanical adaptation to prolonged load redistribution and a marker of longstanding rather than early disease.

Diffuse meniscal degeneration or extrusion beyond the joint line represents the loss of the knee's primary load-distribution architecture. A locally torn meniscus at a focal defect is compatible with preservation candidacy; diffuse meniscal failure is not.

The single sharpest radiological contraindication is bipolar ('kissing') involvement — cartilage damage on opposing articular surfaces within the same compartment. Published outcome data on this pattern, examined in a later section, show substantially higher failure rates than unipolar restoration; the finding alone commonly shifts the clinical conversation away from restorative biology toward symptom management and, in time, joint-replacement planning — not because nothing can be done, but because restorative approaches cannot outpace deterioration that has spread across the entire joint surface.

How radiologists grade cartilage loss — reading your own report

MRI reports often land in patients' inboxes as a string of abbreviated grades and anatomical sub-regions that seem to require a medical dictionary. Two scoring systems account for most of the language you are likely to encounter.

Cartilage defect depth grading

This system describes how far through the cartilage a lesion penetrates. Grade 3 means the defect extends more than 50% of the cartilage's full thickness — with sub-grades (A through D) capturing whether it reaches or breaches the calcified layer immediately above the subchondral bone. Grade 4 means subchondral bone is exposed: the cartilage has been lost entirely at that spot. Grade 3–4 defects are the typical target for restorative procedures; the depth tells a surgeon what tissue environment they are working into and whether an intact bone plate can anchor a repair.

MOAKS area grading

Where depth grading describes how deep, MOAKS describes how much of each sub-region is affected: Grade 0 = no loss; Grade 1 = less than 10% of the sub-region; Grade 2 = 10–75%; Grade 3 = more than 75%, with a separate flag for full-thickness loss. High area grades across several sub-regions produce a very different clinical picture from a single elevated grade in one location — the former is a diffuse-OA pattern, the latter a focal-defect pattern.

Reading the language in practice

A report describing a 'Grade 3–4 focal chondral defect, medial femoral condyle, MOAKS area grade 1' reads as a contained, deep lesion in a limited zone — consistent with preservation candidacy. 'Multi-sub-regional Grade 2–3 full-thickness loss' describes spread that shifts the picture considerably.

One important caveat: both systems involve qualitative judgement, and the same scan can be graded differently depending on the reader's familiarity with cartilage pathology. Depth alone does not determine candidacy, and no single grade substitutes for specialist interpretation of the whole-joint picture.

Why X-rays can mislead — and when MRI changes everything

Plain films catch what bone does — narrowed joint space, sclerosis, osteophytes. What they miss, often entirely, is what cartilage, menisci, and bone marrow are doing long before those bony changes appear. A Kellgren–Lawrence grade 0 or I X-ray is therefore compatible with either a well-preserved joint or meaningful MRI-detectable cartilage loss, bone marrow lesions, or meniscal pathology — two patients with identical films can sit at very different points relative to the preservation window.

The corollary matters equally. Mild MRI abnormalities are not automatically the pain driver. Asymptomatic structural changes are common, and a 2025 OARSI review identified several conditions — meniscal injury, patellofemoral pain, crystal arthritis, and immune arthritis — that can mimic early OA on imaging. A scan showing cartilage signal change does not, on its own, confirm OA as the cause of a patient's symptoms. That determination integrates history, clinical examination, and imaging together; one without the other two leaves the picture incomplete.

MRI earns its place most clearly when X-rays look reassuringly normal yet symptoms or examination point toward focal pathology — the gap between what plain film shows and what the joint has actually done.

At the research frontier, compositional MRI techniques — dGEMRIC, T2 mapping, and T1rho — can detect proteoglycan depletion and collagen disruption before morphological defects appear, and some studies suggest they may predict future cartilage loss. Standardisation across centres remains unresolved, however, and these techniques are currently largely confined to research settings rather than routine clinical assessment.

Gray zones, borderline cases, and what to do next

Candidacy is rarely settled by a single MRI feature. A machine learning analysis of 1,091 cartilage preservation procedures identified the dominant failure predictors as symptom duration, age, BMI, lesion grade, total lesion area, and number of lesions — a reminder that prognosis emerges from the full clinical picture, not from any individual grading score. A relatively contained defect in a younger, lighter patient with a short symptom history carries a different prognosis from the same MRI findings in someone who has been symptomatic for years.

Two structural patterns sit in genuine gray zones. Bipolar involvement — damage to opposing articular surfaces — carries a substantially less predictable prognosis after restoration than a unipolar lesion, and the evidence base for these cases remains limited; outcomes vary considerably across published series. Patellofemoral involvement presents a different complication: it independently worsens pain and functional scores beyond what tibiofemoral loss alone would predict, meaning a scan that looks contained in one compartment may understate the overall burden on the joint.

For patients whose stage feels uncertain, the most useful next step is a specialist assessment that combines MRI review, clinical examination, and symptom history — not a scan report read in isolation. Concrete questions worth raising include how long symptoms have been present, whether the damage involves opposing surfaces, and what conservative measures have already been tried. For those confirmed to be inside the preservation window with a focal, contained defect, restorative options — including injectable scaffold approaches such as ChondroFiller injection, available at the London Cartilage Clinic on Harley Street (londoncartilage.com) — can be discussed in that same consultation.

The structural question this article opened with rarely has a clean, single-variable answer. But the evidence is consistent that it has an answerable one — and that the answer becomes considerably more precise, and more actionable, when imaging, clinical findings, and symptom history are read together by someone who does this regularly.

1. [1] Evidence-Based Machine Learning Algorithm to Predict Failure Following Cartilage Preservation Procedures in the Knee (2023). (2023). https://doi.org/10.1177/2325967123s00019 https://doi.org/10.1177/2325967123s00019