Why the 2–4 cm² defect range is the hardest to match

For a small focal cartilage defect in the knee — roughly the size of a fingernail, under 2 cm² — a single osteochondral plug will often settle the question. For a very large defect, above 4 cm², the evidence points toward cell-based repair or a donor allograft. Neither end of that spectrum demands much deliberation about technique.

The middle is different. Defects in the 2–4 cm² range — approximately the area of a grape — sit squarely within the licensed indications of two distinct techniques: osteochondral autograft transfer (OATS, sometimes performed as mosaicplasty) and autologous matrix-induced chondrogenesis (AMIC). Both are single-stage procedures, both aim to relieve pain and restore load-bearing function, and both have clinical evidence behind them at this size. The choice is a real one rather than a default.

What makes it genuinely difficult is that the two techniques produce different types of repair tissue with different durability profiles — and the 'better' tissue option carries its own trade-offs. Patient age, activity level, defect location, and the condition of the surrounding bone all feed into the decision, which is why two patients with identically sized lesions may reasonably be offered different approaches.

How OATS works and where it reaches its limit

Borrowed from a quieter corner of the same knee, an OATS graft is a cylindrical core of intact bone and cartilage — typically 8–10 mm in diameter — harvested from the peripheral femoral condyle, where load demands are low, and press-fitted into the prepared defect site. The transferred tissue is native hyaline cartilage: the same dense, low-friction material that lines a healthy joint. This matters because hyaline cartilage is stiffer, more fatigue-resistant, and better suited to the repetitive compressive forces the knee generates than the scar-like fibrocartilage produced by marrow stimulation alone.

A single plug covers defects up to roughly 1.5–2.5 cm². When the defect is larger, the surgeon must tile two or more smaller plugs side by side — the mosaicplasty variant. Coverage improves, but the spaces between adjacent plugs inevitably fill with fibrocartilage rather than hyaline tissue, gradually diluting the structural advantage that made OATS attractive in the first place. Technical demands also rise with plug count.

The more fundamental constraint is the donor site itself. Every plug harvested leaves a secondary wound in the same knee. For a small single-plug case that wound is minor; as defect size grows and more tissue must be taken, the harvest zone can become a source of persistent pain or mechanical irritation — a second problem created while solving the first. This caps how large a defect OATS can realistically address in one operation before donor-site burden outweighs the tissue-quality benefit.

Outcome data are reassuring within that envelope. Gudas et al. reported good or excellent results in 86–90% of patients at three years, and OATS has been shown to be superior to microfracture in athletes, with adequate rates of return to sport confirmed in published series.

How AMIC improves on standard microfracture

Plain microfracture — perforating the bone beneath the defect with an awl or drill — releases marrow-derived stem cells into the lesion, but the fibrin clot they form is mechanically fragile. Without structural support, the clot shifts, cell retention is poor, and the repair tissue that eventually forms is predominantly fibrocartilage. Kreuz et al. observed significant score deterioration between 18 and 36 months post-operatively, and Solheim et al. reported less than 60% survivorship at three years, with a mean time to failure of approximately four years — a durability ceiling that limits microfracture's role in the medium-defect range.

AMIC, first introduced by Behrens et al., addresses this directly. After standard microfracture, a bi-layer collagen I/III membrane — such as Chondro-Gide — is sutured or fibrin-glued over the prepared bed. The membrane acts as a biological tent: it stabilises the marrow clot, prevents its displacement under early joint movement, and creates a protected micro-environment that improves retention and chondrogenic differentiation of the progenitor cells drawn up through the perforations.

The repair tissue this produces is best described as fibro-hyaline — a step above the fibrocartilage of unsupported microfracture, but not equivalent to the native hyaline cartilage an OATS graft delivers. That distinction matters over the long term, and it is important to frame AMIC honestly as enhanced marrow stimulation rather than cartilage transplantation.

What AMIC gains in return is scalability. The membrane is cut to match the defect contour, so coverage is not constrained by how much autograft tissue a donor site can yield. This makes it well-suited to the 2–4 cm² medium range — and routinely used up to that ceiling without the harvest burden that mosaicplasty accumulates.

Gille et al.'s prospective registry of 131 patients (mean defect 3.3 ± 1.8 cm², mean age 36.6 years) provides the most robust clinical anchor: statistically significant improvements in Lysholm, KOOS, and VAS pain scores were recorded at one to two years and sustained at seven-year follow-up. Notably, neither age nor defect size significantly altered outcomes, suggesting the technique performs consistently across a realistic range of candidates.

What surgeons weigh when choosing between them

Defect size is the first filter, but within the medium range it is shape and compactness that matter most. A contained, roughly circular lesion of 1.8–2.2 cm² on the femoral condyle is well within what a single OATS plug — or at most two — can cover cleanly, and in that scenario the tissue-quality argument strongly favours autograft hyaline cartilage. The same nominal area spread across an irregular or elongated footprint makes geometric tiling difficult: more plugs are needed, inter-plug gaps accumulate fibrocartilage, and donor-site burden climbs. Once the defect approaches 3–4 cm² or has an awkward outline, AMIC's membrane — trimmed to fit any contour — becomes the more practical single-stage option.

Location adds a separate layer. Femoral condyle defects sit on a convex surface that is relatively forgiving for plug insertion and press-fit stability. Patellar and trochlear defects are geometrically less cooperative: the patella's posterior surface curves in two planes, and plug seating is technically demanding. Debieux et al. (2025) compared OATS and AMIC specifically in patellar defects and found that the OATS patients had a substantially higher preoperative Tegner score (6.59 versus 3.45), reflecting that surgeons were already directing higher-demand patients towards autograft at that site. The residual difference in activity scores at two years therefore partly mirrors who was selected for each technique rather than technique alone.

Bone involvement shifts the calculus further. An OATS plug restores both the cartilage layer and the underlying subchondral bone column — a genuine structural advantage when there is cystic change or bone loss beneath the defect. AMIC addresses the cartilage surface only; significant subchondral pathology may need separate management before or alongside scaffold implantation.

Patient age and physical demand consistently pull towards OATS when the defect size permits it. For an active patient under 50 who places high repetitive loads through the joint, the longer-term durability of hyaline cartilage outweighs the inconvenience of donor-site harvest, provided that harvest remains modest. Where BMI is elevated or there is varus or valgus malalignment, the repair tissue — hyaline or fibrocartilage — will be under disproportionate load regardless of technique, and a concurrent or staged osteotomy may be worth planning alongside either procedure.

What the evidence shows — and where the gaps remain

Taken together, the published literature confirms two things with reasonable confidence: each technique outperforms microfracture alone within its indicated size range, and each has accrued enough mid-to-long-term follow-up to support clinical use in the medium-defect window. What it cannot yet confirm is which technique produces better outcomes when the two are used for the same defect profile in the same patient population.

The evidence base for each comes from separate study streams rather than direct comparison. Gudas et al.'s long-running cohort — the most cited anchor for OATS durability — measured outcomes against microfracture, not against AMIC. Gille et al.'s seven-year registry data establish AMIC's sustained benefit in a cohort whose mean defect size sits squarely in the medium range, but again the comparator is baseline function, not a competing surgical arm. When Debieux et al. (2025) did compare the two techniques directly, the study was limited to patellar defects — a geometrically specific site — and the groups were not matched by pre-operative function at baseline, which constrains how far those findings can be generalised across all medium knee defects.

No large randomised controlled trial has directly compared OATS and AMIC for lesions in the 2–4 cm² window; most available comparative data sit at evidence Level 4 — registry and retrospective cohort studies. Long-term data also remain incomplete: AMIC follow-up beyond seven years is sparse, and arthroscopic OATS series with consistent ten-year reporting are limited.

That gap does not make the decision arbitrary. It reflects that cartilage repair matching depends on variables — defect geometry, bone involvement, patient demand, and donor-site tolerance — that resist standardisation across a single trial design. The existing evidence supports a framework for that matching; it does not yet provide a universal answer.

Getting the right assessment before any repair decision

Accurate defect characterisation is the foundation that any technique selection rests on. A decision reached on symptom pattern alone — without knowing exact defect size, depth, lesion margins, and subchondral bone status — risks matching the wrong procedure to the wrong problem, however sound the underlying reasoning about tissue type and donor-site tolerance may be.

High-quality MRI, with cartilage-specific sequencing rather than a standard joint protocol, provides that foundation. T2 mapping and compositional sequences can resolve detail that routine imaging misses — distinguishing a compact 2 cm² condylar lesion from an area of diffuse grade-II thinning that would fall outside the indication for either OATS or AMIC. AI-assisted cartilage segmentation (such as that available via onMRI™ analysis) can improve consistency of defect measurement across imaging planes, which matters when the size boundary between single-plug OATS and mosaicplasty — or between AMIC and a cell-based pathway — is close.

Functional assessment adds a second layer. How load distributes across the knee in daily movement and sport matters both to technique selection and to whether concurrent alignment correction should be planned; objective gait analysis (MAI Motion® is one validated tool for this) provides measurable data that standing X-ray alignment alone does not capture.

A consultant-led review can draw those findings together with the patient's activity demands and present a personalised options discussion covering OATS, AMIC, and — where defect profile or patient goals warrant it — adjacent pathways such as MACI or ChondroFiller injection. Lincolnshire Knee accepts patients without GP referral; consultations are available at Sleaford NG34 and Grantham NG31.