What OATS and mosaicplasty actually do

Unlike most cartilage procedures, OATS and mosaicplasty do not ask the body to grow replacement tissue from scratch. Instead, the surgeon harvests one or more cylindrical plugs of your own bone and cartilage from a quieter, lower-load area of the same knee — typically the peripheral edge of the femoral condyle — and press-fits them directly into the damaged zone.

The two techniques share that core principle but differ in how many plugs are used. OATS transfers a single, larger cylinder, usually 8–10 mm in diameter, making it well suited to compact focal defects. Mosaicplasty — introduced into clinical practice in 1992 — tiles several smaller plugs side by side in a mosaic arrangement to cover a wider area.

Critically, both are single-stage procedures: harvest and implantation happen in the same operation, with no second surgery for cell culture or laboratory processing. The tissue transferred is genuine hyaline cartilage — the same smooth, load-bearing surface that lines a healthy knee — rather than the fibrocartilage scar tissue produced by microfracture. That distinction matters for durability: fibrocartilage is structurally inferior and tends to break down over time, whereas hyaline cartilage is built to last under repeated joint loading.

The typical candidate is a young, active patient with a focal, full-thickness cartilage defect — ICRS grade III or IV — who has not improved with conservative management such as physiotherapy or injection therapy.

Defect size: the primary filter for surgical choice

Defect area drives the choice between OATS, mosaicplasty, and the alternatives more reliably than any other single variable.

For lesions smaller than approximately 2 cm², OATS — transferring one well-sized plug — is usually sufficient to resurface the damaged zone in a single pass. When a defect falls between roughly 2 cm² and 4 cm², mosaicplasty becomes the more appropriate tool: tiling two or more smaller plugs extends coverage without requiring a second-stage procedure or allograft tissue. The 2021 Cartilage Book algorithm formally positions mosaicplasty as a recognised alternative to microfracture across this window. The key reason is durability: mosaicplasty delivers native hyaline cartilage, whereas microfracture produces structurally inferior fibrocartilage with a tendency to deteriorate within two to three years.

A 2024 institutional registry of 63 patients followed for up to ten years illustrates what this looks like in practice. The mean treated lesion measured 2.3 cm² and required an average of 2.2 plugs — comfortably within the autograft-viable zone. Mean IKDC scores improved from 46.4 before surgery to 70.4 at ten-year follow-up, and only two patients required conversion to joint replacement over the entire study period.

Once a defect reaches 3–4 cm², the calculus shifts. Evidence from the SUMMIT trial supports MACI or ACI as the better primary strategy at this size, partly because the volume of available donor autograft begins to limit how much of the defect can be reliably filled. Defects larger than approximately 4 cm² typically require osteochondral allograft transplantation instead. The precise upper boundary for mosaicplasty remains a matter of clinical judgement rather than a fixed rule, and reported thresholds vary across published series.

Where in the knee — how location changes the decision

Location determines not just whether OATS or mosaicplasty applies, but how the operation is delivered and how robust the supporting evidence actually is.

The medial femoral condyle (MFC) is by far the most common treatment site. A systematic review of 1,139 patients across 24 studies found the MFC accounted for 75.4% of all OAT cases — it is the canonical indication, and the weight of long-term outcome data comes from this location. The lateral femoral condyle (LFC) follows at 12.1%. Both are accessible arthroscopically, meaning the surgeon can work through small portal incisions rather than opening the knee fully.

Patellar lesions (6.7% of cases) and trochlear lesions (5.7%) present a different picture. Neither location is reachable through a standard arthroscopic approach; both require an open arthrotomy, which adds surgical complexity and typically a longer recovery. The same systematic review noted that open-approach cases treated defects roughly three times larger on average than arthroscopic cases (mean 2.96 cm² versus 0.97 cm²) — relevant context when surgeons and patients discuss what the procedure will involve.

Focal defects in the weight-bearing zones of the femoral condyle carry the strongest evidence base and the most straightforward surgical access. Patellar OATS is technically feasible — one small series in adolescents reported good outcomes with 8–10 mm plugs — but patellar-specific outcome data remains considerably thinner than for condylar cases, and individual suitability requires careful assessment.

What MRI shows — and where it falls short for surgical planning

A scan reporting a 1.5 cm² defect is a starting point, not a final measurement. Published accuracy studies confirm that preoperative MRI systematically underestimates cartilage defect size when compared with what the surgeon finds at arthroscopy — meaning the true area, measured after the damaged tissue has been debrided back to a clean margin, is consistently larger than imaging suggests.

Interrater reliability for MRI-based sizing is acceptable: independent reviewers assessing the same scan show good agreement (ICC 0.74–0.78 for full-thickness defect area). Surgeons looking at the same images broadly concur. The limitation is not disagreement between reviewers — it is that all MRI measurements share a common tendency to fall short of the post-debridement intraoperative finding, which remains the operative reference standard.

For practical planning, this matters in two ways. First, the number of plugs and the donor-site preparation may need to be adjusted once the knee is open and the real defect boundary is visible. Second, a defect that appears comfortably within the single-plug OATS range on imaging could cross into mosaicplasty territory once debrided.

AI-assisted MRI analysis — such as onMRI™ cartilage segmentation and T2 mapping — can improve characterisation of defect margins and cartilage thickness at the planning stage, helping the surgical team build a more informed estimate before the procedure. Intraoperative confirmation remains necessary regardless. The surgical plan should be treated as flexible, with the MRI figure as a lower bound rather than a fixed target.

Alignment, donor-site limits, and when a different approach fits better

The technique has clear decision boundaries — understanding them explains why two patients with similar symptoms may be offered different operations.

Alignment comes first. A knee running in varus or valgus alignment concentrates mechanical load on one compartment. Transplanting cartilage plugs into an overloaded zone without correcting that load is likely to accelerate graft failure. When malalignment and a focal lesion coexist, a concurrent osteotomy — typically an opening-wedge high tibial osteotomy (OWHTO) — is required to redistribute load before the repair can integrate. A 2025 study combining OWHTO with OAT in 30 patients with varus knees reported improvement across all KOOS subscale scores and ICRS-CRA Grade 1–2 cartilage repair on second-look arthroscopy at 12 months, supporting the principle that alignment and cartilage repair must be addressed together.

The donor-site ceiling. Only a finite number of plugs can be harvested from non-weight-bearing zones of the patient's own knee before a secondary problem develops at the harvest site. Once that ceiling is approached, collecting more graft becomes counterproductive rather than beneficial.

For defects approaching or exceeding 3–4 cm², this ceiling becomes the binding constraint. The SUMMIT trial showed MACI produced improved KOOS pain and function scores over microfracture at 2 and 5 years in larger defects, and a two-year randomised trial comparing third-generation ACI with mosaicplasty provides direct head-to-head evidence that cell-based repair is a viable next step when autograft volume is insufficient. A hybrid strategy combining ACI with OATS — part plug, part implanted cells — has been described for defects that sit at the autograft boundary. For very large or post-traumatic defects where autograft is wholly inadequate, osteochondral allograft transplantation (OCA) removes the donor-site constraint by using cadaveric graft tissue.

Microfracture is not a reliable fallback: fibrocartilage breakdown within two to three years and damage to the subchondral bone plate are well-documented, and its use has declined accordingly.

Long-term outcomes and the assessment pathway at Lincolnshire Knee

Durability data from published registry follow-up reinforces the case for well-selected autograft repair. A 2024 series tracking 63 patients over up to ten years reported a reoperation rate of 28.6% — the majority for minor procedures such as chondroplasty or loose body removal rather than revision cartilage surgery — with KOS-ADL scores rising from 64.4 to 83.8 and only two patients requiring conversion to arthroplasty. Those figures reflect what the preceding sections describe: small, contained defects at the medial femoral condyle, in a mechanically sound knee with correct alignment, carry a durable prognosis.

Patient selection remains the dominant predictor of success. The questions worth raising at a pre-surgical consultation are practical: Is the defect genuinely focal and contained? Is it within the size range where autograft coverage will be complete? Does alignment need correcting first? Answering them accurately before the operation matters as much as surgical technique.

At Lincolnshire Knee, that planning process is supported by onMRI™ AI cartilage analysis — providing quantitative defect characterisation and T2 mapping — and MAI Motion® biomechanical assessment for objective load-distribution data, both available at the Sleaford NG34 site. Lincolnshire Knee is part of the MSK Doctors group and accepts patients without referral. Book an assessment at lincolnshireknee.co.uk.

1. [1] Clinical experiences with cartilage repair techniques: outcomes, indications, contraindications and rehabilitation. (2015). https://doi.org/10.5606/ehc.2015.19 https://doi.org/10.5606/ehc.2015.19
2. [2] Combination of autologous chondrocyte implantation (ACI) and osteochondral autograft transfer system (OATS) for surgical repair of larger cartilage defects of the knee joint. (2015). https://doi.org/10.3233/THC-151003 https://doi.org/10.3233/THC-151003
3. [3] Arthroscopic versus Open Osteochondral Autograft Transplantation (Mosaicplasty) for Cartilage Damage of the Knee: A Systematic Review. (2019). https://doi.org/10.1055/s-0039-1692999 https://doi.org/10.1055/s-0039-1692999
4. [4] Comparison of Knee Articular Cartilage Defect Size Between Measurements Obtained on Preoperative MRI Versus During Arthrotomy. (2023). https://doi.org/10.1177/23259671231193380 https://doi.org/10.1177/23259671231193380
5. [5] Mosaicplasty/Osteochondral Autograft Transfer Remains a Durable Solution for Symptomatic Chondral Defects of the Knee: Two to Ten-Year Follow-up Analysis. (2024). https://doi.org/10.1177/2325967124s00003 https://doi.org/10.1177/2325967124s00003
6. [6] Combined high tibial osteotomy and osteochondral autograft transplantation for medial osteoarthritis with concomitant lateral femoral condyle cartilage lesions. (2025). https://doi.org/10.1016/j.knee.2025.07.017 https://doi.org/10.1016/j.knee.2025.07.017
7. [7] Commentary on "Third-generation autologous chondrocyte implantation versus mosaicplasty for knee cartilage injury: 2-year randomized trial". (2016). https://doi.org/10.1002/jor.23206 https://doi.org/10.1002/jor.23206
8. [8] Preoperative MRI Underestimates Articular Cartilage Defect Size Compared With Findings at Arthroscopic Knee Surgery. (2013). https://doi.org/10.1177/0363546512472044 https://doi.org/10.1177/0363546512472044