Defect size as the primary decision filter

Surface area of the cartilage lesion is the single strongest predictor of which procedure is appropriate — more so than age, activity level, or symptom duration at the initial decision point.

Under 2 cm² — OATS or mosaicplasty is the standard approach. One or two cylindrical plugs harvested from a non-weight-bearing zone of the knee cover the defect cleanly, restoring both the articular surface and the underlying bone in a single operation. The harvest sites are small, donor-site discomfort is manageable, and the procedure is completed in one stage.

2–4 cm² — this is a genuine grey zone in the evidence. Published data, including a widely cited 2015 review (Richter et al., PMC4789925), indicate that OAT and ACI/MACI produce clinically equivalent outcomes across this range. Neither option has a clear superiority, so surgeon experience, joint access, and individual patient factors tend to drive the decision. If your consultant discusses either route at this size, that reflects the evidence honestly rather than indecision.

Above 4 cm² — MACI becomes the preferred choice. Attempting to fill a large defect with multiple OATS plugs creates two compounding problems: the gaps between plugs heal with fibrocartilage rather than hyaline cartilage, and harvesting enough plugs causes meaningful pain and structural weakness at the donor site. MACI sidesteps both issues; the collagen membrane can be cut to fit defects up to approximately 20 cm², with no structural upper limit on the membrane itself.

Microfracture, the historical benchmark for smaller lesions, is worth a brief note. Current evidence shows that the fibrocartilage it produces tends to break down within two to three years, and the marrow-stimulation process can damage the subchondral bone plate — compromising options for future repair. It is no longer considered a first-line modern choice.

When subchondral bone loss changes the picture

Bone involvement is the second fork — and it can override the size algorithm entirely.

OATS transfers a cylindrical plug that contains both the articular cartilage surface and the column of bone beneath it. When a lesion has eroded through to the subchondral layer — ICRS Grade 4, meaning the damage extends through the full cartilage thickness into bone — OATS restores both tissues in the same step. Osteochondritis dissecans (OCD), a condition in which a fragment of bone and cartilage partially or fully detaches from the joint surface, is a classic example: because the underlying bone is already compromised, a technique that replaces bone as well as cartilage is structurally the right fit.

MACI works differently. The porcine collagen membrane carries chondrocytes to regenerate the cartilage layer, but it has no capacity to rebuild bone. Where there is significant subchondral bone loss or deep bone cysts beneath the defect, the membrane has nothing solid to anchor to and bond formation is impaired. For this reason, significant bone involvement is generally considered a contraindication to MACI — not minor surface irregularity, but meaningful loss of the subchondral plate.

For defects where the bone is intact and only the cartilage layer is damaged, MACI's surface-repair approach is entirely appropriate and avoids donor-site harvesting altogether.

What 10-year follow-up data shows

Long-term follow-up data for both procedures is available, though the study designs differ in ways that shape how the numbers should be read.

OATS at 10 years

A 2024 registry study of 63 patients — mean age 27.4 years, mean defect 2.3 cm² — recorded IKDC scores rising from 46.4 before surgery to 70.4 at 10-year follow-up; 60% of patients exceeded the minimum clinically important difference. Reoperation occurred in 28.6% over that period, but only 2 of the 63 required conversion to knee replacement. Because this is registry data rather than a randomised trial, the figures are best read as real-world durability evidence rather than controlled proof. A 2024 network meta-analysis (Muthu et al.) separately confirmed that OATS demonstrates significantly better functional outcomes at 10 years compared with microfracture — a finding that also held for ACI/MACI.

MACI at 10 years

The strongest controlled evidence for MACI comes from the SUMMIT randomised trial, which showed improved KOOS pain and function scores at both 2 and 5 years compared with microfracture in defects of 3 cm² or larger. That is Level 1 evidence in the size range where MACI is most commonly used.

The longer picture is more cautious. A prospective study tracking MRI fill and biopsy results found the graft covering 90% of the defect at 2 years, declining to 49% by year 10. Biopsy findings were more striking: 73% of samples taken at long-term follow-up showed fibrocartilage rather than true hyaline cartilage — meaning the tissue that forms is biologically different from the original articular surface, even when clinical scores remain acceptable. Scores plateaued around the 5-year mark and declined thereafter. For a patient in their mid-30s, that trajectory is worth weighing: the procedure may deliver a decade of good function, but what happens beyond that horizon is still an open question in the literature.

Bone marrow oedema is common on MRI scans after both procedures and can look concerning, but a multicentre study found no significant difference in clinical or imaging outcomes at 5 years between patients who showed this signal and those who did not.

Patient factors that shift the recommendation

Three practical filters sit alongside defect size and bone involvement: the patient's physiology, the procedure's timeline, and any prior treatment history.

Body weight, age, and osteoarthritis grade narrow the OATS pathway. Relative contraindications include BMI above 40, age above 50, and Kellgren-Lawrence grade 2 or higher osteoarthritis. Beyond those thresholds, plug fixation becomes less predictable and the surrounding joint environment is less favourable for long-term success — selection criteria that guide clinical decision-making rather than form absolute barriers, but that consultants weigh carefully for each patient.

Staging and recovery favour OATS for patients with limited capacity to take extended time away from work or caring responsibilities. OATS is completed in a single operation. MACI, by contrast, requires planning around a two-stage timeline: an initial arthroscopic biopsy to harvest chondrocytes, a laboratory culture period typically lasting six to twelve weeks, then a second procedure for implantation. After tibiofemoral MACI, protected weightbearing continues for seven to nine weeks; patellofemoral MACI allows earlier weightbearing with a brace from the outset.

Prior microfracture is a history detail that affects both routes. Marrow-stimulation alters the subchondral architecture in ways that may compromise OATS plug integration and can reduce the conditions for MACI cell survival. Patients who have already had microfracture should raise this at their first assessment.

All of these factors — physiology, practical timeline, and treatment history — are reviewed together at a consultant-led assessment alongside imaging findings to determine which pathway is the right fit.

Where the evidence still has gaps

No direct randomised comparison of MACI against OATS has been conducted across matched defect sizes. The size-based guidance outlined earlier is extrapolated from separate trial programmes — the SUMMIT RCT for MACI, registry series for OATS — rather than a single study that enrolled equivalent patients and assigned them to one procedure or the other. Most long-term OATS data also come from registries with heterogeneous inclusion criteria, which limits direct comparability even within that evidence base.

The 10-year biopsy trajectory covered in the outcomes section — fibrocartilage dominant in most MACI samples despite acceptable clinical scores — leaves an open question about very-long-term joint preservation that the current literature has not resolved. It does not indicate early failure, but it does mean the biological durability picture is incomplete.

Emerging adjuncts such as mesenchymal stem cells and novel scaffold materials are under active investigation, but none has yet been assessed in high-quality comparative trials against either established technique.

These gaps are precisely why 'it depends on your specific situation' is a clinically accurate answer rather than a hedge. A thorough pre-operative assessment — including MRI characterisation of the lesion's depth, size, and bone involvement — is what converts a general algorithm into a recommendation matched to the individual.

Assessment and next steps at MSK Doctors

The framework this article describes — size first, then bone involvement, then patient factors — gets most patients to the right door. Converting that framework into an actual recommendation requires precise lesion characterisation: measured defect area, confirmed subchondral integrity, and an understanding of loading patterns that imaging alone may not reveal.

At MSK Doctors, that assessment combines detailed MRI review — supported by onMRI™ AI-driven analysis for accurate lesion sizing and bone-layer evaluation — with MAI Motion® biomechanical assessment, which can identify the alignment or load-distribution issues that sometimes indicate a concurrent osteotomy alongside whichever cartilage procedure is chosen.

Consultations take place at the Sleaford and Grantham clinics in Lincolnshire; no GP referral is needed. London-based patients can access the same consultant expertise through the London Cartilage Clinic.

To book an assessment, visit mskdoctors.com.

1. [1] Mosaicplasty/OATS Remains a Durable Solution for Symptomatic Chondral Defects: 2–10 Year Follow-up (2024). (2024). https://doi.org/10.1177/2325967124s00003 https://doi.org/10.1177/2325967124s00003
2. [2] Prospective Outcome, MRI and Biopsy Study of MACI Cartilage Transplantation (2017). (2017). https://doi.org/10.1177/2325967117S00186 https://doi.org/10.1177/2325967117S00186
3. [3] Bone Marrow Edema-Like Signal After Cartilage Repair Does Not Affect Outcomes at 5 Years (2024). (2024). https://doi.org/10.1007/s00330-024-11078-8 https://doi.org/10.1007/s00330-024-11078-8
4. [4] Consensus on Rehabilitation Guidelines Following MACI for Knee Cartilage Lesions (2020). (2020). https://doi.org/10.1177/1947603520968876 https://doi.org/10.1177/1947603520968876