The short answer: neither wins for every patient

The most useful answer to 'which is better — mosaicplasty or ACI?' is that the question itself is framed too broadly. Both techniques restore cartilage; neither is the right choice for every patient. What settles the decision, in most cases, is the size and location of the defect.

For small, contained lesions on the femoral condyle — roughly under 2 cm² — mosaicplasty is a well-established single-stage option that transfers native hyaline cartilage in one operation. For larger defects, and for most lesions on the patella, the long-term evidence favours ACI and its modern form MACI. NICE guidance (TA477, 2017) recommends ACI for symptomatic defects greater than 2 cm² in patients who have not previously had cartilage repair, and clinical consensus (Niemeyer et al., 2016) places the formal ACI threshold at 3–4 cm².

Defect size aside, the decision also turns on the number of prior surgeries, the patient's age and activity level, and whether cell-culture infrastructure is realistically accessible. The sections below set out each consideration in detail.

What a decade of head-to-head evidence shows

The best available direct comparison comes from a single prospective RCT led by Bentley et al. and published in the Journal of Bone and Joint Surgery in 2012. One hundred patients were randomised to either ACI or mosaicplasty and followed for ten years — the only trial in this field to reach that milestone with both groups intact.

The long-term numbers are stark. By ten years, treatment had failed in 17% of ACI patients (10 of 58) compared with 55% of those who received mosaicplasty (23 of 42) — a roughly threefold difference that reached strong statistical significance (p < 0.001). Among patients whose grafts survived, ACI also produced better scores on the modified Cincinnati knee scale and the Stanmore-Bentley functional rating at the ten-year mark (p = 0.02). At the earlier one-year assessment, 87% of ACI patients had a good or excellent result compared with 70% in the mosaicplasty group, and 26% of the mosaicplasty cohort went on to fail clinically and arthroscopically, with peak failure occurring around two years.

Those figures deserve an important qualification. The trial enrolled a demanding cohort: mean defect size was approximately 4.66 cm², mean patient age was 31 years, and participants had undergone an average of 1.5 prior knee operations. These were not the small, isolated, first-presentation lesions where mosaicplasty is currently most used. The structural advantages that make ACI resilient at large-defect scale may be less decisive when the lesion is under 2 cm² and the knee is otherwise intact.

Meta-analytic evidence echoes this nuance. Li et al. (2016) compared OAT with ACI across six outcome measures and found ACI statistically superior in only one of them, suggesting the functional gap narrows considerably over shorter follow-up periods and in studies that include smaller defects. A 2024 network meta-analysis by Muthu et al. reinforces a different conclusion: both mosaicplasty and ACI produce significantly better long-term function than microfracture. In other words, the sharpest evidence divide is between either restoration technique and marrow stimulation — not between the two techniques themselves when defects are small and previously untreated.

Why defect size is the deciding factor

Three variables determine which technique is appropriate: defect size, lesion location, and the availability of donor tissue. Of these, size is the most tractable — and the one that most clearly separates the two procedures in practice.

NICE TA477 (2017) recommends ACI for defects greater than 2 cm² in previously untreated knees; Niemeyer et al. (2016) place the formal ACI threshold at 3–4 cm², defining by implication the upper boundary of mosaicplasty's effective range. The biological logic behind that boundary is practical. Mosaicplasty harvests cylindrical plugs from non-weight-bearing zones of the same knee — a finite supply. As more plugs are needed to cover a larger defect, the gaps between them become proportionally wider, and those gaps fill with fibrocartilage rather than native hyaline tissue. In a small defect, that is a minor compromise; in a larger one, the repair surface is increasingly mixed in quality. Beyond a certain point, the technique simply runs out of road.

Patellar lesions follow a separate rule. Published series consistently report poor outcomes for mosaicplasty on the patella regardless of defect size. When the patellofemoral surface is affected, ACI or MACI is the preferred approach.

One point worth stating honestly: no randomised trial has been designed specifically to compare these two techniques in defects under 2 cm². That is precisely the setting where mosaicplasty is most commonly used today — and the evidence gap there is real.

What mosaicplasty involves and its known risks

During a mosaicplasty, the surgeon harvests one or more cylindrical plugs of bone and cartilage from a relatively low-load area of the same knee — typically the peripheral edge of the femoral condyle — and press-fits them directly into the prepared defect. Everything happens in a single operating session: there is no cell-culture stage, no second procedure, and no waiting several weeks between steps.

Recovery is meaningfully shorter than with ACI. Most patients are fully weight-bearing within roughly a week of surgery, and total rehabilitation to return-to-activity runs to around four to six months, with progress paced by graft integration rather than biological maturation.

The main procedural trade-off is at the harvest site. A systematic review by Andrade et al. (21 studies, 1,726 patients) reported donor-site morbidity in approximately 5.9% of knee-to-knee procedures — a figure worth taking seriously. The most frequently reported problems were patellofemoral disturbance, affecting around 22% of patients, and crepitation, reported in approximately 31%. Surgeons discuss these risks pre-operatively because they cannot be predicted by defect size: the Andrade data showed no significant correlation between the size of the harvest and the likelihood of morbidity, meaning a small graft carries a similar profile of donor-site risk as a larger one.

At the donor site itself, the harvested area fills with fibrocartilage — a natural part of healing rather than a complication, but a biological limitation that means the knee is not simply restored to its original state.

What ACI and MACI involve and their trade-offs

Unlike mosaicplasty, ACI spreads treatment across two separate procedures — and that distinction shapes the entire experience for a patient considering it.

At the first appointment, the surgeon takes a small arthroscopic biopsy of healthy cartilage from a low-load area of the knee. That sample is sent to a specialist cell-culture laboratory, where the harvested chondrocytes are multiplied over several weeks. Once sufficient cells have been grown, the patient returns for the implantation procedure, in which the cultured cells are placed into the prepared defect — in first-generation ACI, secured beneath a periosteal patch taken from the shin.

MACI (matrix-induced ACI) follows the same two-stage principle but seeds the cultured cells onto a Type I/III collagen membrane before implantation. This refinement addresses some of the technical inconsistencies of the periosteal approach and has 10-year follow-up data showing durable functional improvements with low reoperation rates.

The practical burdens of either route are real. Rehabilitation after ACI or MACI is considerably longer than after mosaicplasty, and the total timeline — from initial biopsy to full recovery — typically extends beyond a year, accounting for the culture period and the slower biological maturation of the graft. The main documented complication is graft hypertrophy: overgrowth of the implanted tissue that may require a further arthroscopic debridement procedure to manage.

Access is also a practical constraint. Cell-culture infrastructure is not available at every orthopaedic centre, and patients outside specialist units may face additional referral steps before this pathway opens to them.

How MSK Doctors approaches the assessment

The most practical question to bring to a first consultation is deceptively simple: how large is the defect, precisely? That single MRI-confirmed measurement is the most powerful determinant of which pathway is genuinely appropriate. Below roughly 2 cm² on the femoral condyle, mosaicplasty is a well-supported single-stage option; above 2–4 cm², or where the patella is involved, the weight of evidence shifts decisively towards cell-based repair — a margin that only widens over a decade. Everything else — activity demands, prior surgery, willingness to commit to a longer rehabilitation — is then layered on top of that baseline figure.

At MSK Doctors, consultants assess patients directly, without a GP referral, and without waiting-list delays. On-site MRI at the Sleaford Regeneration Hub means the defect can be sized and characterised at the same appointment. Where objective biomechanical data would inform planning, the MAI Motion® system provides markerless motion analysis to sit alongside imaging findings. London-based patients can be seen through the London Cartilage Clinic. You can book a consultation directly at mskdoctors.com, without a referral.