What comes to mind at the phrase "3D-printed cornea" — a full-thickness artificial tissue — differs from what is actually being tested. Precise Bio's PB-001 is a graft that replaces only the endothelium, the layer at the back of the cornea, and it is registered on ClinicalTrials.gov as a Phase 1 trial under NCT07325097. The planned enrollment is 15 people, in a single-arm, open-label interventional study with no control group.
In an interview with co-founder Aryeh Batt published by Live Science on August 21, 2026, the company stated that it had already implanted the device in five patients. The first implantation took place on October 29, 2025. However, no trial results have yet been posted to the registry. What information exists about these patients so far consists of company statements about progress and individual cases — it does not amount to a comparative conclusion that PB-001 is safer or more effective than donor tissue.
Replacing the water-pumping endothelium, not the whole cornea
The corneal endothelium is a single layer of cells lining the back of the cornea. It pumps out excess fluid to keep the cornea clear. When these cells stop functioning — as in Fuchs' endothelial corneal dystrophy or endothelial damage following cataract surgery — the cornea swells and clouds, reducing vision. PB-001 targets this corneal edema; it is not a product designed to recreate the cornea's shape or its full thickness.
PB-001 is a PVEK (Precise Vision Endothelial Keratoplasty) device: donor-derived, cultured human corneal endothelial cells arranged on a human collagen support layer. It is loaded into a sterile injector or cartridge, transported in corneal storage medium, and inserted into the anterior chamber during surgery. The "printing" process happens at this stage, but what is ultimately transplanted is a thin tissue graft containing cells.
Among existing endothelial transplant techniques, DMEK (Descemet membrane endothelial keratoplasty) transplants only Descemet's membrane and the endothelium, while DSAEK/DSEK also includes a supporting layer of posterior stroma. Precise Bio appears to be aiming for a design closer to DMEK — a cell layer placed on a collagen support that can be handled with existing endothelial transplant instruments. The idea of using cultured endothelial cells in humans is not new. In a single-arm, uncontrolled study reported by Kinoshita and colleagues in the NEJM in 2018, 11 patients with bullous keratopathy received an anterior chamber injection of 1 million cultured human endothelial cells combined with a ROCK inhibitor.
A 15-person single-arm trial measures early safety, not comparative efficacy
The official title of NCT07325097 is "A Prospective, Single Arm Study to Assess the Safety and Tolerability of PVEK Corneal Implant for the Treatment of Corneal Edema," sponsored by Precise Bio. The trial began on September 16, 2025, and was listed as Recruiting as of the January 2026 update. Primary completion is projected for March 2027, with study completion expected in September of that year.
Eligible participants are adults 18 or older with corneal edema due to Fuchs' endothelial corneal dystrophy or pseudophakic bullous keratopathy who require endothelial keratoplasty. The study eye must be pseudophakic, with best-corrected visual acuity between 6/379 and 6/12 (1.8 to 0.3 logMAR) and central corneal thickness greater than 0.6 mm. Because this is a small, narrowly defined early-stage trial, it is not designed to generalize results to corneal disease broadly.
The primary endpoints are adverse events and treatment-related adverse events through 6 months, along with the proportion of participants who complete 6 months of follow-up without seeking alternative treatment due to insufficient efficacy or tolerability. Secondary endpoints include adverse events through 12 months, intraocular pressure above 26 mmHg, change in best-corrected visual acuity from baseline, central corneal thickness measured by OCT, and endothelial cell density at 6 and 12 months. These are planned future measurements, not current results.
Speaking to Live Science, Batt said the first patient, who had pseudophakic bullous keratopathy, could not count fingers before surgery but was able to read menus and TV captions within weeks afterward. Standardized before-and-after values for best-corrected visual acuity, central corneal thickness, and endothelial cell density have not been published. From the course of a single post-intervention case, it is not possible to calculate an improvement rate or a causal treatment effect.
Measured results from 64 rabbits, but still at the conference-abstract stage
While human results remain unpublished, preclinical data include a randomized comparison. In ESCRS 2025 conference abstract FP07.15 (DOI: 10.82333/keye-8t18) by Wessam Salem and Mohamed Hosny, researchers performed 8.25 mm diameter Descemet's membrane stripping in 64 New Zealand White rabbits, randomly assigning 26 to a PVEK group, 22 to a collagen-support-only group, and 16 to a sham group, followed for 6 months. This was an animal study, not a simulation.
At 6 months, mean central corneal thickness was 391±181 μm in the PVEK group versus 646±297 μm in the support-only group, a difference reported as p=0.01. The sham group measured 757±153 μm, with a reported difference from the PVEK group of p<0.001. This is a comparative result showing lower central corneal thickness in the group that included the water-pumping cell layer.
Endothelial cell density in the PVEK group was 2,110±171 cells/mm² at 1 month, 1,894±181 cells/mm² at 3 months, and 1,793±160 cells/mm² at 6 months. Because actual figures exist at each time point, both the presence of cells and their measured values through 6 months can be tracked. However, this rabbit study cannot establish human vision outcomes or safety. The same caveat applies to long-term graft survival, rejection, and outcomes after freezing or international transport.
The form in which this data has been published also warrants caution. The source is an ESCRS 2025 conference abstract; a peer-reviewed full-text publication or independent replication has not been confirmed. Between-group comparisons in animals are a useful next step, but they are distinct from proof of clinical efficacy in humans.
A claim of over 400 grafts per cornea remains, for now, a description of manufacturing capacity
Precise Bio told Live Science that it can produce more than 400 grafts from a single donor cornea. The company states its product's endothelial cell density exceeds 4,000 cells/mm², compared with 2,000–2,500 cells/mm² for standard donor tissue. If cultivation can multiply grafts from a small number of donors, this could be meaningful as a response to supply constraints.
However, the figure of over 400 grafts is a company claim about manufacturing capacity, and independent verification through a peer-reviewed paper or public lot data has not been confirmed. Higher cell density also does not by itself imply superior vision outcomes or long-term graft survival in humans. Live Science itself notes that the significance of higher density needs to be confirmed in clinical trials.
The supply problem itself does have peer-reviewed comparative data. A 2016 paper by Gain et al. in JAMA Ophthalmology (DOI: 10.1001/jamaophthalmol.2015.4776), based on a cross-sectional estimate centered on 2012, reported 184,576 corneal transplants performed across 116 countries, an estimated 12.7 million people on waiting lists, and a ratio of only 1 cornea available for every 70 needed. The paper also estimated that roughly 53% of the world's population lacked access to transplantation — but this reflects activity data from roughly 14 years ago, not current waitlist figures.
Comparison targets should also not be conflated. Kinoshita and colleagues' peer-reviewed NEJM paper (2018, DOI: 10.1056/NEJMoa1712770), involving injection of cells into the anterior chamber in 11 patients, reported that at 24 weeks, all 11 eyes exceeded 500 cells/mm² centrally, 10 eyes exceeded 1,000 cells/mm², and 9 of 11 eyes improved by two or more lines of visual acuity — an improvement rate of 82% (95% CI, 48–98%). Rafat and colleagues' peer-reviewed Nature Biotechnology paper (2023, DOI: 10.1038/s41587-022-01408-w) described a pilot study in 20 patients with keratoconus who received a cell-free stromal implant made from porcine-derived collagen, reporting vision improvement in 19 patients at 24 months. However, both the target disease and the corneal layer being replaced differ from PB-001. Given the differing materials and manufacturing methods, these studies do not support PB-001's efficacy.
By 2027, the key question is whether the graft keeps functioning in humans
In evaluating PB-001, the first thing to confirm is not the progress metric of "five patients implanted," but rather what adverse events the 15 enrolled participants experience through 6 months, and how many complete follow-up without seeking alternative treatment. After that, changes in best-corrected visual acuity from baseline, central corneal thickness by OCT, and endothelial cell density need to be confirmed at the 6- and 12-month marks specified in the trial protocol.
A single-arm, open-label trial lacks a control group for direct comparison against donor corneal transplantation under equivalent conditions. Therefore, even once figures are published, superiority cannot be judged without examining patient characteristics, follow-up completion rates, missing data, and cases that switched to alternative treatment. Endothelial cell density in particular matters not just for its initial post-transplant level but for how well it is maintained over time.
The company projects U.S. commercialization by 2030, but this is a target, contingent on an IND application to the U.S. Food and Drug Administration (FDA), subsequent-phase trials, and the success of manufacturing and regulatory review. Rather than the narrative of a whole cornea being "printed," the real question is whether an endothelial graft built from cultured cells on a support layer can function safely in humans and be manufactured at scale with reliable quality. Only once trial results are published can that question finally be answered with measured data.