Scaffold-based ocular bioengineering versus direct cell injection

It is whether they can be delivered into the eye, remain where they are needed, preserve their function, and survive the first critical hours after surgery.

That distinction separates direct cell injection from scaffold-based ocular bioengineering. A cell suspension can be delivered through a relatively small cannula and may reduce the physical burden of implantation. But once the cells leave the syringe, they have to settle into an injured and highly specialized environment without the architecture they had in culture. In some studies, fewer than 5% of injected cells remain at the target site within the first day, with survival reported as low as approximately 1%. A scaffold takes the opposite approach: it makes the operation more demanding in order to give the cells a defined surface, orientation, and physical address.

For researchers evaluating ocular tissue regeneration strategies, this is not a simple contest between an older and newer technology. It is a question of where to accept risk: during surgical delivery, during early cell retention, or during long-term integration.

The retention crisis in direct cell suspension delivery

Direct injection remains attractive because the surgical concept is straightforward. Cells are expanded or prepared as a suspension, loaded into a micro-injection system, and delivered into the subretinal space or another anatomical compartment. The approach avoids implanting a larger manufactured construct and can be adapted to existing vitreoretinal surgical workflows.

The difficulty begins immediately after the injection.

Cells in suspension are not automatically anchored to the target tissue. They must distribute across the intended surface, establish contact with the host environment, and resist being displaced by fluid movement or pressure changes. The subretinal space is not an empty pocket designed to receive a bolus of cells. It is a narrow, delicate compartment bordered by photoreceptors on one side and the retinal pigment epithelium, or RPE, on the other. The margin for mechanical error is small.

Early cell loss can therefore be profound. Reported studies indicate that less than 5% of injected cells may persist at the target site during the first 24 to 48 hours. In some experimental settings, the final survival of unattached injected suspensions has been as low as 1%. Put plainly, a procedure can be technically successful in placing cells into the eye while still delivering very few viable cells to the surface where they are expected to function.

This is the practical weakness behind the apparent simplicity of the technique. The syringe solves the delivery problem, but it does not solve retention, polarity, or organization.

A cell can reach the subretinal space and still fail to become a graft. Delivery is only the first part of the therapeutic problem.

Why reflux matters

Cell reflux is one of the most consequential hazards of subretinal suspension delivery. During or after the injection, some of the material can move back through the retinotomy into the vitreous cavity. The consequences are not limited to a lower dose at the intended site.

Displaced cells may migrate across the retinal surface, contribute to preretinal membrane formation, or produce pigmented intravitreal deposits. In a vulnerable retina, these changes can increase tractional forces and raise the risk of retinal detachment. The concern is particularly relevant when the transplanted population contains pigment-producing cells or when the suspension is delivered under pressure that is difficult to control.

This does not make direct cell injection useless. Early clinical safety studies have reported preliminary visual preservation in selected patients, and a suspension can still be the most practical route when the therapeutic objective does not require a pre-organized tissue layer. But the burden of proof is higher than the injection step alone suggests. Investigators must ask not only whether cells can be placed beneath the retina, but how many remain there, in what condition, and with what distribution.

What a scaffold adds to the graft

Scaffold-based ocular bioengineering changes the unit of therapy. Instead of delivering cells as individual particles in fluid, the surgeon delivers cells together with a substrate that supports their arrangement.

For RPE replacement, that distinction is important because the RPE is not simply a population of cells occupying space. It is a polarized monolayer with directional functions. The apical and basal surfaces interact differently with photoreceptors, Bruch’s membrane, and the choroidal environment. A therapeutic construct that maintains this orientation is closer to the native tissue organization than a dispersed suspension.

Scaffolds may be fabricated from synthetic materials such as PLGA, PET, or parylene, or from biologically derived materials including decellularized extracellular matrix. Their roles are practical:

• They provide a defined surface for cell attachment before implantation.
• They help preserve a continuous, polarized monolayer.
• They reduce the chance that cells will disperse or reflux after placement.
• They maintain spatial relationships that are difficult to recreate after a free-cell injection.
• They give the surgeon a physical construct that can be positioned relative to the damaged tissue.

The historical logic is not new. Intact RPE sheets were reported in transplantation work as early as 1991. What has changed is the effort to make those sheets more reproducible, manufacturable, and compatible with modern cell-production systems. Contemporary scaffold design is therefore less about creating a biological patch in isolation and more about connecting manufacturing quality to surgical handling.

That connection is where many regenerative programs succeed or fail. A scaffold may preserve cell polarity in the laboratory, but it still has to survive loading, insertion, deployment, and contact with the host tissue. A construct that folds, tears, curls, or lands in the wrong orientation can lose much of the advantage built into its design.

Architecture versus access

The comparison between these platforms becomes clearer when their strengths are placed next to their surgical costs.

The scaffold is not the less invasive option simply because it may improve structural organization. Implanting a patch requires access large enough to introduce the construct and tools capable of controlling it inside the eye. That can mean a larger retinotomy, more complex manipulation near the retina, and a longer learning curve for the surgical team.

The suspension route presents a different physical challenge. The access may be smaller, but the injection itself has to be controlled with precision. In subretinal delivery, specialized protocols use micro-injection pressure calibrated to approximately 10 mmHg to reduce turbulence and limit cell reflux. This number should not be treated as a universal setting for every procedure; it illustrates the degree to which seemingly modest pressure changes can affect where the therapeutic material ends up.

The choice is therefore not between a difficult procedure and an easy one. It is between different forms of difficulty.

Surgical reality: the construct has to reach the right plane

In the operating room, the relevant question is not whether a platform looks elegant in a schematic. It is whether the surgeon can consistently place it in the intended anatomical plane without creating more damage than the therapy is meant to repair.

For direct injection, that means controlling the cannula, the injection volume, the pressure profile, and the size and location of the retinotomy. The surgeon must create a subretinal bleb large enough to distribute the cells while avoiding excessive force against the retina. A poorly controlled delivery can produce reflux into the vitreous or an uneven cell distribution across the target region.

The physical behavior of a suspension is also influenced by fluid dynamics. A bolus delivered too quickly can produce turbulence. A high-pressure stream can move cells away from the desired area. A cannula that is withdrawn before the pressure equalizes may allow material to return through the access point. These are not abstract engineering concerns; they determine whether the delivered dose remains in the subretinal space.

Scaffold implantation moves the critical moment from fluid delivery to device handling. The construct may need to be folded or loaded into an insertion system, advanced through a retinotomy, and released without damaging the cell layer. Orientation becomes a surgical variable. The side carrying the cells must face the correct tissue, and the patch must lie flat enough to maintain contact without creating folds or edge lift.

That is why scaffold programs require close coordination between biomaterials engineers, cell-manufacturing teams, and vitreoretinal surgeons. A scaffold cannot be evaluated only by cell viability before implantation. Its dimensions, flexibility, surface properties, radiopacity or visibility, insertion profile, and deployment behavior all influence whether it can be used reliably.

The strongest construct in a dish is not necessarily the strongest construct in the eye.

The manufacturing-to-surgery handoff

Cell replacement therapy delivery methods are often described as if manufacturing ends when the cells pass quality control. In practice, the product continues to change during the handoff to surgery.

For a suspension, the critical variables include cell concentration, aggregation, carrier composition, time outside controlled culture conditions, and resistance to shear during loading and injection. For a scaffold, the list expands to include:

• uniformity of cell seeding across the substrate;
• maintenance of a confluent and polarized layer;
• mechanical stability during loading and insertion;
• sterility and packaging of the complete construct;
• preservation of orientation during transport;
• compatibility with the surgical delivery device;
• predictable release from the inserter without folding or tearing.

A clinically useful product must be manufactured in a way that anticipates the operating room. If the surgeon needs to manipulate the implant extensively to make it sit correctly, manufacturing consistency may not be enough. The construct must be designed around the actual sequence of surgical movements.

Reflux, migration, and the consequences of losing the target plane

The subretinal space offers a potential route to diseased RPE and photoreceptor regions, but it is also unforgiving when material moves outside that plane.

With suspension injections, reflux can leave cells in the vitreous cavity. Once displaced, cells may migrate, form preretinal membranes, or create pigmented deposits. Those changes can interfere with the retina mechanically and optically. In the worst case, proliferative responses may contribute to retinal traction or detachment.

A scaffold can reduce free-cell translocation because the cells remain attached to a substrate. That is one of its most clinically meaningful advantages. The construct is not immune to displacement, however. A patch that is too large, poorly positioned, or inadequately apposed may create its own problems. The risk profile shifts from individual cell migration to implant placement, edge behavior, and tissue response around the material.

This distinction matters when interpreting preclinical results. A scaffold may show better cell retention because the experimental design gives it a stable landing site. A suspension may appear weaker because its cells are exposed to the full physical challenge of the subretinal environment. Those findings are valuable, but they should not be converted into a general claim that one platform is universally superior. The same scaffold that protects the cells can increase surgical trauma, while the same suspension that is easier to deliver may be more adaptable to irregular anatomy.

The relevant endpoint is not simply cell survival in isolation. It is the combination of graft survival, surgical integration, retinal stability, and functional recovery.

Lessons from corneal cell therapy

The cornea offers a useful comparison because it makes the role of postoperative mechanics especially visible.

Corneal endothelial cell injection therapy depends on the cells settling onto Descemet’s membrane and adhering across the damaged surface. Patients are typically required to maintain face-down positioning for several hours after surgery so gravity can support uniform cell contact. The operation may involve a suspension rather than a preformed scaffold, but the biological principle is similar: cells need a stable relationship with the target surface before they can function as a tissue.

This example clarifies why retention cannot be separated from the surrounding procedure. A cell product may be viable at the time of injection, yet its therapeutic value depends on where it settles and whether it remains attached. In the cornea, positioning helps create that opportunity. In the retina, pressure control, retinotomy management, and the choice between free cells and an organized construct play the same role.

The lesson extends beyond a single tissue. Ocular regenerative therapies are delivered into compartments where fluid movement, gravity, curvature, and tissue adhesion all shape the outcome. A platform that ignores those forces may look efficient at the manufacturing stage and underperform once it reaches the patient.

The unresolved question: how should a scaffold remodel?

Scaffolds solve one problem by creating structure. They introduce another by placing a foreign or modified matrix into a delicate tissue environment.

The long-term goal is not necessarily to preserve the scaffold forever. Depending on the material, the desired outcome may be gradual remodeling, stable support, or eventual replacement by host extracellular matrix. Each path requires control. Degradation that occurs too quickly could compromise cell attachment before the graft integrates. Degradation that is too slow could maintain a chronic interface or alter local biomechanics. The exact kinetics needed to balance matrix remodeling with the avoidance of localized, persistent subretinal inflammation remain unresolved.

Material choice therefore cannot be reduced to synthetic versus natural. PLGA, PET, parylene, and decellularized extracellular matrix offer different mechanical and biological properties, but none is automatically appropriate for every retinal application. The substrate must support cell function without becoming the dominant source of inflammation, fibrosis, or mechanical separation from the host tissue.

This is one reason long-term comparative data remain limited. A construct may retain cells well in the early postoperative period, yet still require years of observation to determine whether the material remains tolerated and whether the cells maintain a useful phenotype. Conversely, a suspension may show poor early retention but still produce a meaningful effect if a small surviving population engages the host tissue or releases beneficial factors.

The field needs head-to-head clinical evidence that measures more than visual acuity at a single follow-up point. Meaningful comparisons should include:

• the number of cells delivered and the number retained at the target site;
• the incidence of reflux, membrane formation, and retinal detachment;
• the integrity and polarity of the graft;
• the need for additional retinal surgery;
• structural changes on imaging;
• durability of functional recovery;
• inflammatory and immune responses over time.

Without those measurements, platform comparisons can become arguments about engineering preference rather than evidence about patient benefit.

The best delivery platform is the one that preserves a usable graft without asking the eye to tolerate an avoidable surgical burden.

Choosing the platform around the disease, not the device

Direct suspension and scaffold-based delivery may ultimately serve different clinical scenarios.

A suspension is likely to remain attractive when the target area is broad, the procedure must be minimally manipulative, or the therapeutic effect may arise from paracrine signaling rather than permanent tissue replacement. It can also be easier to adapt as cell populations, dosing strategies, or carrier formulations change.

A scaffold is more compelling when spatial organization is central to function. RPE replacement is the clearest example because polarity, confluence, and contact with the host interface are not optional details. A scaffold may also be useful where the therapeutic objective requires a defined cell layer rather than dispersed cells.

The decision should be made with several practical questions in view:

1. Does the therapy require an organized tissue architecture?

If the cells must establish directional transport or maintain a continuous epithelial layer, the scaffold offers a structural advantage.

2. Can the surgical team deliver the construct reproducibly?

A technically sophisticated implant is only useful if insertion, orientation, and deployment can be standardized.

3. What is the consequence of early cell loss?

If a therapeutic effect depends on a threshold number of surviving cells, the poor retention of a free suspension becomes especially important.

4. What is the tolerance for a larger retinotomy?

Scaffold delivery cannot be described as less invasive without accounting for the access and instruments required.

5. What will happen if cells leave the target compartment?

Reflux and migration may carry different risks depending on the cell type, pigment profile, and disease state.

6. Can the product be manufactured with surgical consistency?

Cell phenotype, substrate properties, sterility, packaging, and delivery-device compatibility must be controlled as one system.

The most realistic future may not belong to a single platform. Hybrid strategies could combine organized cell layers with materials engineered for controlled handling and remodeling, while more refined suspension systems may improve retention through better carriers, injection control, or temporary adhesion. Those approaches still need evidence; they should not be presented as solved technologies.

A measured path toward translation

Scaffold-based ocular bioengineering versus direct cell injection is ultimately a comparison between organization and access.

Direct injection offers a comparatively familiar route into the eye, but the cells face immediate losses from reflux, dispersion, and lack of attachment. Scaffold-based delivery protects architecture and can maintain a polarized monolayer, yet it asks more from the surgeon and introduces new questions about insertion trauma, implant handling, and long-term matrix behavior.

Neither platform removes the core difficulty of regenerative ophthalmology: the therapeutic cells must survive in a precise location and participate in tissue function without destabilizing the eye. The engineering solution is only successful when it improves that entire chain, from manufacturing and cold-chain handling through delivery, integration, and functional recovery.

For now, the evidence supports a cautious conclusion. Direct cell suspensions remain clinically relevant, particularly where flexibility and procedural simplicity matter. Scaffolds offer a persuasive answer to the retention and organization problem, especially for RPE replacement, but their value will depend on whether the surgical burden can be made reproducible and whether long-term integration remains stable.

The field should not measure progress by how convincingly a construct looks before implantation. It should measure what remains after the operation: viable cells in the right plane, a stable interface with host tissue, and a patient whose retinal function has a durable chance to recover.