3D-Printed Molds Yield Custom Silicone Lenses to Treat Blindness in Guide Dogs

Guide dogs give millions of visually impaired people their independence, yet the animals themselves face a sobering paradox: dogs are among the species most prone to cataracts, the leading cause of preventable blindness in canines. A cross-sectional study of 2,000 dogs found that cataract prevalence reached 50 percent at around 9.4 years of age, and every dog older than 13.5 years showed some degree of lens opacity. Because organizations such as the Canadian National Institute for the Blind and the Thailand Association of the Blind depend on healthy, sighted animals, the shortage of effective, affordable veterinary lens implants has become a genuine bottleneck. A new study published in iScience now reports a low-cost fabrication platform that could change the economics of canine eye surgery.

The research team, led by Suwat Romphosri and Charusluk Viphavakit of Chulalongkorn University together with Sushanta K. Mitra, tackled the problem by rethinking how intraocular lenses, or IOLs, are made. Surgical removal of an opaque lens through phacoemulsification followed by IOL implantation can restore functional vision in up to 79.2 percent of canine cases, and implanted lenses markedly reduce postoperative glare while restoring depth perception compared with leaving the eye aphakic. Yet commercially available veterinary lenses remain limited in range, largely standardized, and constrained by few manufacturers, long lead times, and almost no customization options for the anatomical variation seen across breeds.

Conventional IOL materials such as polymethyl methacrylate, acrylic copolymers, and silicone elastomers offer proven optical clarity but carry inherent trade-offs. Rigid PMMA requires larger surgical incisions, while softer polymers can suffer from protein adsorption, cellular adhesion, and inflammatory deposits that degrade long-term optical performance. Directly 3D-printing a lens is also problematic, because printable photopolymer resins limit material choice and layer-by-layer printing leaves microscale roughness that scatters light. The team’s solution was to separate geometry from material: use a consumer-grade liquid crystal display stereolithography printer to fabricate molds, then cast the final lens from polydimethylsiloxane, or PDMS, a transparent, flexible, medically established silicone.

The lens design itself draws on clinical evidence. The researchers chose a plate-haptic configuration, in which broad plate-like extensions rather than flexible loops support the 6 mm optic, because such designs show significantly lower postoperative decentration, 0.12 plus or minus 0.06 mm, and less tilt than C-loop haptic alternatives. Haptic thickness was set at 1 mm to balance mechanical support with the flexibility needed for folding during implantation. Sharp-edge geometries were incorporated at both the optic and haptic peripheries: a sharp optic edge creates a bend in the posterior capsule that mechanically impedes migration of residual lens epithelial cells, reducing the risk of posterior capsule opacification, while a controlled haptic edge provides anchoring without irritating tissue.

Turning a 3D-printed mold into an optical-grade surface required two clever interventions. First, the printed molds showed radially periodic, ring-like roughness from the layer-by-layer photopolymerization process, so the team spin-coated a thin UV-curable resin layer onto the mold at 1,000, 2,000, and 3,000 rpm. Confocal laser scanning microscopy confirmed that the periodic features vanished after coating, but the measurements revealed a clear trade-off: lower spin speeds smoothed the surface more effectively yet flattened the designed curvature by adding a thicker coating. A speed of 1,000 rpm emerged as the optimal balance, yielding a mold roughness of 0.40 micrometers and a final lens roughness of 0.36 micrometers while preserving the intended geometry.

The second intervention addressed a chemistry problem. Phosphine oxide-based photoinitiators leaching from the printed resin can interfere with the platinum-catalyzed hydrosilylation reaction that cures addition-cure PDMS, causing incomplete or non-uniform curing. The researchers therefore introduced a thin polymethyl methacrylate barrier layer between mold and silicone. Raman spectroscopy of the finished lenses showed only characteristic PDMS bands, with no detectable PMMA ester carbonyl signal near 1,725 to 1,735 inverse centimeters, confirming that the barrier did not transfer to the lens surface. The resulting replication fidelity was striking: curvature radius differed between mold and lens by just 1.32 percent and thickness by 0.74 percent, and the cast lenses were actually smoother than their molds, because conformal PDMS casting fills and levels small-scale irregularities during curing.

Optical testing demonstrated the platform’s core promise: tunability. By changing only the mold curvature, with radii of 3.8, 5.5, 7.3, 8.7, and 9.6 mm, the team produced lenses spanning optical powers from 18 to 42 diopters, measured with the lens mounted in a cuvette of balanced salt solution to simulate the eye’s refractive environment. The measured focal positions agreed closely with predictions from the lensmaker’s equation, confirming that the designed geometry transfers faithfully to the finished silicone. That range covers the clinically required correction for canine pseudophakia and, crucially, allows patient-specific lenses to be generated for different axial eye lengths and corneal curvatures across breeds without altering the material system or processing conditions.

Imaging performance held up under scrutiny as well. All formulations transmitted more than 95 percent of visible light across the 400 to 700 nm band, and the reduced ultraviolet transmission inherent to PDMS may actually benefit ocular tissues. In a resolution test using a USAF 1951 target, the fabricated lens resolved group 4, element 2, corresponding to 20.16 line pairs per millimeter, with a Michelson contrast of 40.6 plus or minus 1.05 percent, statistically comparable to a commercial IOL at 41.2 percent. In a more vivid demonstration, a lens with a 20 mm curvature radius was mounted directly in front of the bare sensor of an APS-C camera, where it produced clearly recognizable images with preserved color fidelity, distinguishable text, and no visible haze or scattering.

Mechanical and biological evaluations narrowed the field to a leading formulation. Varying the Sylgard 184 base-to-curing-agent ratio from 10:1 to 20:1 left maximum compressive stress essentially unchanged near 22 MPa but cut the elastic modulus roughly in half, from 0.97 to 0.50 N per square millimeter, while maximum strain rose from about 85 to 99 percent. Oscillatory shear rheometry at 37 degrees Celsius showed that the 15:1 formulation best preserved its storage modulus under the 50 to 60 percent strains experienced inside an injector cartridge, yet the 10:1 formulation proved the strongest overall candidate once cytocompatibility was weighed. In ISO 10993-5 testing with L929 mouse fibroblasts, the 10:1 lenses achieved 83.97 percent viability before any treatment, and a simple rinse in isopropyl alcohol and deionized water lifted all formulations above the 70 percent acceptance threshold, reaching 97.78 percent for 10:1. A standard commercial injector successfully delivered the folded 10:1 lens with no cracking, tearing, or gross structural damage.

Stability testing under ISO 11979-5 protocols added further reassurance. After 30 days in balanced salt solution at 95 degrees Celsius, equivalent to roughly five years of physiological aging, the lenses retained their structure with only localized surface alterations, and visible-light transmittance remained at 80.94 percent, comparable to a commercial IOL aged under the same conditions. Extraction in distilled water, balanced salt solution, and toluene for 72 hours produced no distinct UV-vis absorption peaks, indicating minimal release of detectable leachable compounds. The authors are careful to frame the work as a prototype platform rather than a commercial replacement, noting that surface quality optimization, cyclic-loading durability, comprehensive biological evaluation, and in vivo validation still lie ahead. Even so, the message is compelling: with a desktop printer, a spin coater, and reusable molds, geometry-tunable silicone lenses spanning 18 to 42 diopters can be fabricated rapidly and affordably, a foundation that could one day keep more guide dogs seeing, and more people partnered with them.

Subject of Research: Fabrication of geometry-tunable PDMS intraocular lenses using LCD 3D-printed molds for canine cataract treatment

Article Title: Low-cost geometry-tunable PDMS intraocular lenses fabricated using LCD 3D-printed molds for canine ophthalmic applications

Article References: Low-cost geometry-tunable PDMS intraocular lenses fabricated using LCD 3D-printed molds for canine ophthalmic applications. (n.d.). https://doi.org/10.1016/j.isci.2026.117890

Image Credits: AI Generated

DOI: Not provided

Keywords: intraocular lens, PDMS, 3D printing, veterinary ophthalmology, canine cataracts, guide dogs, mold casting, biocompatibility, optical power tuning, biomaterials, phacoemulsification, ISO 10993