Simulating Laser Vision Correction Outcomes

Today’s refractive diagnostic and treatment devices are capable of submicron precision. Yet they’re typically guided by generalised surgical planning algorithms that assume population-level responses to biometric measures, with adjustment for just one or two individual variables—primarily drawing from the surgeon’s previous results for the same procedure.

In addition to limiting the precision of refractive outcomes, current surgical planning approaches give minimal guidance on individual risk for ectasia, said William J Dupps Jr MD, PhD. Simplified nomograms for guiding procedures such as intracorneal rings and astigmatic keratotomy are even less reliable, and emerging treatments lack nomograms altogether.

“Every one of these corneal procedures is going to invoke biomechanical responses,” Professor Dupps said. “Some rely on those responses to get the treatment effect. And yet we don’t really have a biomechanical approach to understanding where that patient is likely to end up, and how we could optimise treatment. If we could find a unifying clinical decision tool that considers [individual biomechanics], that would be fantastic.”

A prospective, personalised approach

The current approach is retrospective, Prof Dupps pointed out. “We look at what has happened in the past and try to predict what will happen in the future.” It is probabilistic and population-based, with little attention to the biomechanical realities of the patient to be treated. He proposes the new approach should be more prospective, deterministic, and individualised.

A computational modelling approach using more-individualised patient biomechanical data analysed through a finite element method (FEM) could fit the bill, Prof Dupps said. FEM is a structural engineering method to create a digital twin for interrogating shape and mechanical behaviour under stress. The structure is divided into smaller elements, and the relationships between these elements are mathematically represented, allowing assessment of the global system.

Granular data for constructing such models of individual eyes comes from tomography to define eye surfaces. Demographic data and intraocular pressure are other elements introduced. Variables from emerging biomechanical measurements may also be included. Combining these data creates a 3-D eye model to verify against the existing eye.

This digital twin is then subjected to various procedure simulations, with all responses recorded. Factors such as treatment zone, flap placement, and refractive correction are considered. Outcomes assessed include refraction, astigmatism, and some measures of biomechanical stress and strain.

Models exist for LASIK, PRK, and lenticule extraction—with some more mature than others, Prof Dupps said. Validation studies of some early models showed a strong correlation with refractive outcomes but were limited due to the assumption all eyes have similar biomechanical responses.

Adding corneal hysteresis as a global biomechanical factor improved model performance, Prof Dupps said. “We improve outcomes when we incorporate that simple marker.”

Eventually, more precise biomechanical data might come from dynamic OCT and Brillouin microscopy, which can measure focal corneal stiffness at thousands of points, he added. This could make the analysis even more precise and predictive of individual treatment responses.

Prof Dupps spoke at Refractive Day of the 2025 ASCRS annual meeting in Los Angeles.

William J Dupps Jr MD, PhD, MS, FARVO is professor of ophthalmology and biomedical engineering at the Cole Eye Institute and Cleveland Clinic Lerner College of Medicine of Case Western Reserve University in Cleveland, Ohio, US. duppsw@ccf.org