Multiscale mapping and biomechanics of healthy and pathological Human corneas – CorMecha

Our experimental approach is based on a unique combination of a controlled inflation device for mechanical testing and state-of-the-art imaging systems, primarily a polarization-resolved second harmonic generation (SHG) microscope for imaging the structure of the cornea.

SHG microscopy is the only imaging technique capable of specifically visualizing collagen without any labeling and deep within thick tissues. However, epi-detected SHG images are homogeneous and reveal only stromal striations in negative contrast. Imaging of collagen lamellae is possible in transdetection, but is not compatible with standard corneal chambers. This is why we used polarization-resolved SHG microscopy (P-SHG), which consists of recording a series of SHG images excited by linear polarizations with different orientations and provides the orientation of collagen with sub-micrometric resolution, even in homogeneous images. We then developed image processing pipelines to analyze this P-SHG data. In particular, we segmented the lamellae in order to extract structural measurements across the entire depth of intact corneas.

We have also developed an inflation device to mimic variations in intraocular pressure. It consists of a chamber that can be easily placed under an SHG microscope or optical coherence tomography (OCT) device, where pressure is measured, connected to an injector that allows precise control of the injected volume and injection rate. Digital image correlation (DIC) in 2D and 3D is then performed to obtain corneal deformation maps as a function of pressure. OCT images were also used to design a specific corneal mesh, which we combined with a collagen structure model to perform finite element simulations of the inflation tests. Finally, inflation tests were also performed after various types of refractive surgery to assess any biomechanical alterations in the corneas.

In addition to the digital tools already mentioned, artificial intelligence tools were used to analyze the value of CorvisST indices in the diagnosis of corneal stromal and endothelial disorders. CorvisST is the primary clinical device used to differentiate keratoconus and other corneal ectasias from normal corneas. Preoperative and postoperative clinical data were also collected from keratoconus patients who underwent intrastromal ring segment implantation, and various machine learning models were trained to predict postoperative visual acuity improvements.

We have imaged several control and pathological human corneas using P-SHG microscopy and obtained new structural data about stromal lamellae. The collagen lamellae are mostly oriented along the inferior–superior axis in the anterior stroma and along the nasal-temporal axis in the posterior stroma, with a gradual shift in between, and they exhibit more disorder in the anterior stroma. The angular shift between two sequential lamellae along the stromal depth is close to 90° in the posterior stroma and more dispersed in the anterior stroma. The lamellae size (thickness and area) increases from the anterior to the posterior stroma. This lamellar structure is disrupted in keratoconic corneas.

Inflation assays were performed under OCT or SHG microscope and principal deformation maps were obtained by Digital Volume Correlation (DVC). We observed a strong compression, with depth-dependent heterogeneity, which can be explained only by important outward water fluxes. Focusing on the stromal striae, we observed that their pattern does not change with pressure, even far above physiological pressure. Principal deformations on the striae increase with depth in the cornea. Our results are consistent with the fact that the striae are undulations in the cornea collagenous microstructure, which are progressively unfolded under loading. They decrease the global stiffness of the cornea, in particular in the posterior part, and thus may help in accommodating deformations. Finally, inflation tests conducted after refractive surgery showed that all treated corneas exhibited significant reductions in thickness and Young’s modulus compared to the control group. Among surgical techniques, LASIK resulted in the greatest reduction in corneal stiffness, whereas PRK showed the least impact.

Artificial intelligence analyses showed that the CorvisST indices are relevant for diagnosing CESDs and distinguishing various disorders from each other. Regarding keratoconus, the models demonstrated excellent performance. Key features for accurate predictions included preoperative keratometry values, corneal asphericity, and visual acuity.

The CorMecha project has yielded numerous results that open up several avenues for further research.

Firstly, the detailed characterization of the structure of keratoconus corneas has provided a better understanding of the origin of this pathology and could help to develop a more accurate or earlier diagnosis. This requires the CorMecha project to be supplemented by clinical studies to validate the pathophysiological hypotheses deduced from our experiments on very advanced keratoconus cases.

The various artificial intelligence studies conducted as part of this project could be extended to other types of corneal pathologies and other types of clinical data. Again, these studies will need to be validated by clinical studies.

Finally, a new series of inflation tests combined with high-resolution optical coherence tomography (OCT) imaging or multiphoton microscopy could be used to complement the CorMecha project. These tests would simultaneously quantify deformation and microstructure, particularly on corneas after surgery, with a focus on scarred areas. This data could be used to create a finite element model that simulates the cornea's response to pressure before and after surgery. This would be a key step in creating a digital twin of the cornea by combining mechanical, microstructural, and clinical data. This could improve refractive surgery by incorporating personalized consideration of corneal mechanics, which is currently absent from laser ablation profiles and is a source of residual refractive errors. These perspectives form the basis of a new proposal submitted at the 2026 ANR call.

The cornea is a unique tissue featuring several important physiological properties, mainly transparency and refraction, related to crucial biomechanical properties. It exhibits a viscoelastic behaviour which is important for maintaining the corneal curvature despite changes in intraocular pressure (IOP) and various forces applied on cornea, such as external shocks or eye rubbing. Defective mechanical behaviour is involved in various cornea pathologies, mainly keratoconus, which is characterized by localized corneal thinning and steepening with decreased biomechanical strength of the cornea, and results in decreased vision due to high irregular astigmatism and corneal opacification. Development of keratoconus, sometimes basically called ectasia, is a potential complication of corneal refractive surgery. Moreover, myopia correction by corneal refractive surgery shows variable results, which can be explained by inter-individual variability of wound healing responses and biomechanical behaviours.
The main underlying hypothesis of this project is that the corneal biomechanical properties are closely related to the corneal multiscale structure, which is a common assumption in all connective tissues, based both on experimental evidence and on numerical simulations. This project further assumes that defective corneal biomechanical properties are caused by a defective structure. Corneal stroma consists of several hundred 1-3 µm thick stacked lamellae made of collagen fibrils (25-30 mm of diameter) aligned and regularly packed to ensure cornea transparency. Lamellae are roughly organized parallel to each other with different orientations, but their size and three-dimensional (3D) organization vary along the depth and extent of the cornea. Vertical stromal striae, corresponding to lamellae undulations, are also present in the posterior stroma and have been shown to be related to biomechanical behaviour. However, this anisotropic hierarchical organization of the cornea and its physiological consequences on cornea biomechanics have been poorly characterized because of the limitations of conventional techniques.
This project aims at: (i) setting up an atlas of cornea 3D structure from the sub-micrometer scale (intra-lamellar organization of collagen fibrils) to the millimeter-centimeter scale (lamellae distribution in the full cornea, stromal striae), (ii) accurately measuring the biomechanical properties linked to this structure in physiological conditions and in various pathological conditions, including high IOP, keratoconic corneas and corneas after photo-ablation, and (iii) building a model of corneal biomechanics based on these microstructural and macroscopic data in order to provide an insight into the role of specific stromal structures. It relies on the highly original combination of well-controlled inflation device and state-of-the-art imaging setups, mainly polarization-resolved Second Harmonic Generation microscope. Specific bioimage informatics tools and pipelines will be developed to process the very large data sets (Gb to Tb) generated by this new device and quantify clinically-relevant parameters of interest. Advanced statistical analysis of the series of clinical (corneal topography …), structural (OCT, confocal, SHG…) and mechanical data obtained on the same cornea will then be performed for normal, keratoconic and photo-ablated corneas. The ultimate goals are twofold: (i) to translate the structural features observed with advanced research microscopes into easily-detectable features using commonly used techniques in clinical ophthalmology, in order to enable the diagnosis of structural defects related to defective mechanical properties; (ii) to develop a patient-specific simplified model to serve as a predictive tool by clinicians, mainly to improve refractive surgery procedures.