Predicting corneal cross-linking treatment efficacy with real-time assessment of corneal riboflavin concentration

Purpose: To assess predictability of tissue biomechanical stiffening induced by UV-A light mediated real-time assessment of riboflavin concentration during corneal cross-linking (CXL) of human donor tissues. Setting: Studio Italiano di Oftalmologia, Italy. Design: Laboratory study. Methods: Twenty sclero-corneal tissues were randomly stratified to undergo CXL with either the epithelium intact (n=12) or removed (n=8). Samples underwent corneal soaking with 0.22% riboflavin formulation (RitSight, Regensight srl) with dosing time of t=10 minutes and t=20 minutes in Epi-Off and Epi-On protocols respectively. All tissues underwent 9 minutes UV-A irradiance at 10 mW/cm2 using theranostic device (C4V CHROMO4VIS, Regensight srl). The device utilized controlled UV-A light irradiation to induce both imaging and treatment of the cornea, providing a real time measure of corneal riboflavin concentration and treatment efficacy (i.e., theranostic score) during surgery. Tissue biomechanics was assessed with air-puff device (Corvis, Oculus Optikgeräte GmbH), which was performed before and after treatment. A three-element viscoelastic model was developed to fit the corneal deformation response to air-puff excitation and to calculate the mean corneal stiffness parameter (kc). Results: Significant corneal tissue stiffening (P<0.05) was induced by theranostic UV-A device in either CXL treatment protocol. Significant correlation was found between the theranostic score and the increase in kc (R=0.75; P=0.003). The score showed high accuracy (94%) and precision (94%) to predict correctly samples that had improved tissue biomechanical strengthening. Conclusion: Real-time assessment of corneal riboflavin concentration provided a predictive and precise approach for significant improvement of tissue strength on individual corneas, regardless of CXL treatment protocol.