Effects of corneal hydration on Brillouin measurements

Purpose: Brillouin microscopy is an emerging tool for clinical evaluation of corneal biomechanics. It has been documented that the hydration level of corneal tissues affect their mechanical properties, but it is not understood how the hydration-induced change may affect Brillouin measurements. This study aims to quantify the effects of corneal hydration on the Brillouin frequency shift. Methods: 20 fresh porcine eyeballs within 12 hours post mortem were randomly divided into 5 groups and immersed in Dextran solutions at different concentrations of 15, 7.5, 3.75, 1.86 and 0.93%, respectively. The hydration level of each sample at room temperature (24 C) was determined from the Dextran concentration and the central corneal thickness (CCT) measured by optical coherence tomography. Brillouin spectra of the cornea tissues were obtained with a clinically-viable 780-nm Brillouin system. Correlation between the Brillouin shifts and hydration levels was analyzed. For time-lapse measurement, Brillouin spectra were recorded continuously from porcine corneas (n=3) as they were undergoing artificially-induced hydration changes. As preliminary in vivo tests, Brillouin spectroscopy was performed on 5 healthy volunteers over 12 hours with 2-hour intervals. Ultrasound pachymetry was used to measure diurnal hydration changes in vivo, and correlation with measured Brillouin frequency shifts was determined. Results: The Brillouin shifts in ex vivo porcine corneas decreased with increasing hydration with a dependence of-19.44 MHz per 1% increase in water content. In humans, high correlation (Pearson's r > 0.75) was found between the Brillouin frequency shifts and the diurnal hydration changes (up to 1%). The peak-to-peak diurnal change of Brillouin shift was less than 20 MHz. Conclusions: Brillouin frequency shifts are sensitive to hydration level variations in the cornea. This result suggests that hydration and hydration-induced change of corneal stiffness are non-negligible factors to consider in the interpretation of Brillouin data from patients.