Prediction of compressive stiffness of articular cartilage using Fourier transform infrared spectroscopy

Unique biomechanical behavior of articular cartilage is a result of its structure and composition. Interrelationships of tissue constituents (collagen, proteoglycans (PGs) and water) and tissue biomechanical parameters have been studied, but it is evident that no constituent alone explains the tissue mechanics. Fourier transform infrared (FT-IR) spectra can provide detailed information about the biochemical composition of articular cartilage. In this study, a chemometric approach to predict the biomechanical behavior of articular cartilage directly from the FT-IR spectra, i.e., without converting the data into collagen and PG information, was investigated. Partial least squares regression (PLSR) was used to predict equilibrium modulus (n=32) and dynamic modulus (n=24) of bovine cartilage samples from their average FT-IR spectra. The linear correlation coefficients between the reference and predicted values of Young's modulus and dynamic modulus were r=0.866 (p<0.001) and r=0.898 (p<0.001), respectively. When the compressive biomechanical behavior of AC is predicted, the present study indicates that similar or improved results can be obtained with FT-IR spectroscopy as compared to those of traditional biochemical methods.