Abstract

Degenerative changes in meniscus are diagnosed during surgery by means of mechanical testing and visual evaluation. This method is qualitative and highly subjective, providing very little information on the internal state of the meniscus. Thus, there is need for novel quantitative methods that can support decision-making during arthroscopic surgery. In this study, we investigate the potential of near-infrared spectroscopy (NIRS) for mapping the biochemical constituents of human meniscus, including water, uronic acid, and hydroxyproline contents. Partial least squares regression models were developed using data from 115 measurement locations of menisci samples extracted from 7 cadavers and 11 surgery patient donors. Model performance was evaluated using an independent test set consisting of 55 measurement locations within a meniscus sample obtained from a separate cadaver. The correlation coefficient of calibration (ρ_training), test set (ρ_test), and root-mean-squared error of test set (RMSEP) were as follows: water (ρ_training = 0.61, ρ_test = 0.39, and RMSEP = 2.27 percentage points), uronic acid (ρ_training = 0.68, ρ_test = 0.69, and RMSEP = 6.09 basis points), and hydroxyproline (ρ_training = 0.84, ρ_test = 0.58, and error = 0.54 percentage points). In conclusion, the results suggest that NIRS could enable rapid arthroscopic mapping of changes in meniscus biochemical constituents, thus providing means for quantitative assessment of meniscus degeneration.