Objective: To evaluate infrared (IR) spectroscopy of serum as a screening tool to differentiate dogs affected by naturally occurring osteoarthritis (OA) associated with cranial cruciate ligament rupture (CrCLR) and controls.
Method: 104 adult dogs with CrCLR (affected group) and 50 adult control dogs were recruited for this prospective observational study. Serum samples were collected preoperatively from CrCLR dogs and from a subset of these dogs at 4-, and 12-week post-surgical intervention to stabilize the affected stifles. Serum was collected once from control dogs. Dry films were made from serum samples, and IR absorbance spectra acquired. Data preprocessing, principal component analysis and multivariate analysis of covariance were performed to separate samples from the two groups, and to evaluate temporal differences. Weighted logistic regression with L1 regularization method was used to develop a predictive model. Model performance based on an independent test set was evaluated.
Results: Spectral data analysis revealed significant separation between the sera of CrCLR and control dogs (P < 0.0001), but not amongst different time points in the OA group. The sensitivity, specificity, AUC and accuracy of the test set were 84.62%, 96.15%, 93.20% and 92.31% respectively.
Conclusions: These findings confirm the potential of IR-spectroscopy of serum with chemometrics methods to differentiate controls from dogs with OA associated with CrCLR. This is the first step in development of an economic, and comparatively simple IR-based screening serum test for OA. Utility of this tool as a clinical screening and diagnostic test requires further investigation and validation.
Keywords: Biomarker; Infrared spectroscopy; Osteoarthritis; Stifle; dog; serum.
Published by Elsevier Ltd.