Combination of optical coherence tomography and near infrared spectroscopy enhances determination of articular cartilage composition and structure

Abstract

Conventional arthroscopic evaluation of articular cartilage is subjective and poorly reproducible. Therefore, implementation of quantitative diagnostic techniques, such as near infrared spectroscopy (NIRS) and optical coherence tomography (OCT), is essential. Locations (n = 44) with various cartilage conditions were selected from mature equine fetlock joints (n = 5). These locations and their surroundings were measured with NIRS and OCT (n = 530). As a reference, cartilage proteoglycan (PG) and collagen contents, and collagen network organization were determined using quantitative microscopy. Additionally, lesion severity visualized in OCT images was graded with an automatic algorithm according to International Cartilage Research Society (ICRS) scoring system. Artificial neural network with variable selection was then employed to predict cartilage composition in the superficial and deep zones from NIRS data, and the performance of two models, generalized (including all samples) and condition-specific models (based on ICRS-grades), was compared. Spectral data correlated significantly (p < 0.002) with PG and collagen contents, and collagen orientation in the superficial and deep zones. The combination of NIRS and OCT provided the most reliable outcome, with condition-specific models having lower prediction errors (9.2%) compared to generalized models (10.4%). Therefore, the results highlight the potential of combining both modalities for comprehensive evaluation of cartilage during arthroscopy.

Machine Learning Classification of Articular Cartilage Integrity Using Near Infrared Spectroscopy

“…Furthermore, the approaches adopted in this study would need to be further validated using large animal models and human joints ex vivo. In a recent study, Sarin et al 28 applied shallow neural networks for predicting cartilage properties from its NIR spectrum, thus supporting the outcomes and underlying hypothesis of this study that machine learning methods may offer better approaches for classification of cartilage integrity from NIR spectra than traditional methods.…”

“…We quantified the potential of NIRS to distinguish between cartilage of control and experimental joints, simulating differences between healthy and degenerated cartilage. Although NIRS has been extensively applied for predicting cartilage biochemical 1,22,30 and structural 28 properties, which are the basis of its function, these properties would need to be measured for each location independently because of the site-specific variation in cartilage properties. Furthermore, the functional properties of articular cartilage, alone, are not reliable indicators of tissue integrity.…”

“…This is supported by the observation of Brown et al, 10 who showed significant overlap in the stiffness of normal and degenerated cartilage due to site-specific variations in articular cartilage structure, functional and material properties. More so, besides regression analysis for prediction of cartilage properties, 28,30 it is worth noting that no study has applied machine learning techniques for classification of cartilage integrity based on non-destructive spectroscopic methods like NIR spectroscopy, which has immense diagnostic potential.…”

“…1,2,[5][6][7]19,22,27,28,33 NIRS is capable of characterizing cartilage physical 5,27 and functional 6,27 properties, and composition. 1,22,28 Furthermore, NIRS has been proposed for assessing 4,23,32 cartilage integrity and monitoring the progression of OA using an animal model, 3 which provides in vivo and ex vivo data during methods development and optimization. Thus, the method has the potential to overcome the limitations of current conventional arthroscopy and enable quantitative diagnosis of cartilage condition during repair surgery.…”

“…Analysis of NIR spectra for assessment of cartilage integrity has so far relied on traditional multivariate linear methods, particularly partial least squares regression (PLSR), e.g., for prediction of specific tissue properties from cartilage NIR spectrum. Only few studies have applied machine learning techniques, such as support vector machines (SVM) 3 and neural networks (ANN), 28,30 for analysis of cartilage NIR spectral data, specifically regression analysis. However, no study has applied machine learning techniques for classification of connective tissue integrity based on NIRS.…”