IDENTIFYING DIFFERENTIATING FEATURES BETWEEN NORMAL AND TORN ROTATOR CUFF TENDONS USING FOURIER TRANSFORM INFRARED SPECTROSCOPY

Abstract

Background: Up to one third of adults have been estimated to have rotator cuff tendon (RCT) tears. Larger RCT tears are associated with poorer scores and function, and are more likely to re-rupture after surgical repairs, hence there is a need for earlier identification and treatment. The aim of this study was to identify biomarkers of RCT tear pathologies to aid accurate identification and monitoring of disease progression. FTIR provides unique biochemical fingerprints of tissue specimens. All molecules are excited to higher vibrational states at specific wavelengths, which can be used to identify the chemical composition of tissues.

Methods: The chemical composition of 55 formalin-fixed RCTs was measured from patients aged between 20 and 89. RCT tears were classified according to size (Post et al.); 10 each of small, medium, large and massive and 5 partial tears. These torn RCTs were compared to 10 uninjured RCTs. A diamond attenuated total reflectance accessory was used with a FTIR spectrometer to collect spectra for each sample. The spectra were reduced and classified using standard multivariate analysis; principal component analysis (PCA), partial least square (PLS) and discriminant function analysis (DFA). Data pre-processing was applied to ensure accurate quantitative data analysis.

Results: Hierarchical cluster (HCA) demonstrated that normal and torn tendons could be clearly differentiated, and RCT could also be distinguished by their tear size. Partial tears were clearly distinguishable from normal RCT. Using a genetic algorithm we identified the following spectral regions of importance which accounted for most of the features which discriminated between normal and torn tendons:

1. 1030–1200cm-1: carbohydrates, phospholipids,

2. 1300–1700, 3000–3350cm-1: collagen structural conformation and

3. 2800–3000 cm-1: lipids.

Partial tears were distinguishable from other stages of tendon pathology based on a spectral region which correlated with collagen III.

Conclusions: FTIR can clearly distinguish normal and different sized RCT tears. This prospective non-randomized study indicates that the onset of RCT tear pathology is mainly due to an alteration of the collagen structural arrangements, with associated changes in lipids and carbohydrates. Partial tears show early onset of chemical changes, particularly in collagen III, which could be used to identify earlier stages of disease. The approach described is rapid and has the potential to be used per-operatively to determine the quality of the tendon and extent of disease, thus guiding surgical repairs or allowing monitoring of disease progression or response to treatments.