Investigating changes in the microstructure of calcified tissues using Raman microspectroscopy and chemometrics : a thesis presented in partial fulfilment of the requirements for the degree of Master of Science in Chemistry at Massey University, Manawatū, New Zealand
Loading...
Date
2020
DOI
Open Access Location
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Massey University
Rights
The Author
Abstract
Bone fracture is a growing health concern in medicine. Current clinical assessment methods for bone fracture risk are bone mineral density (BMD)-based, as bone quantity is the only aspect of bone strength that is most readily measured in clinical practice. On their own, these framework methods can not exactly predict the likelihood of individuals to fracture, since bone strength and health are influenced by both quantity and quality; it is a multi-factored probability. A sounder understanding of both facets of bone strength would expedite more accurate fracture risk assessment. Better prevention and treatment of orthopaedic diseases now rest on a greater understanding of bone quality and its underlying factors, since bone quantity has historically received more research attention. One route to confront this challenge in progressing comprehension of the underlying mechanisms of bone quality is to use animal models of human bone diseases like osteoporosis and osteoarthritis (OA). Given that any atypical chemical alterations to bone’s main components are reflected in its microstructure, and therefore contribute to the development of various bone diseases, there is increasing interest in how molecular-level changes to bone affect overall bone quality. Molecular vibrational spectroscopy is often used as a tool in disease diagnosis, as any diseasecausing chemical alterations may be identified and monitored; it also holds the potential to enable prediction of any further complications. As Raman spectroscopy is not as watersensitive as infrared is, it is highly beneficial for characterising biological specimens. Bone tissues and other biological specimens are inherently intricate, as would be the chemical information collected from them; multivariate statistical analysis is required to aid in the simplification, extraction, and classification of these large volumes of chemical information collected. This cataloguing of the actual variation of bone tissue’s chemical information would improve understanding of how damage affects the interplay between bone’s various micro- and macrostructural aspects. Principal component analysis (PCA) – one such dimensionality-reducing statistical technique – was conducted on Raman spectral data collected from two separate sets of equine bone specimens: fracture-prone third metatarsal (Mt3) and induced osteoarthritic (OA) carpal joint sections. The results from both aggregated data sets suggested that some localised microstructural differences were detectable – especially within parts of the subchondral bone. What was unclear, however, was the likely cause of these differences. These differences could potentially be highlighting areas of hypermineralisation or some organic matrix degradation within fracture predilection sites or OA-induced sites that may well be indicative of early development of orthopaedic diseases like osteoporosis or OA. Some of the common questions the PCA results raised were the extent of similarity between individuals with respect to the organic matrix component, and the extent of heterogeneity between individuals with respect to the mineral component. In order for any potential predictions to be applicable, addressing the multi-level nature of the multivariate spectral data obtained would be the first step in preparing this type of work for further validation and classification. Widening the scope of data analysis might then help in clarifying the classification of the spectral data. If not already available, condensed, fibre optic-style instrumentation might enable trialling of this technique in a practical, clinical setting. If it is practically feasible, instrumentation that even combines the two vibrational spectroscopic techniques in tandem with chemometrics to provide simultaneous groups of data from samples, could also be developed.