健康和病变关节软骨的傅里叶变换红外光谱显微成像及Fisher判别

Fourier Transform Infrared Microscopic Imaging and Fisher Discrimination for Identification of Healthy and Degenerated Articular Cartilage

傅里叶变换红外光谱显微成像(FTIRI)可同时获得样品的红外光谱和形貌信息,其与化学计量学方法结合,可实现生物组织中主成分含量及分布的定量研究。本研究采用FTIRI技术结合主成分分析(PCA)以及Fisher判别算法对正常和病变的关节软骨进行鉴别分析。对关节软骨切片进行实现FTIR扫描及光谱分析,再利用SPSS软件对软骨的光谱(矩阵)进行主成分分析,根据主成分得分矩阵构造分类函数,结合Fisher判别算法对样本进行分类识别。正常和病变的关节软骨样品识别准确率高达95.7%(初始案例)和94.3%(交互验证案例)。本方法可准确有效地辨别关节软骨是否发生病变,为监测骨关节炎的发生和修复提供参考。

Fourier transform infrared microscopic imaging（ FTIRI） can be used to collect the infrared spectra and morphology of the samples. As combined with chemometric method,the content and spatial distribution of the principal components in biological tissues can be studied quantitatively. In this study,FTIRI combined with principal component analysis（ PCA） and Fisher discrimination was applied to identify healthy and degenerated articular cartilage samples. First,FTIRI and spectral analysis on articular cartilage specimens were achieved,as well as PCA on the infrared spectra in SPSS software. And then Fisher discrimination was adopted to construct the classification function based on the principal component score matrix for classifying the articular cartilage samples. The healthy and degenerated cartilage samples were effectively discriminated with very high accuracy of 95. 7% for initial samples and 94. 3% for cross validation samples,respectively. It is indicated that the hyphenated technique can effectively discriminate the healthy and degenerated cartilages,which provides a convenient method for monitoring osteoarthritic generation and reparation.