Pathogenesis of different diseases showed that some of them, especially thalassemia (T) and rheumatoid arthritis (RA) have an implicit association with oxidative stress and altered levels of reactive oxygen species (ROS). Introducing ROS level and the balance between ROS and antioxidants as essential metrics, an attempt was made to classify T and RA from normal individuals (treated as controls)(C) using synchronous fluorescence spectroscopy (SFS) and Raman line intensity of water. This non-invasive and label-free approach was backed up by a categorization algorithm that helped in the prediction of disease types from serum samples. The predictive system constituted principal component analysis (PCA) with four parameters, namely spectral intensity ratios of reduced nicotinamide adenine dinucleotide (NADH) to tryptophan (Trp) (NADH/Trp), kynurenine (Kyn) to tryptophan (Kyn/Trp), kynurenine to NADH (Kyn/NADH), and logarithmic changes in Raman line intensity of water (Rline), with the index headers containing the disease types. Rline has a positive correlation with both Kyn/Trp and Kyn/NADH and a negative correlation with NADH/Trp ratio, implying its direct or indirect association with oxidative stress. In addition to the classification of T, RA, and C a sub-classification of T into beta major and E-beta in their post and pre-splenectomized surgical stages could also be realized. Furthermore, receiver operating characteristic (ROC) analysis was deployed to ascertain that the misclassification error (ME) was negligible for the disease types.

A schematic representation of the workflow converging into the categorical classification of disease classes.

不同疾病的发病机理表明，其中一些疾病，尤其是地中海贫血（T）和类风湿性关节炎（RA）与氧化应激和活性氧（ROS）水平的改变有着内在的联系。引入ROS水平以及ROS和抗氧化剂之间的平衡作为基本指标，尝试使用同步荧光光谱（SFS）和水的拉曼光谱强度对正常个体（作为对照）（C）的T和RA进行分类。这种无创且无标签的方法得到分类算法的支持，该算法有助于从血清样本中预测疾病类型。该预测系统由四个参数组成主成分分析（PCA），即还原的烟酰胺腺嘌呤二核苷酸（NADH）与色氨酸（Trp）的光谱强度比（NADH / Trp），犬尿氨酸（Kyn）到色氨酸（Kyn / Trp），犬尿氨酸到NADH（Kyn / NADH）以及水的拉曼谱线强度（Rline）的对数变化，索引标头包含疾病类型。Rline与Kyn / Trp和Kyn / NADH都呈正相关，与NADH / Trp比值呈负相关，这意味着其与氧化应激直接或间接相关。除了对T，RA和C进行分类外，还可以在脾切除后和脾切除前后将T分为亚β大分子和E-β亚型。此外，还采用了接收器工作特性（ROC）分析来确定错误分类错误（ME）对于疾病类型而言可以忽略不计。带有包含疾病类型的索引标题。Rline与Kyn / Trp和Kyn / NADH都呈正相关，与NADH / Trp比值呈负相关，这意味着其与氧化应激直接或间接相关。除了对T，RA和C进行分类外，还可以在脾切除后和切除前的手术阶段中将T细分为beta major和E-beta。此外，还采用了接收器工作特征（ROC）分析来确定错误分类错误（ME）对于疾病类型而言可以忽略不计。带有包含疾病类型的索引标题。Rline与Kyn / Trp和Kyn / NADH都呈正相关，与NADH / Trp比值呈负相关，这意味着其与氧化应激直接或间接相关。除了对T，RA和C进行分类外，还可以在脾切除后和脾切除前后将T分为亚β大分子和E-β亚型。此外，还采用了接收器工作特征（ROC）分析来确定错误分类错误（ME）对于疾病类型而言可以忽略不计。和C，也可以实现在脾切除后和脾切除前手术阶段将T分为β大类和E-β类。此外，还采用了接收器工作特征（ROC）分析来确定错误分类错误（ME）对于疾病类型而言可以忽略不计。和C，也可以实现在脾切除后和脾切除前手术阶段将T分为β大类和E-β类。此外，还采用了接收器工作特征（ROC）分析来确定错误分类错误（ME）对于疾病类型而言可以忽略不计。

融合到疾病类别的分类中的工作流程的示意图。