Rapid discrimination of hepatic echinococcosis patients' serum using vibrational spectroscopy combined with support vector machines

Machine learning-assisted serum SERS strategy for rapid and non-invasive screening of early cystic echinococcosis

Early and accurate diagnosis of cystic echinococcosis (CE) with existing technologies is still challenging. Herein, we proposed a novel strategy based on the combination of label-free serum surface-enhanced Raman scattering (SERS) spectroscopy and machine learning for rapid and non-invasive diagnosis of early-stage CE. Specifically, by establishing early- and middle-stage mouse models, the corresponding CE-infected and normal control serum samples were collected, and silver nanoparticles (AgNPs) were utilized as the substrate to obtain SERS spectra. The early- and middle-stage discriminant models were developed using a support vector machine, with diagnostic accuracies of 91.7% and 95.7%, respectively. Furthermore, by analyzing the serum SERS spectra, some biomarkers that may be related to early CE were found, including purine metabolites and protein-related amide bands, which was consistent with other biochemical studies. Thus, our findings indicate that label-free serum SERS analysis is a potential early-stage CE detection method that is promising for clinical translation.

A Review of Machine Learning Methods Recently Applied to FTIR Spectroscopy Data for the Analysis of Human Blood Cells

Machine learning (ML) is a broad term encompassing several methods that allow us to learn from data. These methods may permit large real-world databases to be more rapidly translated to applications to inform patient-provider decision-making. This paper presents a review of articles that discuss the use of Fourier transform infrared (FTIR) spectroscopy and ML for human blood analysis between the years 2019-2023. The literature review was conducted to identify published research of employed ML linked with FTIR for distinction between pathological and healthy human blood cells. The articles' search strategy was implemented and studies meeting the eligibility criteria were evaluated. Relevant data related to the study design, statistical methods, and strengths and limitations were identified. A total of 39 publications in the last 5 years (2019-2023) were identified and evaluated for this review. Diverse methods, statistical packages, and approaches were used across the identified studies. The most common methods included support vector machine (SVM) and principal component analysis (PCA) approaches. Most studies applied internal validation and employed more than one algorithm, while only four studies applied one ML algorithm to the data. A wide variety of approaches, algorithms, statistical software, and validation strategies were employed in the application of ML methods. There is a need to ensure that multiple ML approaches are used, the model selection strategy is clearly defined, and both internal and external validation are necessary to be sure that the discrimination of human blood cells is being made with the highest efficient evidence.

Rapid detection of cholecystitis by serum fluorescence spectroscopy combined with machine learning

While cholecystitis is a critical public health problem, the conventional diagnostic methods for its detection are time consuming, expensive and insufficiently sensitive. This study examined the possibility of using serum fluorescence spectroscopy and machine learning for the rapid and accurate identification of patients with cholecystitis. Significant differences were observed between the fluorescence spectral intensities of the serum of cholecystitis patients (n = 74) serum and those of healthy subjects (n = 71) at 455, 480, 485, 515, 625 and 690 nm. The ratios of characteristic fluorescence spectral peak intensities were first calculated, and principal component analysis (PCA)-linear discriminant analysis (LDA) and PCA-support vector machine (SVM) classification models were then constructed using the ratios as variables. Compared with the PCA-LDA model, the PCA-SVM model displayed better diagnostic performance in differentiating cholecystitis patients from healthy subjects, with an overall accuracy of 96.55%. This exploratory study showed that serum fluorescence spectroscopy combined with the PCA-SVM algorithm has significant potential for the development of a rapid cholecystitis screening method.