Establishing spectrochemical changes in the natural history of oesophageal adenocarcinoma from tissue Raman mapping analysis

On-chip Raman spectroscopy of live single cells for the staging of oesophageal adenocarcinoma progression

The absence of early diagnosis contributes to oesophageal cancer being the sixth most common cause of global cancer-associated deaths, with a 5-year survival rate of < 20%. Barrett's oesophagus is the main pre-cancerous condition to adenocarcinoma development, characterised by the morphological transition of oesophageal squamous epithelium to metaplastic columnar epithelium. Early tracking and treatment of oesophageal adenocarcinoma could dramatically improve with diagnosis and monitoring of patients with Barrett's Oesophagus. Current diagnostic methods involve invasive techniques such as endoscopies and, with only a few identified biomarkers of disease progression, the detection of oesophageal adenocarcinoma is costly and challenging. In this work, single-cell Raman spectroscopy was combined with microfluidic techniques to characterise the development of oesophageal adenocarcinoma through the progression of healthy epithelial, Barrett's oesophagus and oesophageal adenocarcinoma cell lines. Principal component analysis and linear discriminant analysis were used to classify the different stages of cancer progression. with the ability to differentiate between healthy and cancerous cells with an accuracy of 97%. Whilst the approach could also separate the dysplastic stages from healthy or cancer with high accuracy-the intra-class separation was approximately 68%. Overall, these results highlight the potential for rapid and reliable diagnostic/prognostic screening of Barrett's Oesophagus patients.

Raman spectroscopy and machine learning for the classification of esophageal squamous carcinoma

Early diagnosis of esophageal squamous cell carcinoma (ESCC), a common malignant tumor with a low overall survival rate due to metastasis and recurrence, is critical for effective treatment and improved prognosis. Raman spectroscopy, an advanced detection technology for esophageal cancer, was developed to improve diagnosis sensitivity, specificity, and accuracy. This study proposed a novel, effective, and noninvasive Raman spectroscopy technique to differentiate and classify ESCC cell lines. Seven ESCC cell lines and tissues of an ESCC patient with staging of T3N1M0 and T3N2M0 at low and high differentiation levels were investigated through Raman spectroscopy. Raman spectral data analysis was performed with four machine learning algorithms, namely principal components analysis (PCA)- linear discriminant analysis (LDA), PCA-eXtreme gradient boosting (XGB), PCA- support vector machine (SVM), and PCA- (LDA, XGB, SVM)-stacked Gradient Boosting Machine (GBM). Four machine learning algorithms were able to classifiy ESCC cell subtypes from normal esophageal cells. The PCA-XGB model achieved an overall predictive accuracy of 85% for classifying ESCC and adjacent tissues. Moreover, an overall predictive accuracy of 90.3% was achieved in distinguishing low differentiation and high differentiation ESCC tissues with the same stage when PCA-LDA, XGM, and SVM models were combined. This study illustrated the Raman spectral traits of ESCC cell lines and esophageal tissues related to clinical pathological diagnosis. Future studies should investigate the role of Raman spectral features in ESCC pathogenesis.

Mapping of a Subgingival Dual-Species Biofilm Model Using Confocal Raman Microscopy

Techniques for continuously monitoring the formation of subgingival biofilm, in relation to the determination of species and their accumulation over time in gingivitis and periodontitis, are limited. In recent years, advancements in the field of optical spectroscopic techniques have provided an alternative for analyzing three-dimensional microbiological structures, replacing the traditional destructive or biofilm staining techniques. In this work, we have demonstrated that the use of confocal Raman spectroscopy coupled with multivariate analysis provides an approach to spatially differentiate bacteria in an in vitro model simulating a subgingival dual-species biofilm. The present study establishes a workflow to evaluate and differentiate bacterial species in a dual-species in vitro biofilm model, using confocal Raman microscopy (CRM). Biofilm models of Actinomyces denticolens and Streptococcus oralis were cultured using the “Zürich in vitro model” and were analyzed using CRM. Cluster analysis was used to spatially differentiate and map the biofilm model over a specified area. To confirm the clustering of species in the cultured biofilm, confocal laser scanning microscopy (CLSM) was coupled with fluorescent in vitro hybridization (FISH). Additionally, dense bacteria interface area (DBIA) samples, as an imitation of the clusters in a biofilm, were used to test the developed multivariate differentiation model. This confirmed model was successfully used to differentiate species in a dual-species biofilm and is comparable to morphology. The results show that the developed workflow was able to identify main clusters of bacteria based on spectral “fingerprint region” information from CRM. Using this workflow, we have demonstrated that CRM can spatially analyze two-species in vitro biofilms, therefore providing an alternative technique to map oral multi-species biofilm models.

Deconstruction of Obscure Features in SVD-Decomposed Raman Images from P. chrysogenum Reveals Complex Mixing of Spectra from Five Cellular Constituents

Raman imaging has transcended in recent times from being an analytical tool to a molecular profiling technique. Biomedical applications of this technique often rely on singular-value decomposition (SVD), principal component analysis (PCA), etc. for data analysis. These methods, however, obliterate the molecular information contained in the original Raman data leading to speculative interpretations based on relative intensities. In the present study, SVD analysis of the Raman images from Penicillium chrysogenum resulted in 11 spectral components and corresponding images with highly distorted spectral features and complex image contrast, respectively. To interpret the SVD results in molecular terms, we have developed a combined multivariate approach. By applying this methodology, we have successfully extracted the contribution of five biomolecular constituents of the P. chrysogenum filamentous cell to the SVD vectors. Molecular interpretability will help SVD/PCA surpass the realm of variance-based classification to a more meaningful molecular domain.