Discriminating isogenic cancer cells and identifying altered unsaturated fatty acid content as associated with metastasis status, using k-means clustering and partial least squares-discriminant analysis of Raman maps

Enhancing Bone Infection Diagnosis with Raman Handheld Spectroscopy: Pathogen Discrimination and Diagnostic Potential

Osteomyelitis is a bone disease caused by bacteria that can damage bone. Raman handheld spectroscopy has emerged as a promising diagnostic tool for detecting bone infection and can be used intraoperatively during surgical procedures. This study involved 120 bone samples from 40 patients, with 80 samples infected with either Staphylococcus aureus or Staphylococcus epidermidis. Raman handheld spectroscopy demonstrated successful differentiation between healthy and infected bone samples and between the two types of bacterial pathogens. Raman handheld spectroscopy appears to be a promising diagnostic tool in bone infection and holds the potential to overcome many of the shortcomings of traditional diagnostic procedures. Further research, however, is required to confirm its diagnostic capabilities and consider other factors, such as the limit of pathogen detection and optimal calibration standards.

Rapid multi-task diagnosis of oral cancer leveraging fiber-optic Raman spectroscopy and deep learning algorithms

Introduction

Oral cancer, a predominant malignancy in developing nations, represents a global health challenge with a five-year survival rate below 50%. Nonetheless, substantial reductions in both its incidence and mortality rates can be achieved through early detection and appropriate treatment. Crucial to these treatment plans and prognosis predictions is the identification of the pathological type of oral cancer.

Methods

Toward this end, fiber-optic Raman spectroscopy emerges as an effective tool. This study combines Raman spectroscopy technology with deep learning algorithms to develop a portable intelligent prototype for oral case analysis. We propose, for the first time, a multi-task network (MTN) Raman spectroscopy classification model that utilizes a shared backbone network to simultaneously achieve different clinical staging and histological grading diagnoses.

Results

The developed model demonstrated accuracy rates of 94.88%, 94.57%, and 94.34% for tumor staging, lymph node staging, and histological grading, respectively. Its sensitivity, specificity, and accuracy compare closely with the gold standard: routine histopathological examination.

Discussion

Thus, this prototype proposed in this study has great potential for rapid, non-invasive, and label-free pathological diagnosis of oral cancer.

Comparison of Mid-Infrared Handheld and Benchtop Spectrometers to Detect Staphylococcus epidermidis in Bone Grafts

Bone analyses using mid-infrared spectroscopy are gaining popularity, especially with handheld spectrometers that enable on-site testing as long as the data quality meets standards. In order to diagnose Staphylococcus epidermidis in human bone grafts, this study was carried out to compare the effectiveness of the Agilent 4300 Handheld Fourier-transform infrared with the Perkin Elmer Spectrum 100 attenuated-total-reflectance infrared spectroscopy benchtop instrument. The study analyzed 40 non-infected and 10 infected human bone samples with Staphylococcus epidermidis, collecting reflectance data between 650 cm-1 and 4000 cm-1, with a spectral resolution of 2 cm-1 (Agilent 4300 Handheld) and 0.5 cm-1 (Perkin Elmer Spectrum 100). The acquired spectral information was used for spectral and unsupervised classification, such as a principal component analysis. Both methods yielded significant results when using the recommended settings and data analysis strategies, detecting a loss in bone quality due to the infection. MIR spectroscopy provides a valuable diagnostic tool when there is a tissue shortage and time is of the essence. However, it is essential to conduct further research with larger sample sizes to verify its pros and cons thoroughly.

Molecular Planarization of Raman Probes to Avoid Background Interference for High-Precision Intraoperative Imaging of Tumor Micrometastases and Lymph Nodes

The intraoperative imaging applications of a large number of Raman probes are hampered by the overlap of their signals with the background Raman signals generated by biological tissues. Here, we describe a molecular planarization strategy for adjusting the Raman shift of these Raman probes to avoid interference. Using this strategy, we modify the backbone of thiophene polymer-poly(3-hexylthiophene) (P3HT), and obtain the adjacent thiophene units planarized polycyclopenta[2,1-b;3,4-b']dithiophene (PCPDT). Compared with P3HT whose signal is disturbed by the Raman signal of lipids in tissues, PCPDT exhibits a 60 cm^-1 blueshift in its characteristic signal. Therefore, the PCPDT probe successfully avoids the signal of lipids, and achieves intraoperative imaging of lymph nodes and tumor micrometastasis as small as 0.30 × 0.36 mm. In summary, our study presents a concise molecular planarization strategy for regulating the signal shift of Raman probes, and brings a tunable thiophene polymer probe for high-precision intraoperative Raman imaging.

Unsupervised segmentation of biomedical hyperspectral image data: tackling high dimensionality with convolutional autoencoders

Information about the structure and composition of biopsy specimens can assist in disease monitoring and diagnosis. In principle, this can be acquired from Raman and infrared (IR) hyperspectral images (HSIs) that encode information about how a sample's constituent molecules are arranged in space. Each tissue section/component is defined by a unique combination of spatial and spectral features, but given the high dimensionality of HSI datasets, extracting and utilising them to segment images is non-trivial. Here, we show how networks based on deep convolutional autoencoders (CAEs) can perform this task in an end-to-end fashion by first detecting and compressing relevant features from patches of the HSI into low-dimensional latent vectors, and then performing a clustering step that groups patches containing similar spatio-spectral features together. We showcase the advantages of using this end-to-end spatio-spectral segmentation approach compared to i) the same spatio-spectral technique not trained in an end-to-end manner, and ii) a method that only utilises spectral features (spectral k-means) using simulated HSIs of porcine tissue as test examples. Secondly, we describe the potential advantages/limitations of using three different CAE architectures: a generic 2D CAE, a generic 3D CAE, and a 2D convolutional encoder-decoder architecture inspired by the recently proposed UwU-net that is specialised for extracting features from HSI data. We assess their performance on IR HSIs of real colon samples. We find that all architectures are capable of producing segmentations that show good correspondence with HE stained adjacent tissue slices used as approximate ground truths, indicating the robustness of the CAE-driven spatio-spectral clustering approach for segmenting biomedical HSI data. Additionally, we stress the need for more accurate ground truth information to enable a precise comparison of the advantages offered by each architecture.

Raman microscopic spectroscopy as a diagnostic tool to detect Staphylococcus epidermidis in bone grafts

INTRODUCTION

Raman microscopic spectroscopyis a new approach for further characterization and detection of molecular features in many pathological processes. This technique has been successfully applied to scrutinize the spatial distribution of small molecules and proteins within biological systems by in situ analysis. This study uses Raman microscopic spectroscopyto identify any in-depth benefits and drawbacks in diagnosing Staphylococcus epidermidis in human bone grafts.

MATERIAL AND METHODS

40 non-infected human bone samples and 10 human bone samples infected with Staphylococcus epidermidis were analyzed using Raman microscopic spectroscopy. Reflectance data were collected between 200 cm^-1 and 3600 cm^-1 with a spectral resolution of 4 cm^-1 using a Senterra II microscope (Bruker, Ettlingen, Germany). The acquired spectral information was used for spectral and unsupervised classification, such as principal component analysis.

RESULTS

Raman measurements produced distinct diagnostic spectra that were used to distinguish between non-infected human bone samples and Staphylococcus epidermidis infected human bone samples by spectral and principal component analyses. A substantial loss in bone quality and protein conformation was detected by human bone samples co-cultured with Staphylococcus epidermidis. The mineral-to-matrix ratio using the phosphate/Amide I ratio (p = 0.030) and carbonate/phosphate ratio (p = 0.001) indicates that the loss of relative mineral content in bones upon bacterial infection is higher than in non-infected human bones. Also, an increase of alterations in the collagen network (p = 0.048) and a decrease in the structural organization and relative collagen in infected human bone could be detected. Subsequent principal component analyses identified Staphylococcus epidermidis in different spectral regions, respectively, originating mainly from CH₂ deformation (wagging) of protein (at 1450 cm^-1) and bending and stretching modes of C-H groups (∼2800-3000 cm^-1).

CONCLUSION

Raman microscopic spectroscopyis presented as a promising diagnostic tool to detect Staphylococcus epidermidis in human bone grafts. Further studies in human tissues are warranted.

Deep Learning for Raman Spectroscopy: A Review

Raman spectroscopy (RS) is a spectroscopic method which indirectly measures the vibrational states within samples. This information on vibrational states can be utilized as spectroscopic fingerprints of the sample, which, subsequently, can be used in a wide range of application scenarios to determine the chemical composition of the sample without altering it, or to predict a sample property, such as the disease state of patients. These two examples are only a small portion of the application scenarios, which range from biomedical diagnostics to material science questions. However, the Raman signal is weak and due to the label-free character of RS, the Raman data is untargeted. Therefore, the analysis of Raman spectra is challenging and machine learning based chemometric models are needed. As a subset of representation learning algorithms, deep learning (DL) has had great success in data science for the analysis of Raman spectra and photonic data in general. In this review, recent developments of DL algorithms for Raman spectroscopy and the current challenges in the application of these algorithms will be discussed.

Atrial fibrillation designation with micro-Raman spectroscopy and scanning acoustic microscope

Atrial fibrillation (AF) is diagnosed with the electrocardiogram, which is the gold standard in clinics. However, sufficient arrhythmia monitoring takes a long time, and many of the tests are made in only a few seconds, which can lead arrhythmia to be missed. Here, we propose a combined method to detect the effects of AF on atrial tissue. We characterize tissues obtained from patients with or without AF by scanning acoustic microscopy (SAM) and by Raman spectroscopy (RS) to construct a mechano-chemical profile. We classify the Raman spectral measurements of the tissue samples with an unsupervised clustering method, k-means and compare their chemical properties. Besides, we utilize scanning acoustic microscopy to compare and determine differences in acoustic impedance maps of the groups. We compared the clinical outcomes with our findings using a neural network classification for Raman measurements and ANOVA for SAM measurements. Consequently, we show that the stiffness profiles of the tissues, corresponding to the patients with chronic AF, without AF or who experienced postoperative AF, are in agreement with the lipid-collagen profiles obtained by the Raman spectral characterization.

Classifying breast cancer tissue by Raman spectroscopy with one-dimensional convolutional neural network

As the most common cancer in women, breast cancer is becoming lethal worldwide. However, the current breast diagnosis technologies are not enough to meet the requirements in clinic due to some shortages of early-stage insensitiveness, time consumption and relying on the doctor's experience, etc. It's necessary to develop a creative method for the automatical diagnosis of breast cancer. Therefore, Raman spectroscopy and one-dimensional convolutional neural network (1D-CNN) algorithm were combined together for the first time to classify the healthy and cancerous breast tissues in this study. First, a number of Raman spectra were collected from breast samples of 20 patients for spectral analysis. Then, a 1D-CNN model was developed and trained for classification. In addition, the Fisher Discrimination Analysis (FDA) and Support Vector Machine (SVM) classifiers were trained and tested with the same spectral data for comparison. The best classification performance, namely the overall diagnostic accuracy of 92%, the sensitivity of 98% and the specificity of 86%, has been achieved by using 1D-CNN model. This study proves that 1D-CNN combined with Raman spectroscopy can classify breast tissues effectively and automatically and lay the foundation for automated cancer diagnosis in the future.

Multivariate analysis of Brillouin imaging data by supervised and unsupervised learning

Brillouin imaging relies on the reliable extraction of subtle spectral information from hyperspectral datasets. To date, the mainstream practice has been to use line fitting of spectral features to retrieve the average peak shift and linewidth parameters. Good results, however, depend heavily on sufficient signal-to-noise ratio and may not be applicable in complex samples that consist of spectral mixtures. In this work, we thus propose the use of various multivariate algorithms that can be used to perform supervised or unsupervised analysis of the hyperspectral data, with which we explore advanced image analysis applications, namely unmixing, classification and segmentation in a phantom and live cells. The resulting images are shown to provide more contrast and detail, and obtained on a timescale ∼10² faster than fitting. The estimated spectral parameters are consistent with those calculated from pure fitting.

Predicting Metastasis from Cues of Metastatic Cancer Stem-like Cells-3D-Ultrasensitive Metasensor at a Single-Cell Level

Metastasis is the primary reason for treatment failure and cancer-related deaths. Hence forecasting the disease in its primary state can advance the prognosis. However, existing techniques fail to reveal the tumor heterogeneity or its evolutionary cascades; hence they are not feasible to predict the onset of metastatic cancer. The key to metastasis originates from the primary tumor cells, evolving by inheriting multistep sequential cue signals. We have identified this specific population, termed metastatic cancer stem-like cells (MCSCs), to foresee cancer's ability to metastasize. An invasive property renders MCSCs nonadherent, summoning a powerful technique to forecast metastasis. Thus, we have generated an ultrasensitive 3D-metasensor to efficiently capture and investigate MCSCs and magnify the vital premetastatic signals from a single cell. We developed 3D-metasensor by an ultrafast laser ionization technique, consisting of self-assembled three-dimensionally organized nanoprobes incorporated with dopant functionalities. This distinct methodology establishes attachment with nonadherent MCSCs, elevates Raman activity, and enables probing of consequent signals (metabolic, proliferation, and metastatic) specifically altered in MCSCs. Extensive analysis using prediction tools-the area under the curve (AUC) and principal component analysis (PCA)-revealed high sensitivity (100%) and specificity (80%) to differentiate the MCSCs from other populations. Further, investigation reveals that the cue signal level from MCSCs of primary cancer is analogous to MCSCs from higher-level tumors, disclosing the relative dependence to estimate the primary tumor's capacity to metastasize. Multiple spectrum evaluation using the metasensor pinpoint the dynamic cues in MCSCs predict the onset of metastasis; thus, exploring these metastasis hallmarks can enhance prognosis and revolutionize therapy strategies.

Infrared spectroscopic imaging study of BV-2 microglia altering tumor cell biological activity and cellular fraction

Tumor brain metastasis is a severe threat to patients' neurological function, in which microglia may be involved in the process of tumor cell metastasis among nerve cells. Our study focused on the interaction between microglia and breast and lung cancer cells. Changes in the proliferation and migration ability of cocultured tumor cells were examined; synchrotron radiation-based fourier transform infrared microspectroscopy (SR-FTIR) was used to detect changes in the structures and contents of biomolecules within the tumor cells. The experimental results showed that the proliferation and migration ability of tumor cells increased after coculture, and the structures and contents of biological macromolecules in tumor cells changed. The absorption peak positions of the amide Ⅱ and amide Ⅰ bands observed for the four kinds of tumor cells changed, and the absorption intensities were significantly enhanced, indicating changes in the secondary structures and contents of proteins in tumor cells, which may be the root cause of the change in tumor cell characteristics. Therefore, the metabolites of microglia may be involved in the progression of tumor cells in the nervous system. In this study, we focused on the interaction between microglia and tumor cells by using SR-FTIR and provided a new understanding of the mechanism of brain metastasis.

Fiber-based SORS-SERDS system and chemometrics for the diagnostics and therapy monitoring of psoriasis inflammatory disease in vivo

Psoriasis is considered a widespread dermatological disease that can strongly affect the quality of life. Currently, the treatment is continued until the skin surface appears clinically healed. However, lesions appearing normal may contain modifications in deeper layers. To terminate the treatment too early can highly increase the risk of relapses. Therefore, techniques are needed for a better knowledge of the treatment process, especially to detect the lesion modifications in deeper layers. In this study, we developed a fiber-based SORS-SERDS system in combination with machine learning algorithms to non-invasively determine the treatment efficiency of psoriasis. The system was designed to acquire Raman spectra from three different depths into the skin, which provide rich information about the skin modifications in deeper layers. This way, it is expected to prevent the occurrence of relapses in case of a too short treatment. The method was verified with a study of 24 patients upon their two visits: the data is acquired at the beginning of a standard treatment (visit 1) and four months afterwards (visit 2). A mean sensitivity of ≥85% was achieved to distinguish psoriasis from normal skin at visit 1. At visit 2, where the patients were healed according to the clinical appearance, the mean sensitivity was ≈65%.

Identification of Molecular Basis for Objective Discrimination of Breast Cancer Cells (MCF-7) from Normal Human Mammary Epithelial Cells by Raman Microspectroscopy and Multivariate Curve Resolution Analysis

Raman spectroscopy (RS), a non-invasive and label-free method, has been suggested to improve accuracy of cytological and even histopathological diagnosis. To our knowledge, this novel technique tends to be employed without concrete knowledge of molecular changes in cells. Therefore, identification of Raman spectral markers for objective diagnosis is necessary for universal adoption of RS. As a model study, we investigated human mammary epithelial cells (HMEpC) and breast cancer cells (MCF-7) by RS and employed various multivariate analyses (MA) including principal components analysis (PCA), linear discriminant analysis (LDA), and support vector machine (SVM) to estimate diagnostic accuracy. Furthermore, to elucidate the underlying molecular changes in cancer cells, we utilized multivariate curve resolution analysis–alternating least squares (MCR-ALS) with non-negative constraints to extract physically meaningful spectra from complex cellular data. Unsupervised PCA and supervised MA, such as LDA and SVM, classified HMEpC and MCF-7 fairly well with high accuracy but without revealing molecular basis. Employing MCR-ALS analysis we identified five pure biomolecular spectra comprising DNA, proteins and three independent unsaturated lipid components. Relative abundance of lipid 1 seems to be strictly regulated between the two groups of cells and could be the basis for excellent discrimination by chemometrics-assisted RS. It was unambiguously assigned to linoleate rich glyceride and therefore serves as a Raman spectral marker for reliable diagnosis. This study successfully identified Raman spectral markers and demonstrated the potential of RS to become an excellent cytodiagnostic tool that can both accurately and objectively discriminates breast cancer from normal cells.

Use of Raman spectroscopy to evaluate the biochemical composition of normal and tumoral human brain tissues for diagnosis

Raman spectroscopy was used to identify biochemical differences in normal brain tissue (cerebellum and meninges) compared to tumors (glioblastoma, medulloblastoma, schwannoma, and meningioma) through biochemical information obtained from the samples. A total of 263 spectra were obtained from fragments of the normal cerebellum (65), normal meninges (69), glioblastoma (28), schwannoma (8), medulloblastoma (19), and meningioma (74), which were collected using the dispersive Raman spectrometer (830 nm, near infrared, output power of 350 mW, 20 s exposure time to obtain the spectra), coupled to a Raman probe. A spectral model based on least squares fitting was developed to estimate the biochemical concentration of 16 biochemical compounds present in brain tissue, among those that most characterized brain tissue spectra, such as linolenic acid, triolein, cholesterol, sphingomyelin, phosphatidylcholine, β-carotene, collagen, phenylalanine, DNA, glucose, and blood. From the biochemical information, the classification of the spectra in the normal and tumor groups was conducted according to the type of brain tumor and corresponding normal tissue. The classification used in discrimination models were (a) the concentrations of the biochemical constituents of the brain, through linear discriminant analysis (LDA), and (b) the tissue spectra, through the discrimination by partial least squares (PLS-DA) regression. The models obtained 93.3% discrimination accuracy through the LDA between the normal and tumor groups of the cerebellum separated according to the concentration of biochemical constituents and 94.1% in the discrimination by PLS-DA using the whole spectrum. The results obtained demonstrated that the Raman technique is a promising tool to differentiate concentrations of biochemical compounds present in brain tissues, both normal and tumor. The concentrations estimated by the biochemical model and all the information contained in the Raman spectra were both able to classify the pathological groups.

Micro-Raman spectroscopy in medicine

A potential role of optical technologies in medicine including micro-Raman spectroscopy is diagnosis of bacteria, cells and tissues which is covered in this chapter. The main advantage of Raman-based methods to complement and augment diagnostic tools is that unsurpassed molecular specificity is achieved without labels and in a nondestructive way. Principles and applications of micro-Raman spectroscopy in the context of medicine will be described. First, Raman spectra of biomolecules representing proteins, nucleic acids, lipids and carbohydrates are introduced. Second, microbial applications are summarized with the focus on typing on species and strain level, detection of infections, antibiotic resistance and biofilms. Third, cytological applications are presented to classify single cells and study cell metabolism and drug–cell interaction. Fourth, applications to tissue characterization start with discussion of lateral resolution for Raman imaging followed by Raman-based detection of pathologies and combination with other modalities. Finally, an outlook is given to translate micro-Raman spectroscopy as a clinical tool to solve unmet needs in point-of-care applications and personalized treatment of diseases.

Applications of Raman spectroscopy in cancer diagnosis

Novel approaches toward understanding the evolution of disease can lead to the discovery of biomarkers that will enable better management of disease progression and improve prognostic evaluation. Raman spectroscopy is a promising investigative and diagnostic tool that can assist in uncovering the molecular basis of disease and provide objective, quantifiable molecular information for diagnosis and treatment evaluation. This technique probes molecular vibrations/rotations associated with chemical bonds in a sample to obtain information on molecular structure, composition, and intermolecular interactions. Raman scattering occurs when light interacts with a molecular vibration/rotation and a change in polarizability takes place during molecular motion. This results in light being scattered at an optical frequency shifted (up or down) from the incident light. By monitoring the intensity profile of the inelastically scattered light as a function of frequency, the unique spectroscopic fingerprint of a tissue sample is obtained. Since each sample has a unique composition, the spectroscopic profile arising from Raman-active functional groups of nucleic acids, proteins, lipids, and carbohydrates allows for the evaluation, characterization, and discrimination of tissue type. This review provides an overview of the theory of Raman spectroscopy, instrumentation used for measurement, and variation of Raman spectroscopic techniques for clinical applications in cancer, including detection of brain, ovarian, breast, prostate, and pancreatic cancers and circulating tumor cells.

Hierarchical deep convolutional neural networks combine spectral and spatial information for highly accurate Raman-microscopy-based cytopathology

Hierarchical variants of so-called deep convolutional neural networks (DCNNs) have facilitated breakthrough results for numerous pattern recognition tasks in recent years. We assess the potential of these novel whole-image classifiers for Raman-microscopy-based cytopathology. Conceptually, DCNNs facilitate a flexible combination of spectral and spatial information for classifying cellular images as healthy or cancer-affected cells. As we demonstrate, this conceptual advantage translates into practice, where DCNNs exceed the accuracy of both conventional classifiers based on pixel spectra as well as classifiers based on morphological features extracted from Raman microscopic images. Remarkably, accuracies exceeding those of all previously proposed classifiers are obtained while using only a small fraction of the spectral information provided by the dataset. Overall, our results indicate a high potential for DCNNs in medical applications of not just Raman, but also infrared microscopy.

Unravelling the Metabolic Progression of Breast Cancer Cells to Bone Metastasis by Coupling Raman Spectroscopy and a Novel Use of Mcr-Als Algorithm

Raman spectroscopy (RS) has shown promise as a tool to reveal biochemical changes that occur in cancer processes at the cellular level. However, when analyzing clinical samples, RS requires improvements to be able to resolve biological components from the spectra. We compared the strengths of Multivariate Curve Resolution (MCR) versus Principal Component Analysis (PCA) to deconvolve meaningful biological components formed by distinct mixtures of biological molecules from a set of mixed spectra. We exploited the flexibility of the MCR algorithm to easily accommodate different initial estimates and constraints. We demonstrate the ability of MCR to resolve undesired background signals from the RS that can be subtracted to obtain clearer cancer cell spectra. We used two triple negative breast cancer cell lines, MDA-MB 231 and MDA-MB 435, to illustrate the insights obtained by RS that infer the metabolic changes required for metastasis progression. Our results show that increased levels of amino acids and lower levels of mitochondrial signals are attributes of bone metastatic cells, whereas lung metastasis tropism is characterized by high lipid and mitochondria levels. Therefore, we propose a method based on the MCR algorithm to achieve unique biochemical insights into the molecular progression of cancer cells using RS.

Confocal Raman spectroscopic imaging for in vitro monitoring of active ingredient penetration and distribution in reconstructed human epidermis model

Topically applied active cosmetic ingredients (ACI) or active pharmaceutical ingredients (API) efficacy is directly related to their efficiency of penetration in the skin. In vitro reconstructed human epidermis surrogate models offer in vivo like skin samples for transdermal studies. Using Delipidol®, an ACI currently used in the cosmetics industry, the capabilities to deliver accurate distribution maps and penetration profiles of this molecule by means of confocal Raman spectroscopic imaging have been demonstrated. Using a non-negative constrained least squares (NCLS) approach, contribution of specific molecules can be estimated at each point of spectral maps in order to deliver semi-quantitative heat maps representing the ACI levels in the different skin layers. The concentration profiles obtained are approximately single exponential for all 3 time points evaluated, with a consistent decay constant, which is independent of the sublayer structure. Notably, however, there is no significant penetration into the lower basal layers until a critical concentration is built up, after 3 hours. Combination of Raman confocal imaging with spectral unmixing methods such as NCLS is demonstrated to be a relevant approach for in vitro biological evaluation of cosmetic and pharmaceutical active ingredients and could easily be implemented as a screening tool for industrial use.

Raman spectroscopy for accurately characterizing biomolecular changes in androgen-independent prostate cancer cells

Metastatic prostate cancer resistant to hormonal manipulation is considered the advanced stage of the disease and leads to most cancer-related mortality. With new research focusing on modulating cancer growth, it is essential to understand the biochemical changes in cells that can then be exploited for drug discovery and for improving responsiveness to treatment. Raman spectroscopy has a high chemical specificity and can be used to detect and quantify molecular changes at the cellular level. Collection of large data sets generated from biological samples can be employed to form discriminatory algorithms for detection of subtle and early changes in cancer cells. The present study describes Raman finger printing of normal and metastatic hormone-resistant prostate cancer cells including analyses with principal component analysis and linear discrimination. Amino acid-specific signals were identified, especially loss of arginine band. Androgen-resistant prostate cancer cells presented a higher content of phenylalanine, tyrosine, DNA and Amide III in comparison to PNT2 cells, which possessed greater amounts of L-arginine and had a B conformation of DNA. The analysis utilized in this study could reliably differentiate the 2 cell lines (sensitivity 95%; specificity 88%).

Label-Free Molecular Imaging of Biological Cells and Tissues by Linear and Nonlinear Raman Spectroscopic Approaches

Raman spectroscopy is an emerging technique in bioanalysis and imaging of biomaterials owing to its unique capability of generating spectroscopic fingerprints. Imaging cells and tissues by Raman microspectroscopy represents a nondestructive and label-free approach. All components of cells or tissues contribute to the Raman signals, giving rise to complex spectral signatures. Resonance Raman scattering and surface-enhanced Raman scattering can be used to enhance the signals and reduce the spectral complexity. Raman-active labels can be introduced to increase specificity and multimodality. In addition, nonlinear coherent Raman scattering methods offer higher sensitivities, which enable the rapid imaging of larger sampling areas. Finally, fiber-based imaging techniques pave the way towards in vivo applications of Raman spectroscopy. This Review summarizes the basic principles behind medical Raman imaging and its progress since 2012.

Evaluation of the diagnostic potential of ex vivo Raman spectroscopy in gastric cancers: fingerprint versus high wavenumber

The aim of this study was to apply Raman spectroscopy in the high wavenumber (HW) region (2800 to 3000??cm?1) for ex vivo detection of gastric cancer and compare its diagnostic potential with that of the fingerprint (FP) region (800 to 1800??cm?1). Raman spectra were collected in the FP and HW regions to differentiate between normal mucosa (n=38) and gastric cancer (n=37). The distinctive Raman spectral differences between normal and cancer tissues are observed at 853, 879, 1157, 1319, 1338, 1448, and 2932??cm?1 and are primarily related to proteins, lipids, nucleic acids, collagen, and carotenoids in the tissue. In FP and HW Raman spectroscopy for diagnosis of gastric cancer, multivariate diagnostic algorithms based on partial-least-squares discriminant analysis, together with leave-one-sample-out cross validation, yielded diagnostic sensitivities of 94.59% and 81.08%, and specificities of 86.84% and 71.05%, respectively. Receiver operating characteristic analysis further confirmed that the FP region model performance is superior to that of the HW region model. Better differentiation between normal and gastric cancer tissues can be achieved using FP Raman spectroscopy and PLS-DA techniques, but the complementary natures of the FP and HW regions make both of them useful in diagnosis of gastric cancer.

Raman Spectroscopic Analysis Reveals Abnormal Fatty Acid Composition in Tumor Micro- and Macroenvironments in Human Breast and Rat Mammary Cancer

Fatty acids play essential roles in the growth and metastasis of cancer cells. To facilitate their avid growth and proliferation, cancer cells not only alter the fatty acid synthesis and metabolism intracellularly and extracellularly, but also in the macroenvironment via direct or indirect pathways. We report here, using Raman micro-spectroscopy, that an increase in the production of polyunsaturated fatty acids (PUFAs) was identified in both cancerous and normal appearing breast tissue obtained from breast cancer patients and tumor-bearing rats. By minimizing confounding effects from mixed chemicals and optimizing the signal-to-noise ratio of Raman spectra, we observed a large-scale transition from monounsaturated fatty acids to PUFAs in the tumor while only a small subset of fatty acids transitioned to PUFAs in the tumor micro- and macroenvironment. These data have important implications for further clarifying the macroenvironmental effect of cancer progression and provide new potential approaches for characterizing the tumor micro- and macroenvironment of breast cancer in both pre-clinical animal studies and clinical applications.

Epigenetic changes in cancer by Raman imaging, fluorescence imaging, AFM and scanning near-field optical microscopy (SNOM). Acetylation in normal and human cancer breast cells MCF10A, MCF7 and MDA-MB-231

This paper examines epigenetic changes in breast cancer by Raman imaging, fluorescence imaging, AFM and SNOM and discusses how they contribute to different aspects of tumourigenesis in malignant human breast epithelial cell lines MCF7 and MDA-MB-231 compared with non-malignant MCF10A cell lines. The paper focuses on information that can be extracted from Raman microscopy and Raman imaging for the biological material of nucleoli contained within the cell nucleus and lipid droplets within the cell cytoplasm. The biochemical composition of the nuclei and lipid droplets in the non-malignant and malignant human breast epithelial cell lines has been monitored. The potential of Raman microspectroscopy to monitor acetylation processes and a prognostic value of Raman biomarkers in breast cancer have been discussed.

Quantitative multi-image analysis for biomedical Raman spectroscopic imaging

Imaging by Raman spectroscopy enables unparalleled label-free insights into cell and tissue composition at the molecular level. With established approaches limited to single image analysis, there are currently no general guidelines or consensus on how to quantify biochemical components across multiple Raman images. Here, we describe a broadly applicable methodology for the combination of multiple Raman images into a single image for analysis. This is achieved by removing image specific background interference, unfolding the series of Raman images into a single dataset, and normalisation of each Raman spectrum to render comparable Raman images. Multivariate image analysis is finally applied to derive the contributing 'pure' biochemical spectra for relative quantification. We present our methodology using four independently measured Raman images of control cells and four images of cells treated with strontium ions from substituted bioactive glass. We show that the relative biochemical distribution per area of the cells can be quantified. In addition, using k-means clustering, we are able to discriminate between the two cell types over multiple Raman images. This study shows a streamlined quantitative multi-image analysis tool for improving cell/tissue characterisation and opens new avenues in biomedical Raman spectroscopic imaging.

Ratiometric Array of Conjugated Polymers-Fluorescent Protein Provides a Robust Mammalian Cell Sensor

Supramolecular complexes of a family of positively charged conjugated polymers (CPs) and green fluorescent protein (GFP) create a fluorescence resonance energy transfer (FRET)-based ratiometric biosensor array. Selective multivalent interactions of the CPs with mammalian cell surfaces caused differential change in FRET signals, providing a fingerprint signature for each cell type. The resulting fluorescence signatures allowed the identification of 16 different cell types and discrimination between healthy, cancerous, and metastatic cells, with the same genetic background. While the CP-GFP sensor array completely differentiated between the cell types, only partial classification was achieved for the CPs alone, validating the effectiveness of the ratiometric sensor. The utility of the biosensor was further demonstrated in the detection of blinded unknown samples, where 121 of 128 samples were correctly identified. Notably, this selectivity-based sensor stratified diverse cell types in minutes, using only 2000 cells, without requiring specific biomarkers or cell labeling.

Label-free identification and characterization of living human primary and secondary tumour cells

We used three label-free minimally invasive methods to characterize individual cells derived from primary and secondary tumours from the same patient, and of the same type – colorectal. Raman spectroscopy distinguished cells by their biochemical 'fingerprint' in a vibrational spectrum with 100% accuracy, and revealed that the primary cell line contains more lipids and alpha-helix proteins, whereas the secondary cell line contains more porphyrins and beta-sheet proteins. Stimulated Raman scattering (SRS) microscopy distinguished cells in chemically-specific images of CH2 bonds which revealed lipid droplets in secondary tumour cells. Atomic force microscopy (AFM) was used to distinguish cells with 80% accuracy by measuring their elasticity – secondary tumour cells (SW620) are around 3 times softer than primary ones (SW480). As well as characterizing the physical and biochemical differences between cell lines in vitro, these techniques offer three novel methods which could potentially be used for diagnosis – to assign a tumour as primary or secondary.

The lipid-reactive oxygen species phenotype of breast cancer. Raman spectroscopy and mapping, PCA and PLSDA for invasive ductal carcinoma and invasive lobular carcinoma. Molecular tumorigenic mechanisms beyond Warburg effect

Vibrational signatures of human breast tissue (invasive ductal carcinoma and invasive lobular carcinoma) were used to identify, characterize and discriminate structures in normal (noncancerous) and cancerous tissues by confocal Raman imaging, Raman spectroscopy and IR spectroscopy. The most important differences between normal and cancerous tissues were found in regions characteristic for vibrations of carotenoids, fatty acids, proteins, and interfacial water. Particular attention was paid to the role played by unsaturated fatty acids and their derivatives. K-means clustering and basis analysis followed by PCA and PLSDA is employed to analyze Raman spectroscopic maps of human breast tissue and for a statistical analysis of the samples (82 patients, 164 samples). Raman maps successfully identify regions of carotenoids, fatty acids, and proteins. The intensities, frequencies and profiles of the average Raman spectra differentiate the biochemical composition of normal and cancerous tissues. The paper demonstrates that Raman imaging has reached a clinically relevant level in regard to breast cancer diagnosis applications. The sensitivity and specificity obtained directly from PLSLD and cross validation are equal to 90.5% and 84.8% for calibration and 84.7% and 71.9% for cross-validation respectively.

Discrimination and classification of liver cancer cells and proliferation states by Raman spectroscopic imaging

Discrimination of nodular lesions in cirrhotic liver is a challenge in the histopathologic diagnostics. For this reason, there is an urgent need for new detection methods to improve the accuracy of the diagnosis of liver cancer. Raman imaging allows to determine the spatial distribution of a variety of molecules in cells or tissue label-free and to correlate this molecular information with the morphological structures at the same sample location. This study reports investigations of two liver cancer cell lines, - HepG2 and SK-Hep1, - as well as HepG2 cells in different cellular growth phases using Raman micro-spectroscopic imaging. Spectral data of all cells were recorded as a color-coded image and subsequentially analyzed by hierarchical cluster and principal component analysis. A support vector machine-based classification algorithm reliably predicts previously unknown cancer cells and cell cycle phases. By including selectively the Raman spectra of the cytoplasmic lipids in the classifier, the accuracy has been improved. The main spectral differences that were found in the comparative analysis can be attributed to a higher expression of unsaturated fatty acids in the hepatocellular carcinoma cells and during the proliferation phase. This corresponds to the already examined de novo lipogenesis in cells of liver cancer.

Identifying the lineages of individual cells in cocultures by multivariate analysis of Raman spectra

The cellular and matrix cues that induce stem cell differentiation into distinct cell lineages must be identified to permit the ex vivo expansion of desired cell populations for clinical applications. Combinatorial biomaterials enable screening multiple different microenvironments while using small numbers of rare stem cells. New methods to identify the phenotypes of individual cells in cocultures with location specificity would increase the efficiency and throughput of these screening platforms. Here, we demonstrate that partial least-squares discriminant analysis (PLS-DA) models of calibration Raman spectra from cells in pure cultures can be used to identify the lineages of individual cells in more complex culture environments. The calibration Raman spectra were collected from individual cells of four different lineages, and a PLS-DA model that captured the Raman spectral profiles characteristic of each cell line was created. The application of these models to Raman spectra from test sets of cells indicated individual, fixed and living cells in separate monocultures, as well as those in more complex culture environments, such as cocultures, could be identified with low error. Cells from populations with very similar biochemistries could also be identified with high accuracy. We show that these identifications are based on reproducible cell-related spectral features, and not spectral contributions from the culture environment. This work demonstrates that PLS-DA of Raman spectra acquired from pure monocultures provides an objective, noninvasive, and label-free approach for accurately identifying the lineages of individual, living cells in more complex coculture environments.

Applications of Raman micro-spectroscopy to stem cell technology: label-free molecular discrimination and monitoring cell differentiation

Stem cell therapy is widely acknowledged as a key medical technology of the 21st century which may provide treatments for many currently incurable diseases. These cells have an enormous potential for cell replacement therapies to cure diseases such as Parkinson's disease, diabetes and cardiovascular disorders, as well as in tissue engineering as a reliable cell source for providing grafts to replace and repair diseased tissues. Nevertheless, the progress in this field has been difficult in part because of lack of techniques that can measure non-invasively the molecular properties of cells. Such repeated measurements can be used to evaluate the culture conditions during differentiation, cell quality and phenotype heterogeneity of stem cell progeny. Raman spectroscopy is an optical technique based on inelastic scattering of laser photons by molecular vibrations of cellular molecules and can be used to provide chemical fingerprints of cells or organelles without fixation, lysis or use of labels and other contrast enhancing chemicals. Because differentiated cells are specialized to perform specific functions, these cells produce specific biochemicals that can be detected by Raman micro-spectroscopy. This mini-review paper describes applications of Raman micro-scpectroscopy to measure moleculare properties of stem cells during differentiation in-vitro. The paper focuses on time- and spatially-resolved Raman spectral measurements that allow repeated investigation of live stem cells in-vitro.

Surface-enhanced Raman spectroscopy of saliva proteins for the noninvasive differentiation of benign and malignant breast tumors

The capability of saliva protein analysis, based on membrane protein purification and surface-enhanced Raman spectroscopy (SERS), for detecting benign and malignant breast tumors is presented in this paper. A total of 97 SERS spectra from purified saliva proteins were acquired from samples obtained from three groups: 33 healthy subjects; 33 patients with benign breast tumors; and 31 patients with malignant breast tumors. Subtle but discernible changes in the mean SERS spectra of the three groups were observed. Tentative assignments of the saliva protein SERS spectra demonstrated that benign and malignant breast tumors led to several specific biomolecular changes of the saliva proteins. Multiclass partial least squares–discriminant analysis was utilized to analyze and classify the saliva protein SERS spectra from healthy subjects, benign breast tumor patients, and malignant breast tumor patients, yielding diagnostic sensitivities of 75.75%, 72.73%, and 74.19%, as well as specificities of 93.75%, 81.25%, and 86.36%, respectively. The results from this exploratory work demonstrate that saliva protein SERS analysis combined with partial least squares–discriminant analysis diagnostic algorithms has great potential for the noninvasive and label-free detection of breast cancer.

The role of lipid droplets and adipocytes in cancer. Raman imaging of cell cultures: MCF10A, MCF7, and MDA-MB-231 compared to adipocytes in cancerous human breast tissue

We discussed the potential of lipid droplets in nonmalignant and malignant human breast epithelial cell lines as a prognostic marker in breast cancer.

Maximizing throughput in label-free microspectroscopy with hybrid Raman imaging

Raman spectroscopy is an optical method providing sample molecular composition, which can be analyzed (by point measurements) or spatially mapped by Raman imaging. These provide different information, signal-to-noise ratios, and require different acquisition times. Here, we quantitatively assess Raman spectral features and compare the two measurement methods by multivariate analysis. We also propose a hybrid method: scanning the beam through the sample but optically binning the signal at one location on the detector. This approach generates significantly more useful spectral signals in terms of peak visibility and statistical information. Additionally, by combination with a complementary imaging mode such as quantitative phase microscopy, hybrid imaging allows high throughput and robust spectral analysis while retaining sample spatial information. We demonstrate the improved ability to discriminate between cell lines when using hybrid scanning compared to typical point mode measurements, by quantitatively evaluating spectra taken from two macrophage-like cell lines. Hybrid scanning also provides better classification capability than the full Raman imaging mode, while providing higher signal-to-noise signals with shorter acquisition times. This hybrid imaging approach is suited for various applications including cytometry, cancer versus noncancer detection, and label-free discrimination of cell types or tissues.

Label-free determination of lipid composition and secondary protein structure of human salivary noncancerous and cancerous tissues by Raman microspectroscopy

The applications of optical spectroscopic methods in cancer detection open new possibilities in oncological diagnostics.

Dynamic molecular monitoring of retina inflammation by in vivo Raman spectroscopy coupled with multivariate analysis

Retinal tissue is damaged during inflammation in Multiple Sclerosis. We assessed molecular changes in inflamed murine retinal cultures by Raman spectroscopy. Partial Least Squares-Discriminant analysis (PLS-DA) was able to classify retina cultures as inflamed with high accuracy. Using Multivariate Curve Resolution (MCR) analysis, we deconvolved 6 molecular components suffering dynamic changes along inflammatory process. Those include the increase of immune mediators (Lipoxygenase, iNOS and TNFα), changes in molecules involved in energy production (Cytochrome C, phenylalanine and NADH/NAD+) and decrease of Phosphatidylcholine. Raman spectroscopy combined with multivariate analysis allows monitoring the evolution of retina inflammation.

Complexity of fatty acid distribution inside human macrophages on single cell level using Raman micro-spectroscopy

Macrophages are phagocytic cells which are involved in the non-specific immune defense. Lipid uptake and storage behavior of macrophages also play a key role in the development of atherosclerotic lesions within walls of blood vessels. The allocation of exogenous lipids such as fatty acids in the blood stream dictates the accumulation and quantity of lipids within macrophages. In case of an overexposure, macrophages transform into foam cells because of the large amount of lipid droplets in the cytoplasm. Raman micro-spectroscopy is a powerful tool for studying single cells due to the combination of microscopic imaging with spectral information. With a spatial resolution restricted by the diffraction limit, it is possible to visualize lipid droplets within macrophages. With stable isotopic labeling of fatty acids with deuterium, the uptake and storage of exogenously provided fatty acids can be investigated. In this study, we present the results of time-dependent Raman spectroscopic imaging of single THP-1 macrophages incubated with deuterated arachidonic acid. The polyunsaturated fatty acid plays an important role in the cellular signaling pathway as being the precursor of icosanoids. We show that arachidonic acid is stored in lipid droplets but foam cell formation is less pronounced as with other fatty acids. The storage efficiency in lipid droplets is lower than in cells incubated with deuterated palmitic acid. We validate our results with gas chromatography and gain information on the relative content of arachidonic acid and its metabolites in treated macrophages. These analyses also provide evidence that significant amounts of the intracellular arachidonic acid is elongated to adrenic acid but is not metabolized any further. The co-supplementation of deuterated arachidonic acid and deuterated palmitic acid leads to a non-homogenous storage pattern in lipid droplets within single cells.

Allosteric transition induced by Mg²⁺ ion in a transactivator monitored by SERS

We demonstrate the utility of the surface-enhanced Raman spectroscopy (SERS) to monitor conformational transitions in protein upon ligand binding. The changes in protein's secondary and tertiary structures were monitored using amide and aliphatic/aromatic side chain vibrations. Changes in these bands are suggestive of the stabilization of the secondary and tertiary structure of transcription activator protein C in the presence of Mg(2+) ion, whereas the spectral fingerprint remained unaltered in the case of a mutant protein, defective in Mg(2+) binding. The importance of the acidic residues in Mg(2+) binding, which triggers an overall allosteric transition in the protein, is visualized in the molecular model. The present study thus opens up avenues toward the application of SERS as a potential tool for gaining structural insights into the changes occurring during conformational transitions in proteins.

Model based variable selection as a tool to highlight biological differences in Raman spectra of cells

Here we present a novel approach to analyse cells using Partial Least Squares – Discriminant Analysis (PLS-DA) Variable Importance Projection (VIP) scores normally used for variable selection as heat maps combined with group difference spectra to highlight significant differences in Raman band shapes and position.

Raman microspectroscopic discrimination of TCam-2 cultures reveals the presence of two sub-populations of cells

TCam-2 cells are the main in vitro model for investigations into seminomatous tumors. However, despite their widespread use, questions remain regarding the cells' homogeneity and consequently how representative they are of seminomas. We assess the TCam-2 cell line using routine and novel authentication methods to determine its homogeneity, identify any cellular sub-populations and resolve whether any changes could be due to generational differentiation. TCam-2, embryonal carcinoma cells (2102EP) and breast cancer cell (MCF7) lines were assessed using qRT-PCR, immunocytochemistry, flow cytometry and short tandem repeat analyses. Raman maps of individual cells (minimum of 10) and single scan spectra from 200 cells per culture were obtained. TCam-2s displayed the characteristic marker gene expression pattern for seminoma, were uniform in size and granularity and short tandem repeat analysis showed no contamination. However, based only on physical parameters, flowcytometry was unable to differentiate between TCam-2 and 2102EPs. Raman maps of TCam-2s comprised three equally distributed, distinct spectral patterns displaying large intercellular single spectral variation. All other cells showed little variation. Principal component, cluster and local spectral angle analyses indicated that the TCam-2s contained two different types of cells, one of which comprised two subgroups and was similar to some 2102EP cells. Protein expression corroborated the presence of different cells and generational differences. The detailed characterization provided by the Raman spectra, augmented by the routine methods, provide substantiation to the long-held suspicion that TCam-2 are not homogeneous but comprise differing cell populations, one of which may be embryonal carcinoma in origin.

Segmentation of confocal Raman microspectroscopic imaging data using edge-preserving denoising and clustering

Over the past decade, confocal Raman microspectroscopic (CRM) imaging has matured into a useful analytical tool to obtain spatially resolved chemical information on the molecular composition of biological samples and has found its way into histopathology, cytology, and microbiology. A CRM imaging data set is a hyperspectral image in which Raman intensities are represented as a function of three coordinates: a spectral coordinate λ encoding the wavelength and two spatial coordinates x and y. Understanding CRM imaging data is challenging because of its complexity, size, and moderate signal-to-noise ratio. Spatial segmentation of CRM imaging data is a way to reveal regions of interest and is traditionally performed using nonsupervised clustering which relies on spectral domain-only information with the main drawback being the high sensitivity to noise. We present a new pipeline for spatial segmentation of CRM imaging data which combines preprocessing in the spectral and spatial domains with k-means clustering. Its core is the preprocessing routine in the spatial domain, edge-preserving denoising (EPD), which exploits the spatial relationships between Raman intensities acquired at neighboring pixels. Additionally, we propose to use both spatial correlation to identify Raman spectral features colocalized with defined spatial regions and confidence maps to assess the quality of spatial segmentation. For CRM data acquired from midsagittal Syrian hamster ( Mesocricetus auratus ) brain cryosections, we show how our pipeline benefits from the complex spatial-spectral relationships inherent in the CRM imaging data. EPD significantly improves the quality of spatial segmentation that allows us to extract the underlying structural and compositional information contained in the Raman microspectra.

Raman spectroscopy in biomedicine - non-invasive in vitro analysis of cells and extracellular matrix components in tissues

Raman spectroscopy is an established laser-based technology for the quality assurance of pharmaceutical products. Over the past few years, Raman spectroscopy has become a powerful diagnostic tool in the life sciences. Raman spectra allow assessment of the overall molecular constitution of biological samples, based on specific signals from proteins, nucleic acids, lipids, carbohydrates, and inorganic crystals. Measurements are non-invasive and do not require sample processing, making Raman spectroscopy a reliable and robust method with numerous applications in biomedicine. Moreover, Raman spectroscopy allows the highly sensitive discrimination of bacteria. Rama spectra retain information on continuous metabolic processes and kinetics such as lipid storage and recombinant protein production. Raman spectra are specific for each cell type and provide additional information on cell viability, differentiation status, and tumorigenicity. In tissues, Raman spectroscopy can detect major extracellular matrix components and their secondary structures. Furthermore, the non-invasive characterization of healthy and pathological tissues as well as quality control and process monitoring of in vitro-engineered matrix is possible. This review provides comprehensive insight to the current progress in expanding the applicability of Raman spectroscopy for the characterization of living cells and tissues, and serves as a good reference point for those starting in the field.