Raman Hyperspectral Imaging Research Articles

Machine Learning for Prediction, Classification, and Identification of Immobilized Enzymes for Biocatalysis.

Enzyme immobilization is a beneficial component involved in biocatalytic strategies. Understanding and evaluating the enzyme immobilization system plays an important role in the successful development and implementation of the biocatalysis route. Ensuring the implementation of a successful enzyme immobilization process is vital for realizing a highly functioning and well suited biocatalytic process within pharmaceutical development. To develop a method which can accurately and objectively identify and classify differences within enzyme immobilization systems, sample preparation methods, and data collection parameters. Raman hyperspectral imaging was used to obtain a total of eight spectral data sets from enzyme immobilization samples. Partial least squares discriminant analysis (PLS-DA) was used to classify and identify the samples based on their differences. Several two-class, four-class, and eight-class PLS-DA models were built to classify the different sample data sets. All models reached between 92-100% accuracy after cross-validation and external validation, illustrating great success of the models for identifying differences between the samples. Raman hyperspectral imaging with machine learning can be used to investigate, interpret, and classify different data collection parameters, sample preparation methods, and enzyme immobilization supports, providing crucial insight into enzyme immobilization process development.

Rapid coherent Raman hyperspectral imaging based on delay-spectral focusing dual-comb method and deep learning algorithm.

Rapid coherent Raman hyperspectral imaging shows great promise for applications in sensing, medical diagnostics, and dynamic metabolism monitoring. However, the spectral acquisition speed of current multiplex coherent anti-Stokes Raman scattering (CARS) microscopy is generally limited by the spectrometer integration time, and as the detection speed increases, the signal-to-noise ratio (SNR) of single spectrum will decrease, leading to a terrible imaging quality. In this Letter, we report a dual-comb coherent Raman hyperspectral microscopy imaging system developed by integrating two approaches, a rapid delay-spectral focusing method and deep learning. The spectral refresh rate is exploited by focusing the relative delay scanning in the effective Raman excitation region, enabling a spectral acquisition speed of 36 kHz, ≈4 frames/s, for a pixel resolution of 95 × 95 pixels and a spectral bandwidth no less than 200 cm-1. To improve the spectral SNR and imaging quality, the deep learning models are designed for spectral preprocessing and automatic unsupervised feature extraction. In addition, by changing the relative delay focusing region of the comb pairs, the detected spectral wavenumber region can be flexibly tuned to the high SNR region of the spectrum.

Packaged butter adulteration evaluation based on spatially offset Raman spectroscopy coupled with FastICA

Optical detection technology has been widely used in unpackaged food adulteration detection. However, due to the interference of packaging materials on the internal food optical signal, including signal occlusion, mixing and overlap precluded the accurate detection of internal food quality. In this study, a method of packaged butter adulteration evaluation based on spatially offset Raman spectroscopy (SORS) combined with fast independent component analysis (FastICA) was proposed. The adulterated butter from 0% to 100% w/w margarine at 10% intervals was covered with packaging sheets as test samples. A line-scan Raman hyperspectral imaging system was used to obtain a scattering spectral image of the packaged butter samples. The region of interest of the scattering image is extracted as the input of FastICA model to separate the internal butter signals. The extracted butter Raman features were input into four quantitative analysis models to assess the content of butter adulteration. The results showed that the ensemble model Extra-tree has the best performance with RMSEp, Rp2, and RPD values of 0.6, 0.93, and 4.73, respectively. Additionally, the applicability of the method was validated with four types of packaging materials. This rapid non-destructive testing method is beneficial to the effective testing method of packaged butter and other products industry.

Fast Raman imaging through the combination of context-aware matrix completion and low spectral resolution.

Raman hyperspectral imaging is an effective method for label-free imaging with chemical specificity, yet the weak signals and correspondingly long integration times have hindered its wide adoption as a routine analytical method. Recently, low resolution Raman imaging has been proposed to improve the spectral signal-to-noise ratio, which significantly improves the speed of Raman imaging. In this paper, low resolution Raman spectroscopy is combined with "context-aware" matrix completion, where regions of the sample that are not of interest are skipped, and the regions that are measured are under-sampled, then reconstructed with a low-rank constraint. Both simulations and experiment show that low-resolution Raman boosts the speed and image quality of the computationally-reconstructed Raman images, allowing deeper sub-sampling, reduced exposure time, and an overall >10-fold improvement in imaging speed, without sacrificing chemical specificity or spatial image quality. As the method utilizes traditional point-scan imaging, it retains full confocality and is "backwards-compatible" with pre-existing traditional Raman instruments, broadening the potential scope of Raman imaging applications.

Acceleration of high-quality Raman imaging via a locality enhanced transformer network.

Raman imaging (RI) is an outstanding technique that enables molecular-level medical diagnostics and therapy assessment by providing characteristic fingerprint and morphological information about molecules. However, obtaining high-quality Raman images generally requires a long acquisition time, up to hours, which is prohibitive for RI applications of timely cytopathology and histopathology analyses. To address this issue, image super-resolution (SR) based on deep learning, including convolutional neural networks and transformers, has been widely recognized as an effective solution to reduce the time required for achieving high-quality RI. In this study, a locality enhanced transformer network (LETNet) is proposed to perform Raman image SR. Specifically, the general architecture of the transformer is adopted with the replacement of self-attention by convolution to generate high-fidelity and detailed SR images. Additionally, the convolution in the LETNet is further optimized by utilizing depth-wise convolution to improve the computational efficiency of the model. Experiments on hyperspectral Raman images of breast cancer cells and Raman images of a few channels of brain tumor tissues demonstrate that the LETNet achieves superior 2×, 4×, and 8× SR with fewer parameters compared with other SR methods. Consequently, high-quality Raman images can be obtained with a significant reduction in time, ranging from 4 to 64 times. Overall, the proposed method provides a novel, efficient, and reliable solution to expedite high-quality RI and promote its application in real-time diagnosis and therapy.

Combination of scattering-projection interleaving and random down-sampling for compressive confocal Raman imaging.

Parallel excitation with an array of foci is one way to improve the speed of Raman hyperspectral imaging, and random interleaving of its projection has been proved to be a successful strategy for reconstructing the compressed data cube. The so-called SIRI method allows single-acquisition compressive confocal Raman imaging and provides excellent reconstruction fidelity at a high compression ratio. Here, we demonstrate that, when scattering-projection interleaving and randomly down-sampling in the spatial domain are combined, the modified SIRI allows a further reduction in the data acquisition time and an expansion of the imaging region. At a moderate down-sampling rate, the modified SIRI is even superior to its precursor in terms of reconstruction fidelity. A maximum compression ratio of 80 is also reported experimentally with the proposed method.

Raman Spectroscopy in Open-World Learning Settings Using the Objectosphere Approach.

Raman spectroscopy, combined with machine learning techniques, holds great promise for many applications as a rapid, sensitive, and label-free identification method. Such approaches perform well when classifying spectra of chemical species that were encountered during the training phase. That is, species that are known to the neural network. However, in real-world settings, such as in clinical applications, there will always be substances whose spectra have not yet been taken. When the neural network encounters these new species during the testing phase, the number of false positives becomes uncontrollable, limiting the usefulness of these techniques, especially in public safety applications. To overcome these barriers, we implemented the recently introduced Entropic Open Set and Objectosphere loss functions. To demonstrate the efficacy and efficiency of this approach, we compiled a database of hyperspectral Raman images of 40 chemical species separating them into three class categorizations. The known class consisted of 20 biologically relevant species comprising amino acids, the ignored class was 10 "irrelevant" species comprising bio-related chemicals, and the never seen before class was 10 various chemical species that the neural network had not seen before. We show that this approach not only enables the network to effectively separate the unknown species while preserving high accuracy on the known ones and reducing false positives but also performs better than the current gold standards in machine learning techniques. This opens the door to using Raman spectroscopy, combined with our novel machine learning algorithm, in a variety of practical applications. Availability and implementation: freely available on the web at https://github.com/BalytskyiJaroslaw/RamanOpenSet.git.

Accurate and fast identification of minimally prepared bacteria phenotypes using Raman spectroscopy assisted by machine learning

The worldwide increase of antimicrobial resistance (AMR) is a serious threat to human health. To avert the spread of AMR, fast reliable diagnostics tools that facilitate optimal antibiotic stewardship are an unmet need. In this regard, Raman spectroscopy promises rapid label- and culture-free identification and antimicrobial susceptibility testing (AST) in a single step. However, even though many Raman-based bacteria-identification and AST studies have demonstrated impressive results, some shortcomings must be addressed. To bridge the gap between proof-of-concept studies and clinical application, we have developed machine learning techniques in combination with a novel data-augmentation algorithm, for fast identification of minimally prepared bacteria phenotypes and the distinctions of methicillin-resistant (MR) from methicillin-susceptible (MS) bacteria. For this we have implemented a spectral transformer model for hyper-spectral Raman images of bacteria. We show that our model outperforms the standard convolutional neural network models on a multitude of classification problems, both in terms of accuracy and in terms of training time. We attain more than 96% classification accuracy on a dataset consisting of 15 different classes and 95.6% classification accuracy for six MR–MS bacteria species. More importantly, our results are obtained using only fast and easy-to-produce training and test data.

Raman hyperspectral imaging detects novel and combinations of crystals in synovial fluids of patients with a swollen joint

AbstractThe identification of synovial crystals is important for diagnosing rheumatic diseases. Currently, rheumatologists worldwide use compensated polarized light microscopy (CPLM) for crystal identification, but this technique is flawed. Raman spectroscopy might offer an objective, accurate alternative. We have tested Raman hyperspectral imaging on synovial fluid samples of 28 patients with swollen joints, measuring 5–10 crystals in each of these. Reference spectra for identification were measured using patient material, synthetic compounds, and oxalate kidney stones. Additionally, RRUFF and PANGAEA Raman databases were used to identify spectra. We identified the classical pathological crystals monosodium urate for gout and calcium pyrophosphate for calcium pyrophosphate deposition disease (CPPD). Hydroxyapatite, lipid spherules, and calcium oxalate monohydrate have also been observed and had been previously identified in synovial fluids. Other crystals, which were not previously observed in synovial fluid using CPLM, have been identified as well: calcite, aragonite, anatase, rutile, thenardite, dolomite, and a carotenoid. In addition, we found combinations of crystals in 13 out of 28 patients. We propose that the observed large variation of detectable crystals from a small population of patients and a small number of crystals per patient has significant ramifications for the use of compensated polarized light microscopy for the diagnosis of gout and CPPD. Further studies are required to learn the clinical significance of these crystals in human arthritis.

Simulated fine-needle aspiration diagnosis of follicular thyroid nodules by hyperspectral Raman microscopy and chemometric analysis.

.Significance: Follicular thyroid carcinoma carries a substantially poor prognosis due to its unique biological behavior and less favorable outcomes. In particular, fine-needle aspiration (FNA) biopsies, which play a key role in screening thyroid nodules, cannot differentiate benign from malignant follicular neoplasm.Aim: We report on the use of hyperspectral Raman microscopy in combination with chemometric analysis for identifying and classifying single cells obtained from clinical samples of human follicular thyroid neoplasms.Approach: We used a method intended to simulate the FNA procedure to obtain single cells from thyroid nodules. A total of 392 hyperspectral Raman images of single cells from follicular thyroid neoplasms were collected.Results: Malignant cells were identified based on their intrinsic Raman spectral signatures with an overall diagnostic accuracy of up to 83.7%.Conclusions: Our findings indicate that hyperspectral Raman microscopy can potentially be developed into an ancillary test for analyzing single cells from thyroid FNA biopsies to better stratify “indeterminate” nodules and other cytologically challenging cases.

Screening naturally mildewed maize kernels based on Raman hyperspectral imaging coupled with machine learning classifiers

AbstractMildewed maize kernels were common in nature and they were harmful to humans and livestocks. Therefore, it is essential to screen the mildewed maize kernels and improve food security. In this study, Raman hyperspectral imaging technique was used to screen the mildewed maize kernels. Hyperspectral imaging of both embryo‐up and embryo‐down side were taken into consideration and the effect of their weight ratio on the screening accuracy was compared. Pixel‐wise approach and object‐wise approach were both analyzed and different variables selection methods were applied to optimize the input variables. Six machine learning classifiers including random forest (RF), support vector machine (SVM), linear discriminant analysis (LDA), extreme learning machine (ELM), partial least squares discriminant analysis (PLSDA), and naive Bayes classifier (NBC) were used to establish the screening models. The results exhibited 19 variables selected by competitive adaptive reweighted sampling (CARS) method from the weight ratio of 3:7 based on object‐wise approach were the most suitable for establishment of screening model with accuracy of 90.63% in the testing set. The microscopic mechanism of mildewed maize kernels was illustrated by micro‐fluorescence imaging. The study showed that Raman hyperspectral imaging combined with both embryo‐up and embryo‐down sides of maize kernels could be used as effective methods to improve the accuracy of screening mildewed maize kernels.Practical ApplicationsThis study provides a comprehensive analysis about the pixels and both sides of maize kernels and gives a nondestructive screening method based on both sides of maize kernels for the mildewed detection. According the mildew of maize kernels, enterprises can effectively improve the storage efficiency of maize kernels to reduce economic losses.

High-temperature phase transformations of hydroxylapatite and the formation of silicocarnotite in the hydroxylapatite–quartz–lime system studied in situ and in operando by Raman spectroscopy

Silica-/calcium phosphate ceramics are of high interest in various aspects. On the one hand, they play an important role in medical applications due to their excellent biocompatibility. Therefore, detailed knowledge of the formation and stability properties of the high-temperature products ensures production under controlled conditions. On the other hand, they were identified as sinter deposits in industrial kilns, where it can indicate problems caused by too high combustion temperatures during the thermal combustion processes. Here, we report the results of two Raman heating studies to ~ 1300 °C in 10 °C-steps with nano-crystalline hydroxylapatite (HAp) and tricalcium phosphate (TCP), and a Raman heating study of natural silicocarnotite (to ~ 1200 °C, 50 °C-steps). The Raman experiments were complemented with thermal analyses. The Raman spectra of nano-crystalline HAp recorded at high temperatures revealed the stepwise loss of adsorbed water and surface-bound OH groups until ~ 570 °C. Significant loss of structural OH started at ~ 770 °C and was completed at ~ 850 °C, when HAp transformed to β-TCP. Between ~ 1220 and ~ 1270 °C, β-TCP was found to transform to α-TCP. The room temperature Raman spectrum of silicocarnotite is characterized by an intense v1(PO4) band at 951 ± 1 cm−1 that shifts to ~ 930 cm−1 at ~ 1200 °C. Using hyperspectral Raman imaging with a micrometer-scale spatial resolution, we were able to monitor in operando and in situ the solid-state reactions in the model system Ca10(PO4)6(OH)2-SiO2-CaO, in particular, the formation of silicocarnotite. In these multi-phase experiments, silicocarnotite was identified at ~ 1150 °C. The results demonstrate that silicocarnotite can form by a reaction between β-TCP and α′L-Ca2SiO4, but also between β-TCP and CaSiO3 with additional formation of quartz.

A New Supramolecular Tetraruthenated Cobalt (II) Porphyrazine Displaying Outstanding Electrocatalytical Performance in Oxygen Evolution Reaction.

A new supramolecular electrocatalyst for Oxygen Evolution Reaction (OER) was synthesized from a central multibridging cobalt tetrapyridylporphyrazine (CoTPyPz) species by attaching four [Ru(bpy)2Cl]+ groups. Both CoTPyPz and the tetraruthenated cobalt porphyrazine species, TRuCoTPyPz, form very homogenous molecular films just by dropcasting their methanol solutions onto GCE electrodes. Such films exhibited low overpotentials for O2 evolution, e.g., 560 e 340 mV, respectively, displaying high stability, typically exceeding 15 h. The kinetic parameters obtained from the Tafel plots showed that the peripheral complexes are very important for the electrocatalytic activity. Hyperspectral Raman images taken along the electrochemical process demonstrated that the cobalt center is the primary active catalyst site, but its performance is enhanced by the ruthenium complexes, which act as electron-donating groups, in the supramolecular system.

A New Alternative Tool to Analyse Glycosylation in Monoclonal Antibodies Based on Drop-Coating Deposition Raman imaging: A Proof of Concept.

Glycosylation is considered a critical quality attribute of therapeutic proteins as it affects their stability, bioactivity, and safety. Hence, the development of analytical methods able to characterize the composition and structure of glycoproteins is crucial. Existing methods are time consuming, expensive, and require significant sample preparation, which can alter the robustness of the analyses. In this context, we developed a fast, direct, and simple drop-coating deposition Raman imaging (DCDR) method combined with multivariate curve resolution alternating least square (MCR-ALS) to analyze glycosylation in monoclonal antibodies (mAbs). A database of hyperspectral Raman imaging data of glycoproteins was built, and the glycoproteins were characterized by LC-FLR-MS as a reference method to determine the composition in glycans and monosaccharides. The DCDR method was used and allowed the separation of excipient and protein by forming a “coffee ring”. MCR-ALS analysis was performed to visualize the distribution of the compounds in the drop and to extract the pure spectral components. Further, the strategy of SVD-truncation was used to select the number of components to resolve by MCR-ALS. Raman spectra were processed by support vector regression (SVR). SVR models showed good predictive performance in terms of RMSECV, R2CV.

Spatiotemporal-Resolved Hyperspectral Raman Imaging of Plasmon-Assisted Reactions at Single Hotspots.

Raman spectroscopy facilitates the study of reacting molecules on single nanomaterials. In recent years, the temporal resolution of Raman spectral measurement has been remarkably reduced to the millisecond level. However, the classic scan-based imaging mode limits the application in the dynamical study of reactions at multiple nanostructures. In this paper, we propose a spatiotemporal-resolved Raman spectroscopy (STRS) technology to achieve fast (∼40 ms) and high spatial resolution (∼300 nm) hyperspectral Raman imaging of single nanostructures. With benefits of the outstanding electromagnetic field enhancement factor by surface plasmon resonance (∼1012) and the snapshot hyperspectral imaging strategy, we demonstrate the observation of stepwise Raman signals from single-particle plasmon-assisted reactions. Results reveal that the reaction kinetics is strongly affected by not only the surface plasmon-polariton generation but also the density of Raman molecules. In consideration of the spatiotemporal resolving capability of STRS, we anticipate that it provides a potential platform for further extending the application of Raman spectroscopy methods in the dynamic study of 1D or 2D nanostructures.

Feedbacks and non-linearity of silicate glass alteration in hyperalkaline solution studied by in operando fluid-cell Raman spectroscopy

Mechanisms, feedbacks and resulting non-linearity during silicate glass alteration in a hyperalkaline carbonate solution were studied through hyperspectral Raman imaging of heated fluid-cells. Our experimental setup enabled in operando visualization and rate measurements of glass dissolution and secondary phase precipitation, complemented by spectral characterization of the phases involved and semi-quantitative monitoring of the ionic strength of the solution close to the glass interface. After initial congruent dissolution of the Ba-bearing soda-lime boroaluminosilicate glass, the formation of a crystalline, saponite-based surface alteration layer (SAL), as well as subsequent zeolite precipitation, witherite coating, and carbonate precipitation within pore spaces of the saponite layer were observed. Two in operando experiments were conducted at ∼ 90 °C for 180 and 260 h that otherwise solely differed in the solution volume (SV) while keeping the surface area constant. The high SV experiment exhibited a transient upward excursion of initial dissolution rates, followed by continuously rapid glass dissolution along with slow SAL growth and sustained oscillations in ionic strength. Contrastingly, in the low SV experiment, glass dissolution monotonically decreased after the onset of rapid SAL growth and no sustained oscillations were observed. We find that growth conditions and resulting properties of the SAL exert dominant, non-linear effects on the evolution of glass dissolution rates. In turn, SAL formation depends on nucleation/growth kinetics and the accumulation of glass-derived solutes at the reaction front. Both, dissolution and precipitation, feedback with solution chemistry and transport processes, together controlling the evolution of the corrosion process. Additionally, fracturing, delamination, and the evolution of surface morphology may affect glass dissolution rates and transport pathways. Such interpretations of decelerating reaction rates in response to the growth of a protective layer are consistent in micro-scale experiments and in outcrop- to global-scale observations, as is the accelerating effect of surface area creation by physical disruption and morphology. Thus, these µm-scale mechanistic insights could help elucidating local to global environmental feedbacks (e.g., erosion or weathering patterns) as well as process dynamics in engineered environments (e.g., nuclear waste disposal) and may assist the improvement predictive models.

Hyperspectral Raman imaging and multivariate statistical analysis for the reconstruction of obliterated serial numbers in polymers

AbstractAs recent years have shown an increasing use of polymers for the fabrication of firearms, it is necessary to develop techniques for the reconstruction of obliterated serial numbers that are stamped in these materials. Hyperspectral Raman imaging has proven to be a suitable technique for this purpose, as it is sensitive to residual strain. The extraction of relevant information however requires an advanced two‐step fitting procedure (i.e., the identification of strain‐sensitive peaks followed by the fitting itself) that may be somewhat time consuming. In this study, principal component analysis (PCA), an exploratory method of the Raman data, is proposed to overcome this deficit. The results show that PCA offers better visual contrast in comparison to the previously reported mathematical modeling technique, as it is able to highlight pertinent variance in the original dataset, for multiple polymers, such as polycarbonate, polyethylene, nylon, and nylatron. Results obtained by limiting acquisition times and spectral ranges have displayed no significant loss of information and therefore reconstruction abilities in polyethylene. A normal density function model and receiver operating characteristic (ROC) curves have been used to show that score matrices obtained from PCA are suitable for separation of distinct strained and unstrained pixel populations. Additionally, binary images favoring minimization of the false positive rate are created to enhance observable contrast allowing for easier character recognition. Finally, a recommended routine analysis is offered to forensic scientists wanting to apply these methods in order to aid criminal investigations or trials.

RamanLIGHT—a graphical user-friendly tool for pre-processing and unmixing hyperspectral Raman spectroscopy images

Raman spectroscopy is a valuable tool for non-destructive vibrational analysis of chemical compounds in various samples. Through 2D scanning, it one can map the chemical surface distribution in a heterogeneous sample. These hyperspectral Raman images typically contain spectra of pure compounds that are hidden within thousands of sum spectra. Inspecting each spectrum to find the pure compounds in the dataset is impractical, and several algorithms have been described in the literature to help analyze such complex datasets. However, choosing the best approach(es) and optimizing the parameters is often difficult, and the necessary software was not yet combined in a single program. Therefore, we introduce RamanLIGHT, a fast and simple app to pre-process Raman mapping datasets and apply up to eight unsupervised unmixing algorithms to find endmember spectra of pure compounds. The user can select from six smoothing methods, four fluorescence baseline-removal methods, four normalization methods, and cosmic-ray and outlier removal to generate a uniform dataset prior to the unmixing. We included the most promising pre-processing methods, since there is no routine that perfectly fits all types of samples. Unmixed endmember spectra can be further used to visualize the distribution of compounds in a sample by creating abundance maps for each endmember separately, or a single labeled image containing all endmembers. It is also possible to create a mean spectrum for each endmember, which better describes the true compound spectrum. We tested RamanLIGHT on three samples: an aspirin-paracetamol-caffeine tablet, Alzheimer’s disease brain tissue and a phase-separated polymer coating. The datasets were pre-processed and unmixed within seconds to gain endmembers of known and unknown chemical compounds. The unmixing algorithms are sensitive to noisy spectra and strong fluorescence backgrounds, so it is important to apply pre-processing methods to a suitable degree. RamanLIGHT is freely available as an MATLAB and soon as standalone app.

Multivariate curve resolution for analysis of Raman hyperspectral imaging data sets for enzyme immobilization

• Raman hyperspectral imaging of an enzyme immobilized onto two resins was accomplished. • Multivariate data analysis spectrally resolved chemical species involved in immobilization. • Chemical images spatially resolved chemical species involved in immobilization. • The Raman hyperspectral imaging data sets can be used for alternative applications. Biocatalytic processes are an essential component in achieving efficient and productive chemical syntheses in the scientific and industrial communities. Immobilization of the biocatalyst offers distinct advantages for the process, including increasing the stability and reusability of the biocatalyst as well as increasing control over the process. Herein, directly engineered pantothenate kinase (PanK) was immobilized to two commercially available microporous resins. Ralbovsky et al. (2021) utilized Raman hyperspectral imaging in combination with multivariate curve resolution alternating least squares (MCR-ALS) to spatially and spectrally resolve the chemical species involved in the enzyme immobilization process. Here, we provide eight Raman hyperspectral imaging data sets collected from the enzyme immobilized to the resin beads under different collection parameters, encompassing over 26,000 raw spectra. A brief discussion of the results, in the form of chemical maps and chemical images, indicates both resins, PanK, the glass substrate, and a Bis-Tris wash solution were resolved using MCR-ALS analysis of a subset of hyperspectral images, thus providing crucial chemical and spatial information about the system. The eight total Raman hyperspectral imaging data sets are shared to provide the scientific community with a thoroughly characterized set of spectral data which can be used for a variety of applications, including for further understanding and optimization of enzyme immobilization systems as well as for individual learning and development applications in the fields of chemical imaging and chemometrics.

Raman hyperspectral imaging coupled to three-dimensional discriminant analysis: Classification of meningiomas brain tumour grades

Meningiomas remains a clinical dilemma. They are the commonest “benign” types of brain tumours and, although being typically benign, they are divided into three WHO grades categories (I, II and III) which are associated with the tumour growth rate and likelihood of recurrence. Recurrence depends on extend of surgery as well as histopathological diagnosis. There is a marked variation amongst surgeons in the follow-up arrangements for their patients even within the same unit which has a significant clinical, and financial implication. Knowing the tumour grade rapidly is an important factor to predict surgical outcomes and adequate patient treatment. Clinical follow up sometimes is haphazard and not based on clear evidence. Spectrochemical techniques are a powerful tool for cancer diagnostics. Raman hyperspectral imaging is able to generate spatially-distributed spectrochemical signatures with great sensitivity. Using this technique, 95 brain tissue samples (66 meningiomas WHO grade I, 24 meningiomas WHO grade II and 5 meningiomas that reoccurred) were analysed in order to discriminate grade I and grade II samples. Newly-developed three-dimensional discriminant analysis algorithms were used to process the hyperspectral imaging data in a 3D fashion. Three-dimensional principal component analysis quadratic discriminant analysis (3D-PCA-QDA) was able to distinguish grade I and grade II meningioma samples with 96% test accuracy (100% sensitivity and 95% specificity). This technique is here shown to be a high-throughput, reagent-free, non-destructive, and can give accurate predictive information regarding the meningioma tumour grade, hence, having enormous clinical potential with regards to being developed for intra-operative real-time assessment of disease.