No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Tumor margin identification and prediction of the primary tumor from brain metastases using FTIR imaging and support vector machines
Abstract
Infrared spectroscopy enables the identification of tissue types based on their inherent vibrational fingerprint without staining in a nondestructive way. Here, Fourier transform infrared microscopic images were collected from 22 brain metastasis tissue sections of bladder carcinoma, lung carcinoma, mamma carcinoma, colon carcinoma, prostate carcinoma and renal cell carcinoma. The scope of this study was to distinguish the infrared spectra of carcinoma from normal tissue and necrosis and to use the infrared spectra of carcinoma to determine the primary tumor of brain metastasis. Data processing follows procedures that have previously been developed for the analysis of Raman images of these samples and includes the unmixing algorithm N-FINDR, segmentation by k-means clustering, and classification by support vector machines (SVMs). Upon comparison with the subsequent hematoxylin and eosin stained tissue sections of training specimens, correct classification rates of the first level SVM were 98.8% for brain tissue, 98.4% for necrosis and 94.4% for carcinoma. The primary tumors were correctly predicted with an overall rate of 98.7% for FTIR images of the training dataset by a second level SVM. Finally, the two level discrimination models were applied to four independent specimens for validation. Although the classification rates are slightly reduced compared to the training specimens, the majority of the infrared spectra of the independent specimens were assigned to the correct primary tumor. The results demonstrate the capability of FTIR imaging to complement histopathological tools for brain tissue diagnosis.
... Additionally, ] (C � O) at 1750 cm −1 is connected to lipids [32,68,69]. e human brain tissue consists of fatty acids and lipids [7], and the formation of gliomas is linked to significant changes in fat metabolism [32]. ...
... erefore, the variation of lipid concentration and composition during tumour genesis also affects the spectroscopic properties tumour. e prominent bands around the amide I and amide II vibrations (1655 cm −1 , 1582 cm −1 , and 1547 cm −1 ) are assigned to proteins and peptides [32,[67][68][69][70][71]]. An example is collagen fibers for the use of brain tumour detection with IR spectroscopy [70]. ...
Characterization of brain tumours requires neuropathological expertise and is generally performed by histological evaluation and molecular analysis. One emerging technique to assist pathologists in future tumour diagnostics is multimodal optical spectroscopy. In the current clinical routine, tissue preprocessing with formalin is widely established and suitable for spectroscopic investigations since degradation processes impede the measurement of native tissue. However, formalin fixation results in alterations of the tissue chemistry and morphology for example by protein cross-linking. As optical spectroscopy is sensitive to these variations, we evaluate the effects of formalin fixation on multimodal brain tumour data in this proof-of-concept study. Nonfixed and formalin-fixed cross sections of different common human brain tumours were subjected to analysis of chemical variations using ultraviolet and Fourier-transform infrared microspectroscopy. Morphological changes were assessed by elastic light scattering microspectroscopy in the visible wavelength range. Data were analysed with multivariate data analysis and compared with histopathology. Tissue type classifications deduced by optical spectroscopy are highly comparable and independent from the preparation and the fixation protocol. However, formalin fixation leads to slightly better classification models due to improved stability of the tissue. As a consequence, spectroscopic methods represent an appropriate additional contrast for chemical and morphological information in neuropathological diagnosis and should be investigated to a greater extent. Furthermore, they can be included in the clinical workflow even after formalin fixation.
... The imaging capabilities of FTIR spectroscopy allow visual classification of tissue samples, with clear differentiation between tissue features, as well as the presence of diseased or cancerous tissue. This approach has proven valuable in exploring prostate tissue, differentiating between benign and malignant epithelium, as well as other tissues including colon, lung and brain tissues [45,[46][47][48]. However, it is worth noting that tissue analysis often requires significant sample preparation steps, particularly when using pathlength dependent FTIR modalitiessuch as transmission or transflection. ...
... The rapid determination of cancerous tissue is useful for determining tumour margins and could be a useful intra-operative aid [48]. However, it is noticeable that FTIR spectroscopy have been relatively limited to ex vivo applications, although novel light sources may extend this technique to other applications [49]. ...
In recent years, the application of vibrational spectroscopy in biomedical research has rapidly expanded; covering aspects of pharmaceutical development, to point-of-care technologies. Vibrational spectroscopy techniques such as Fourier-transform IR (FTIR), and Raman spectroscopy have been at the forefront of this movement, with their complementary information able to shine light onto a range of medical applications. As a relative newcomer to biomedical applications, two-dimensional (2D)-IR is also gaining traction in the field. Here we describe the recent development of these techniques as analytical tools in medical science, and their relative advancements towards the clinic.
... Similar results were found by many other studies. 18,23,61,64,65,[67][68][69][70][71][74][75][76][77] In Figure 2B, absorbance values are clearly larger in High grade spectra than Low grade spectra. Furthermore, spectral differences are mostly apparent in bands attributed to amide I, II, and III and protein regions (1400-1585 cm −1 ), followed by DNA/RNA (O-P-O symmetric stretch) (1080 cm −1 ) and DNA (O-P-O asymmetric stretch) (1230 cm −1 ) regions; RNA Ribose and DNA (C-O stretching) regions (1120-1180 cm −1 ); glycogen (C-O-H bend) (1030 cm −1 ) region; and protein phosphorylation region (970 cm −1 ) ( Table 5 and Figure 6). ...
... Some research has also pointed to high performance of variable reduction and selection coupled to SVM in the classification of cancer. 61,[67][68][69][70][71] In fact, a study by Baker et al 71 confirmed the success of cancer classification by GA-SVM, suggesting to include this algorithm on a standard list of options since this method often provides optimum classification. However, this pilot study has as limitation the fact that all samples were prepared in the same lot; therefore, it does not reflect completely the variability encountered by analyzing different samples over a long-time period, which would happen in the real clinical theater. ...
... Typical imaging methods applied for tumor diagnosis are Fourier-transform infrared spectroscopy [25], fluorescence [26] and Raman imaging [27], which are mainly associated with statistical tools, such as PCA, DA, support vector machine, the k-nearest neighbor algorithm or artificial neural network analysis [25,[28][29][30]. The spectroscopy-based models are used for the identification or distinction of brain [31], colon [32], breast [33] or head and neck tumors [34,35]. ...
Salivary gland tumors (SGTs) are a relevant, highly diverse subgroup of head and neck tumors whose entity determination can be difficult. Confocal Raman imaging in combination with multivariate data analysis may possibly support their correct classification. For the analysis of the translational potential of Raman imaging in SGT determination, a multi-stage evaluation process is necessary. By measuring a sample set of Warthin tumor, pleomorphic adenoma and non-tumor salivary gland tissue, Raman data were obtained and a thorough Raman band analysis was performed. This evaluation revealed highly overlapping Raman patterns with only minor spectral differences. Consequently, a principal component analysis (PCA) was calculated and further combined with a discriminant analysis (DA) to enable the best possible distinction. The PCA-DA model was characterized by accuracy, sensitivity, selectivity and precision values above 90% and validated by predicting model-unknown Raman spectra, of which 93% were classified correctly. Thus, we state our PCA-DA to be suitable for parotid tumor and non-salivary salivary gland tissue discrimination and prediction. For evaluation of the translational potential, further validation steps are necessary.
... IR images were first collected using Fourier transform infrared (FTIR) spectrometers with 64 × 64 or 128 × 128 MCT-based, liquid nitrogen cooled focal plane array detectors. Classification models were trained to determine the tumor grade of glioma brain tumors [54] and the primary tumor of brain metastases [55,56] using FTIR images. Relying on FTIR spectroscopic imaging and computation, stainless computed histopathology enabled a rapid, digital, quantitative, and non-perturbing visualization of morphology and multiple molecular epitopes simultaneously that mimics H&E and various immunohistochemical stains [57]. ...
Modern optical and spectral technologies represent powerful approaches for a molecular characterization of tissues enabling delineating pathological tissues but also a label-free grading and staging of tumors in terms of computer-assisted histopathology. First, currently used tools for intraoperative tumor assessment are described. Next, the requirements for intraoperative tissue visualization from a medical and optical point of view are specified. Then, optical and spectral techniques are introduced that are already approved or close to being used in standard clinical practice for ex vivo and in vivo monitoring, and proof-of concept studies utilizing linear and nonlinear spectroscopy and imaging modalities are presented. Combining several spectroscopic mechanisms in multi-contrast approaches constitutes further advances. Modern artificial intelligence and deep learning concepts have emerged to analyze spectroscopic and imaging datasets and have contributed to the progress of each technique. Finally, an outlook for opportunities and prospects of clinical translation is given.
... Classification using genetic algorithm with LDA on ATR-IR spectra of frozen sections resulted in an accuracy of 79.2% for glioma [32]. Using IR imaging on frozen tissue sections, tissue components were predicted with an accuracy of 85.2% for brain metastases [33] and human glioma grade III and IV were identified with correct rate of 81.7% [10]. For classification of glioma WHO I-III, glioblastoma and control tissue, a success rate of 89% of spectra on 22 samples was obtained [22]. ...
Purpose
Infrared (IR) spectroscopy has the potential for tumor delineation in neurosurgery. Previous research showed that IR spectra of brain tumors are generally characterized by reduced lipid-related and increased protein-related bands. Therefore, we propose the exploitation of these common spectral changes for brain tumor recognition.
Methods
Attenuated total reflection IR spectroscopy was performed on fresh specimens of 790 patients within minutes after resection. Using principal component analysis and linear discriminant analysis, a classification model was developed on a subset of glioblastoma (n = 135) and non-neoplastic brain (n = 27) specimens, and then applied to classify the IR spectra of several types of brain tumors.
Results
The model correctly classified 82% (517/628) of specimens as “tumor” or “non-tumor”, respectively. While the sensitivity was limited for infiltrative glioma, this approach recognized GBM (86%), other types of primary brain tumors (92%) and brain metastases (92%) with high accuracy and all non-tumor samples were correctly identified.
Conclusion
The concept of differentiation of brain tumors from non-tumor brain based on a common spectroscopic tumor signature will accelerate clinical translation of infrared spectroscopy and related technologies. The surgeon could use a single instrument to detect a variety of brain tumor types intraoperatively in future clinical settings. Our data suggests that this would be associated with some risk of missing infiltrative regions or tumors, but not with the risk of removing non-tumor brain.
... Hence, combining vibrational spectroscopy and machine learning is a rapidly growing field with a lot of promise, especially in medical diagnosis [21,22]. A considerably wide range of biological and medical studies have covered FTIR methods in recent years, including the fields of dentistry [23], bone [23,24], DNA [25], and cancer diagnosis, including the tissues of skin [26], brain [27], head, neck [28], breast [29], prostate [30], cervix [31], and digestive system [32][33][34]. FTIR has also been employed in the field of forensic science for identification of trace substances [35], noninvasive detection of latent fingerprints [36,37], biological stain analysis [38,39], postmortem interval estimation [40][41][42], determination of the cause of death [43], identification of injuries [44,45], and so on. ...
Determination of the cause of death for diabetic ketoacidosis (DKA), a common and fatal acute complication of diabetes mellitus, is a challenging forensic task owing to the lack of characteristic morphological findings at autopsy. In this study, Fourier-transform infrared (FTIR) microspectroscopy coupled with chemometrics was employed to characterize biochemical differences in pulmonary edema fluid from different causes of death to supplement conventional methods and provide an efficient postmortem diagnosis of DKA. With this aim, FTIR spectra in three different situations (DKA-caused death, other causes of death with diabetes history, and other causes of death without diabetes history) were measured. The results of principal component analysis indicated different spectral profiles between these three groups, which mainly exhibited variations in proteins. Subsequently, two binary classification models were established using an algorithm of partial least squares discriminant analysis (PLS-DA) to determine whether decedents had diabetes and whether the diabetic patients died from DKA. Satisfactory prediction results of PLS-DA models demonstrated good differentiation among these three groups. Therefore, it is feasible to make a postmortem diagnosis of DKA and detect diabetes history via FTIR microspectroscopic analysis of the pulmonary edema fluid.
... To parse IR data to cardiac pathology knowledge, supervised classification is needed to map the input spectra of each pixel to a desired histologic class. Popular supervised learning techniques are artificial neural networks, 28,29 support vector machines, 30 Bayesian inference methods, 21 among others. In this study, we employed the emerging and powerful methods of deep learning that have proven effective in various image processing applications, including super-resolution, 31 image reconstruction, 32 classification, 33 and estimation of spatial details beyond IR diffraction limit. ...
Context.—
Myocardial fibrosis underpins a number of cardiovascular conditions and is difficult to identify with standard histologic techniques. Challenges include imaging, defining an objective threshold for classifying fibrosis as mild or severe, as well as understanding the molecular basis for these changes.
Objective.—
To develop a novel, rapid, label-free approach to accurately measure and quantify the extent of fibrosis in cardiac tissue using infrared spectroscopic imaging.
Design.—
We performed infrared spectroscopic imaging and combined that with advanced machine learning–based algorithms to assess fibrosis in 15 samples from patients belonging to the following 3 classes: (1) nonpathologic (control) donor hearts; (2) patients receiving transplant; and (3) tissue from patients undergoing implantation of ventricular assist device.
Results.—
Our results show excellent sensitivity and accuracy for detecting myocardial fibrosis as demonstrated by high area under the curve of 0.998 in the receiver-operating characteristic curve measured from infrared imaging. Fibrosis of various morphologic subtypes are then demonstrated with virtually generated picrosirius red images, which show good visual and quantitative agreement (correlation coefficient = 0.92, ρ = 7.76 × 10−15) with stained images of the same sections. Underlying molecular composition of the different subtypes were investigated with infrared spectra showing reproducible differences presumably arising from differences in collagen subtypes and/or crosslinking.
Conclusions.—
Infrared imaging can be a powerful tool in studying myocardial fibrosis and gleaning insights into the underlying chemical changes that accompany it. Emerging methods suggest that the proposed approach is compatible with conventional optical microscopy and its consistency makes it translatable to the clinical setting for real-time diagnoses as well as for objective and quantitative research.
... Renal cell carcinoma (RCC) is not a single organism but rather a group of tumors that arise from the highly heterogeneous epithelium of kidney tissue [1][2][3][4][5]. According to the Heidelberg description of renal cell cancer, histopathological subtypes of RCC includes clear cell adenocarcinoma, the most prevalent form of RCC, chromophobe accumulating duct carcinoma, papillary carcinoma, and unclassi ed carcinoma [6][7][8][9][10]. Among urological tumors, RCC has the highest risk of cancer-speci c mortality and the 5-year survival rate for patients with metastatic disease is just 12%. ...
Renal cell carcinoma (RCC) is a widespread type of urological tumor that derives from the highly heterogeneous epithelium of the kidney tissue. For the past decade, the treatment of kidney cancer cells has changed clinical care for RCC. Herein, we present a very easy and cost-effective method that incorporates tumor-specific targeting supramolecular nanoassembly, and therapeutically to overcome the different challenges raised by the distribution of the pharmaceutical potential anticancer drug Cisplatin (CIS-PT). On covalent conjugations of hydrophobic linoleic acid by carboxylic group, the CIS-PT prodrugs were skilled in impulsively nanoassembly into extremely steady nanoparticles size (~100 nm). Electron microscopic techniques have verified the newly synthesized morphology of CIS-PT-NPs. The anticancer properties of CIS-PT and CIS-PT-NPs against Caki-1 and A-498 (renal carcinoma) cancer cell lines have been evaluated after successful synthesis. Other research, such as dual staining acridine orange/ethidium bromide, Hoechst 33344 and flow cytometry study on the apoptosis mechanisms, have shown that proliferation in renal cancer cells is associated with apoptosis. Further the In vivo toxicity results displays the CIS-PT-NPs remarkably alleviated the toxicity of the potential anticancer drug CIS-PT In vivo while conserving the Pharmaceutical activity. Compared to CIS-PT, CIS-PT-NPs demonstrate excellent In vitro and In vivo property, this study clarified the CIS-PT-NPs as a healthy and positive RCC care chemotherapeutics technique and deserve further clinical evaluations.
... Support vector machine (SVM) has been widely used in FTIR analysis of cancer tissues. 21,22 It can easily deal with the limited number of samples and large number of data. Finally, different definitions of protein secondary structures have been proposed in the course of years, and DSSP appears to be best related to FTIR spectral features. ...
The paper introduces a new method designed for high-throughput protein structure determination. It is based on spotting proteins as microarrays at a density of ca. 2000-4000 samples per cm2 and recording Fourier transform infrared (FTIR) spectra by FTIR imaging. It also introduces a new protein library, called cSP92, which contains 92 well-characterized proteins. It has been designed to cover as well as possible the structural space, both in terms of secondary structures and higher level structures. Ascending stepwise linear regression (ASLR), partial least square (PLS) regression, and support vector machine (SVM) have been used to correlate spectral characteristics to secondary structure features. ASLR generally provides better results than PLS and SVM. The observation that secondary structure prediction is as good for protein microarray spectra as for the reference attenuated total reflection spectra recorded on the same samples validates the high throughput microarray approach. Repeated double cross-validation shows that the approach is suitable for the high accuracy determination of the protein secondary structure with root mean square standard error in the cross-validation of 4.9 ± 1.1% for α-helix, 4.6 ± 0.8% for β-sheet, and 6.3 ± 2.2% for the "other" structures when using ASLR.
... Imaging technologies exploiting the vibrational transition of biomolecules, 1,2 such as Fourier transform infrared (FTIR) spectroscopy 3 and stimulated Raman scattering (SRS) imaging, 4 provide rich and specific information about tissue's biochemical composition in a label-free manner. [5][6][7][8][9][10] Because of their ability to classify biomolecules (e.g., glycogen, proteins, lipids, or nucleic acids), 11,12 these technologies have many biomedical applications, such as analyzing clinical biopsy samples ex vivo [13][14][15][16] and studying disease progression with tissue sections taken from animal models. 17,18 However, their broader translation is still inhibited by certain intrinsic limitations. ...
Significance:
Mid-infrared (IR) imaging based on the vibrational transition of biomolecules provides good chemical-specific contrast in label-free imaging of biology tissues, making it a popular tool in both biomedical studies and clinical applications. However, the current technology typically requires thin and dried or extremely flat samples, whose complicated processing limits this technology's broader translation.
Aim:
To address this issue, we report mid-IR photoacoustic microscopy (PAM), which can readily work with fresh and thick tissue samples, even when they have rough surfaces.
Approach:
We developed a transmission-mode mid-IR PAM system employing an optical parametric oscillation laser operating in the wavelength range from 2.5 to 12 μm. Due to its high sensitivity to optical absorption and the low ultrasonic attenuation of tissue, our PAM achieved greater probing depth than Fourier transform IR spectroscopy, thus enabling imaging fresh and thick tissue samples with rough surfaces.
Results:
In our spectroscopy study, the CH2 symmetric stretching at 2850 cm - 1 (3508 nm) was found to be an excellent source of endogenous contrast for lipids. At this wavenumber, we demonstrated label-free imaging of the lipid composition in fresh, manually cut, and unprocessed tissue sections of up to 3-mm thickness.
Conclusions:
Our technology requires no time-consuming sample preparation procedure and has great potential in both fast clinical histological analysis and fundamental biological studies.
... Machine learning is frequently applied to spectroscopy and spectroscopic imaging to differentiate distinct tissue types [52][53][54][55][56][57][58][59][60][61][62][63][64][65][66][67][68][69], while recent methods attempt to map molecular spectra to conventional stains [70,71] to produce images that can be interpreted by pathologists without additional training. ...
Deep neural networks are emerging as a popular choice for hyperspectral image analysis—compared with other machine learning approaches, they are more effective for a variety of applications in hyperspectral imaging. Part I (Chap. 3) introduces the fundamentals of deep learning algorithms and techniques deployed with hyperspectral images. In this chapter (Part II), we focus on application-specific nuances and design choices with respect to deploying such networks for robust analysis of hyperspectral images. We provide quantitative and qualitative results with a variety of deep learning architectures, and compare their performance to baseline state-of-the-art methods for both remote sensing and biomedical image analysis tasks. In addition to surveying recent developments in these areas, our goal in these two chapters is to provide guidance on how to utilize such algorithms for multichannel optical imagery. With that goal, we also provide code and example datasets used in this chapter.
... Only three specific cases are presented here, but we can see the similarity in all solid tumors. Previous studies also used HSI as a diagnostic tool in cancers of the prostate (34,(65)(66)(67), lung (43,(68)(69)(70), cervix (71), colon (39,44,67,(72)(73)(74)(75)(76), kidney (77), brain (78,79), skin (80,81), oral cavity (82), and bladder (83), etc. ...
Hyperspectral imaging (HSI) is an emerging new technology in solid tumor diagnosis and detection. It incorporates traditional imaging and spectroscopy together to obtain both spatial and spectral information from tissues simultaneously in a non-invasive manner. This imaging modality is based on the principle that different tissues inherit different spectral reflectance responses that present as unique spectral fingerprints. HSI captures those composition-specific fingerprints to identify cancerous and normal tissues. It becomes a promising tool for performing tumor diagnosis and detection from the label-free histopathological examination to real-time intraoperative assistance. This review introduces the basic principles of HSI and summarizes its methodology and recent advances in solid tumor detection. In particular, the advantages of HSI applied to solid tumors are highlighted to show its potential for clinical use.
... In fact, approximately 10% of pathologic evaluation could not result in a firm diagnosis because either certain tumors are histologically similar or the tissue of origin could not be identified from the poorly differentiated cells [4]. The method also involves the complex process of histochemical staining techniques for the tissue samples, whereby the most commonly used hematoxylin and eosin (H&E) dyes are non-specific for cancer [5]. ...
Infrared spectroscopy has long been used to characterize chemical compounds, but the applicability of this technique to the analysis of biological materials containing highly complex chemical components is arguable. However, recent advances in the development of infrared spectroscopy have significantly enhanced the capacity of this technique in analyzing various types of biological specimens. Consequently, there is an increased number of studies investigating the application of infrared spectroscopy in screening and diagnosis of various diseases. The lack of highly sensitive and specific methods for early detection of cancer has warranted the search for novel approaches. Being more simple, rapid, accurate, inexpensive, non-destructive and suitable for automation compared to existing screening, diagnosis, management and monitoring methods, Fourier transform infrared spectroscopy can potentially improve clinical decision-making and patient outcomes by detecting biochemical changes in cancer patients at the molecular level. Besides the commonly analyzed blood and tissue samples, extracellular vesicle-based method has been gaining popularity as a non-invasive approach. Therefore, infrared spectroscopic analysis of extracellular vesicles could be a useful technique in the future for biomedical applications. In this review, we discuss the potential clinical applications of Fourier transform infrared spectroscopic analysis using various types of biological materials for cancer. Additionally, the rationale and advantages of using extracellular vesicles in the spectroscopic analysis for cancer diagnostics are discussed. Furthermore, we highlight the challenges and future directions of clinical translation of the technique for cancer.
... Supervised machine learning (ML) approaches have been successfully applied to circumvent the problem of explicitly and analytically describing the specific segmentation procedure and related parameters, lying to a learning stage the charge of inducing the classifier from supervised data available. The proposed techniques make use of a single image or multispectral pattern and are interactive or fully automatic [3,[10][11][12][13][14][15][16]. Among the most promising methods, we found the support vector machine (SVM) [17,18], discriminative models based on Random Forest and logistic regression [19,20]. ...
This work aimed to investigate whether automated classifiers belonging to feature-based and deep learning may approach brain metastases segmentation successfully. Support Vector Machine and V-Net Convolutional Neural Network are selected as representatives of the two approaches. In the experiments, we consider several configurations of the two methods to segment brain metastases on contrast-enhanced T1-weighted magnetic resonance images. Performances were evaluated and compared under critical conditions imposed by the clinical radiotherapy domain, using in-house dataset and public dataset created for the Multimodal Brain Tumour Image Segmentation (BraTS) challenge. Our results showed that the feature-based and the deep network approaches are promising for the segmentation of Magnetic Resonance Imaging (MRI) brain metastases achieving both an acceptable level of performance. Experimental results also highlight different behaviour between the two methods. Support vector machine (SVM) improves performance with a smaller training set, but it is unable to manage a high level of heterogeneity in the data and requires post-processing refinement stages. The V-Net model shows good performances when trained on multiple heterogeneous cases but requires data augmentations and transfer learning procedures to optimise its behaviour. The paper illustrates a software package implementing an integrated set of procedures for active support in segmenting brain metastases within the radiotherapy workflow.
... The positions of the bands and their intensity and shape carry detailed information about the molecular composition of the specimen [13]. Previous studies have demonstrated the versatile use of vibrational spectroscopy to analyse lung [14], cervix [15], brain [16][17][18], breast [19], prostate [20] and separate diseased specimens from normal, identification of tumour margins and correlation of spectra to tumour grade. Brain tumour studies have been carried out in a number of areas which including spectro-imaging of collagen scaffolds during glioma tumour development [21], separation of normal and neoplastic tissue using both Raman and IR [17] amongst others. ...
Diagnosis of a brain tumour is not always a straight forward process with the signs and symptoms overlapping with other neurological diseases which can mimic brain tumours on neuroimaging and histological examination. The current techniques used lack the desired level of precision in both diagnosis and cytoreductive surgery. Classification of tumours such as gliomas which are on a continuous spectrum of histology and malignancy into distinct categories is still a challenge using histopathology. Histopathology is the gold standard and studies have shown it to be subjective with pathological discrepancies reported to be over 45 % during first histological diagnoses and modified at second reading resulting in different treatment decisions. This study evaluated the diagnostic application of Raman spectroscopy to distinguish between gliomas and normal brain tissue as well as degree of malignancy based on two groups' high grade and low grade gliomas. Formalin fixed paraffin embedded (FFPE) tissue blocks of both gliomas and normal brain were sectioned at 10 microns and mounted on low-E slides, dewaxed using Xylene, washed with alcohol and water before storage at room temperature until analysis. Raman spectroscopy coupled with multivariate statistics was able to distinguish between normal brain tissue and gliomas. Classification of gliomas based on degree of malignancy was also apparent with very high classification accuracy above 90 %. We present for the first time spectral panels that can be used as spectral biomarkers in the diagnosis of gliomas and their degree of malignancy.
... Results showed tumor primary site could be delineated; however, there was an overlap between breast, lung, and colorectal carcinomas. A later study by the same group, again using imaging methods but a broader range of cancers, also demonstrated similar overlap within the adenocarcinomas (Bergner et al. 2013). Given the relatively similar morphological appearances and IHC staining results overlaps, this is not surprising. ...
Metastatic brain tumors represent a significant proportion of tumors identified intraoperatively. A rapid diagnostic method, circumventing the need for histopathology studies, could prove clinically useful. As many spectroscopic studies have shown ability to differentitate between different tumor types, this technique was evaluated for use within metastatic brain tumors. Spectrochemical approaches [Raman and attenuated total reflection Fourier-transform infrared spectroscopy (ATR-FTIR) spectroscopy] were applied to determine how readily they may identify the primary site for the metastatic tumor. Metastases were from primary adenocarcinomas of lung (n = 7) and colorectum (n = 7), and for comparison, metastatic melanoma (n = 7). The objective was to determine if Raman or ATR-FTIR spectroscopy could delineate the origin of the primary tumor. The results demonstrate that there are marked similarities between the two adenocarcinoma groups and whilst Raman and ATR-FTIR can distinguish the three groups with limited success, classification accuracy is greatly improved when combining the adenocarcinoma groups. The use of such techniques in the clinical setting is more likely to be found intraoperatively, determining the presence of a tumor and suggesting the tumor class; however, traditional histopathology would still be needed to identify the primary origin of the tumor.
... There are several reviews and studies [4,12e14] in the literature on the application of supervised analysis and IR imaging in the diagnosis and characterization of tissue. For example, Lasch et al. [15] have employed Artificial Neural Networks (ANN) to discriminate histological structures of human colorectal adenocarcinoma, Wald and Goormaghtigh [16] used SVMs for identifying the main cell types present in melanoma tumours, Bergner et al. [17] employed PLS-DA for predicting brain metastasis and Pilling and Gardner [18] have used Random Forest algorithms to diagnose breast cancer using a Quantum Cascade Laser (QCL) source to generate discrete wavenumber value images. ...
Infrared (IR) imaging is an emerging and powerful approach for studying the molecular composition of cells and tissues. It is a non-destructive and phenotypic technique which combines label-free molecular specific information from cells and tissues provided by IR with spatial resolution, offering great potential in biochemical and biomedical research and routine applications. The application of multivariate discriminant analysis using bilinear models such as Partial Least Squares-Discriminant Analysis (PLS-DA) to IR images requires to unfold the spatial directions in a two-way matrix, resulting in a loss of spatial information and structure. In this article, first we evidence that that internal validation methods such as repeated k-fold cross-validation (CV) can be overly optimistic when the pixel size of the image is lower than the lateral spatial resolution. Secondly, we propose a new approach for the unbiased internal evaluation of the model performance named COnstrained Repeated Random Subsampling–Cross Validation (CORRS-CV). This method is based on the generation of q training and test sub-sets using a constrained random sampling of n training pixels without replacement and it circumvents overly optimistic effects due to oversampling, providing more accurate and robust images. The approach can be applied in IR microscopy for the development of discriminant models to analyse underlying biochemical differences associated to anatomical and histopathological features in cells and tissues.
... The study of the bio-molecular alterations in the biochemistry of retinal layers with different duration of diabetes would be greatly facilitated by vibrational spectroscopic techniques without any need for isolating the cell compartments. FT-IR chemical imaging has been used to investigate various tissue types including adipose 23 , prostate 24 , breast 25,26 , kidney 27 , brain 28,29 , pancreas 30 , pleura 31 , and colon 32 . ...
To discover the mechanisms underlying the progression of diabetic retinopathy (DR), a more comprehensive understanding of the biomolecular processes in individual retinal cells subjected to hyperglycemia is required. Despite extensive studies, the changes in the biochemistry of retinal layers during the development of DR are not well known. In this study, we aimed to determine a more detailed understanding of the natural history of DR in Akita/+ (type 1 diabetes model) male mice with different duration of diabetes. Employing label-free spatially resolved Fourier transform infrared (FT-IR) chemical imaging engaged with multivariate analysis enabled us to identify temporal-dependent reproducible biomarkers of the individual retinal layers from mice with 6 weeks,12 weeks, 6 months, and 10 months of age. We report, for the first time, the nature of the biochemical alterations over time in the biochemistry of distinctive retinal layers namely photoreceptor retinal layer (PRL), inner nuclear layer (INL), and plexiform layers (OPL, IPL). Moreover, we present the molecular factors associated with the changes in the protein structure and cellular lipids of retinal layers induced by different duration of diabetes. Our paradigm provides a new conceptual framework for a better understanding of the temporal cellular changes underlying the progression of DR.
... The greatest benefit of this technique lies in the high molecular sensitivity, which enables the researcher to potentially detect molecular changes that precede any morphological change, thus enabling earlier diagnosis of tissue dysplasia [15]. Currently, the FTIR microspectroscopy technique is widely used for studying cancer diagnosis, such as colon [16], skin [17], cervical [18], breast [19,20], brain [21], and prostate [22] tumors. Another advantage of this technique is the ability to offer rapid, labelfree, and non-destructive measurements of the samples [15]. ...
Anaphylaxis is a rapid allergic reaction that may cause sudden death. Currently, postmortem diagnosis of anaphylactic shock is sometimes difficult and often achieved through exclusion. The aim of our study was to investigate whether Fourier transform infrared (FTIR) microspectroscopy combined with pattern recognition methods would be complementary to traditional methods and provide a more accurate postmortem diagnosis of fatal anaphylactic shock. First, the results of spectral peak area analysis showed that the pulmonary edema fluid of the fatal anaphylactic shock group was richer in protein components than the control group, which included mechanical asphyxia, brain injury, and acute cardiac death. Subsequently, principle component analysis (PCA) was performed and showed that the anaphylactic shock group contained more turn and α-helix protein structures as well as less tyrosine-rich proteins than the control group. Ultimately, a partial least-square discriminant analysis (PLS-DA) model combined with a variables selection method called the genetic algorithm (GA) was built and demonstrated good separation between these two groups. This pilot study demonstrates that FTIR microspectroscopy has the potential to be an effective aid for postmortem diagnosis of fatal anaphylactic shock.
... Learning-based methods consider segmentation tasks as classification problems and require a certain number of expert-segmented images to train classification models. On those expert-segmented images, manually-designed image features, such as mean, standard deviation, gray level co-occurrence matrix (GLCM), and local binary pattern features (LBP), are extracted and fed into machine learning models, such as support vector machine (SVM) or artificial neural network (ANN), to classify target abnormalities [12][13][14][15][16][17]. Automatic brain metastases segmentation requires special considerations in its clinical implementation. ...
Accurate and automatic brain metastases target delineation is a key step for efficient and effective stereotactic radiosurgery (SRS) treatment planning. In this work, we developed a deep learning convolutional neural network (CNN) algorithm for segmenting brain metastases on contrast-enhanced T1-weighted magnetic resonance imaging (MRI) datasets. We integrated the CNN-based algorithm into an automatic brain metastases segmentation workflow and validated on both Multimodal Brain Tumor Image Segmentation challenge (BRATS) data and clinical patients' data. Validation on BRATS data yielded average DICE coefficients (DCs) of 0.75±0.07 in the tumor core and 0.81±0.04 in the enhancing tumor, which outperformed most techniques in the 2015 BRATS challenge. Segmentation results of patient cases showed an average of DCs 0.67±0.03 and achieved an area under the receiver operating characteristic curve of 0.98±0.01. The developed automatic segmentation strategy surpasses current benchmark levels and offers a promising tool for SRS treatment planning for multiple brain metastases.
... date, oncology is the largest field of application to be explored by VS. Several types of cancer including melanoma and 170 brain metastasis, 171 as well as breast, 172 colon, 173 and lung cancer, 174 can be diagnosed by VS. In kidney research, VS has been used to evaluate renal cancers, which represent almost 2% of cancers worldwide. ...
Correspondence: Richard M. Caprioli, Mass Spectrometry Research Center, Department of Biochemistry, 9160 MRB III, Vanderbilt University, Nashville, Tennessee 37232, USA.
... Bergner et. al. [111] did predict of primary brain tumor metastasis. Here they worked on 22 brain metastasis tissue sections and in comparison, with H&E stained tissue sections, the correct classification rates 98.8% for brain tissue, 98.4% for necrosis and 94.4% for carcinoma were achieved. ...
Background:
Cancer is a major global health issue. It causes extensive individual suffering and gives a huge burden on the health care in society. Despite extensive research and different tools have been developed it still remains a challenge for early detection of this disease. FTIR imaging has been used to diagnose and differentiate the molecular differences between normal and diseased tissues.
Methods:
Fourier Transform Infrared Spectroscopy (FTIR) is able to measure biochemical changes in tissue, cell and biofluids based on the vibrational signature of their components. This technique enables to the distribution and structure of lipids, proteins, nucleic acids as well as other metabolites. These differences depended on the type and the grade of cancer.
Results:
We emphasize here, that the FTIR spectroscopy and imaging can be considered as a promising technique and will find its place on the detection of this dreadful disease because of high sensitivity, accuracy and inexpensive technique. Now the medical community started using and accepting this technique for early stage cancer detection. We discussed this technique and the several challenges in its application for the diagnosis of cancer in regards of sample preparations, data interpretation, and data analysis. The sensitivity of chemotherapy drugs on individual specific has also discussed.
Conclusion:
So far progressed has done with the FTIR imaging in understanding of cancer disease pathology. However, more research is needed in this field and it is necessary to understand the morphology and biology of the sample before using the spectroscopy and imaging because invaluable information to be figured out.
Background: In brain tumor surgery, maximal tumor resection is typically desired. This is complicated by infiltrative tumor cells which cannot be visually distinguished from healthy brain tissue. Optical methods are an emerging field that can potentially revolutionize brain tumor surgery through intraoperative differentiation between healthy and tumor tissues. Methods: This study aimed to systematically explore and summarize the existing literature on the use of Raman Spectroscopy (RS), Hyperspectral Imaging (HSI), Optical Coherence Tomography (OCT), and Diffuse Reflectance Spectroscopy (DRS) for brain tumor detection. MEDLINE, Embase, and Web of Science were searched for studies evaluating the accuracy of these systems for brain tumor detection. Outcome measures included accuracy, sensitivity, and specificity. Results: In total, 44 studies were included, covering a range of tumor types and technologies. Accuracy metrics in the studies ranged between 54 and 100% for RS, 69 and 99% for HSI, 82 and 99% for OCT, and 42 and 100% for DRS. Conclusions: This review provides insightful evidence on the use of optical methods in distinguishing tumor from healthy brain tissue.
Cervical cancer was considered the fourth most common cancer worldwide in 2020. In order to reduce mortality, an early diagnosis of the tumor is required. Currently, this type of cancer occurs mostly in developing countries due to the lack of vaccination and screening against the Human Papillomavirus. Thus, there is an urgent clinical need for new methods aiming at a reliable screening and an early diagnosis of precancerous and cancerous cervical lesions. Vibrational spectroscopy has provided very good results regarding the diagnosis of various tumors, particularly using Fourier transform infrared microspectroscopy, which has proved to be a promising complement to the currently used histopathological methods of cancer diagnosis. This spectroscopic technique was applied to the analysis of cryopreserved human cervical tissue samples, both squamous cell carcinoma (SCC) and non-cancer samples. A dedicated Support Vector Machine classification model was constructed in order to categorize the samples into either normal or malignant and was subsequently validated by cross-validation, with an accuracy higher than 90%.
Purpose
Infrared (IR) spectroscopy has the potential for tumor delineation in neurosurgery. Previous research showed that IR spectra of brain tumors are generally characterized by reduced lipid-related and increased protein-related bands. Therefore, we propose the exploitation of these common spectral changes for brain tumor recognition.
Methods
Attenuated total reflection IR spectroscopy was performed on fresh specimens of 790 patients within minutes after resection. Using principal component analysis and linear discriminant analysis, a classification model was developed on a subset of glioblastoma (n = 135) and non-neoplastic brain (n = 27) specimens, and then applied to classify the IR spectra of several types of brain tumors.
Results
The model correctly classified 82% (517/628) of specimens as “tumor” or “non-tumor”, respectively. While the sensitivity was limited for infiltrative glioma, this approach recognized GBM (86%), other types of primary brain tumors (92%) and brain metastases (92%) with high accuracy and all non-tumor samples were correctly identified.
Conclusion
The concept of differentiation of brain tumors from non-tumor brain based on a common spectroscopic tumor signature will accelerate clinical translation of infrared spectroscopy and related technologies. The surgeon could use a single instrument to detect a variety of brain tumor types intraoperatively in future clinical settings. Our data suggests that this would be associated with some risk of missing infiltrative regions or tumors, but not with the risk of removing non-tumor brain.
Brain metastases are the most prevalent intracranial neoplasm that causes excessive morbidity and mortality in most cancer patients. The current medical model for brain metastases is focused on the physical condition of the affected individual, the anatomy of the main tumor, and the number and proximity of brain lesions. In this paper, a new hybrid Metastases Fast Fourier Transformation with SVM (MFFT-SVM) method is proposed that can classify high dimensional magnetic resonance imaging as tumor and predicts lung cancer from given protein primary sequences. The goal is to address the associated issues stated with the treatment targeted at unique molecular pathways to the tumor, together with those involved in crossing the blood-brain barrier and migrating cells to the lungs. The proposed method identifies the place of the lung damage by the Fast Fourier Technique (FFT). FFT is the principal statistical approach for frequency analysis which has many engineering and scientific uses. Moreover, Differential Fourier Transformation (DFT) is considered for focusing the brain metastases that migrate into the lungs and create non-small lungs cancer. However, Support Vector Machine (SVM) is used to measure the accuracy of control patient's datasets of sensitivity and specificity. The simulation results verified the performance of the proposed method is improved by 92.8% sensitivity, of 93.2% specificity and 95.5% accuracy respectively.
Aristolochic acid is a potent carcinogenic and nephrotoxic chemical found in all Aristolochia plants and the toxicity issue of aristolochic acid has gained increasing attention in recent years. To better understand the toxicity of aristolochic acid, we here explored to use Fourier transform infrared (FTIR) spectroscopy combined with chemometric analysis to investigate the biochemical changes of the liver and kidney tissues as well as plasma of a rat model due to aristolochic acid poisoning. We first collected the FTIR spectra of the plasma and tissue samples from the treatment group and the control group of the model rats. We then performed the chemometric analysis of these FTIR spectra using principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA). The interesting discovery was that when there was no distinctive change in the plasma samples due to aristolochic acid poisoning, there were already spectral changes in the liver and kidney samples which could be effectively distinguished by PCA and PLS-DA. Moreover, our chemometric analysis indicated that the aristolochic acid-induced damages on the liver tissue of model rats featured with lipid metabolism changes, while the biochemical changes caused by aristolochic acid on the kidney tissue seemed to be more complex. To the best of our knowledge, this work is the first application of FTIR spectroscopy combined with chemometric analysis in the investigation of the toxicity of aristolochic acid.
Head and neck tumors can be very challenging to treat because of the risk of problems or complications after surgery. Therefore, prompt and accurate diagnosis is extremely important to drive appropriate treatment decisions, which may reduce the chance of recurrence. This paper presents the original research exploring the feasibility of Fourier transform infrared (FT-IR) and Raman spectroscopy (RS) methods to investigate biochemical alterations upon the development of the pleomorphic adenoma. Principal component analysis (PCA) was used for a detailed assessment of the observed changes and to determine the spectroscopic basis for salivary gland neoplastic pathogenesis. It is implied that within the healthy margin, as opposed to the tumoral tissue, there are parts that differ significantly in lipid content. This observation shed new light on the crucial role of lipids in tissue physiology and tumorigenesis. Thus, a novel approach that eliminates the influence of lipids on the elucidation of biochemical changes is proposed. The performed analysis suggests that the highly heterogeneous healthy margin contains more unsaturated triacylglycerols, while the tumoral section is rich in proteins. The difference in protein content was also observed for these two tissue types, i.e. the healthy tissue possesses more proteins in the anti-parallel β-sheet conformation, whereas the tumoral tissue is dominated by proteins rich in unordered random coils. Furthermore, the pathogenic tissue shows a higher content of carbohydrates and reveals noticeable differences in nucleic acid content. Finally, FT-IR and Raman spectroscopy methods were proposed as very promising methods in the discrimination of tumoral and healthy tissues of the salivary gland.
Renal cell carcinoma (RCC) is a widespread type of urological tumor that derives from the highly heterogeneous epithelium of the kidney tissue. For the past decade, the treatment of kidney cancer cells has changed clinical care for RCC. Herein, we present a very easy and cost-effective method that incorporates tumor-specific targeting supramolecular nanoassembly, and therapeutically to overcome the different challenges raised by the distribution of the pharmaceutical potential anticancer drug Cisplatin (CIS-PT). On covalent conjugations of hydrophobic linoleic acid by carboxylic group, the CIS-PT prodrugs were skilled in impulsively nanoassembly into extremely steady nanoparticles size (∼100 nm). Electron microscopic techniques have verified the newly synthesized morphology of CIS-PT-NPs. The anticancer properties of CIS-PT and CIS-PT-NPs against Caki-1 and A-498 (renal carcinoma) cancer cell lines have been evaluated after successful synthesis. Other research, such as dual staining acridine orange/ethidium bromide (AO-EB), Hoechst 33344 and flow cytometry study on the apoptosis mechanisms, have shown that proliferation in renal cancer cells is associated with apoptosis. Further the In vivo toxicity results displays the CIS-PT-NPs remarkably alleviated the toxicity of the potential anticancer drug CIS-PT In vivo while conserving the Pharmaceutical activity. Compared to CIS-PT, CIS-PT-NPs demonstrate excellent In vitro and In vivo property, this study clarified the CIS-PT-NPs as a healthy and positive RCC care chemotherapeutics technique and deserve further clinical evaluations.
Carcinogenesis is a multifaceted process of cancer formation. The transformation of normal cells into cancerous ones may be difficult to determine at a very early stage. Therefore, methods enabling identification of initial changes caused by cancer require novel approaches. Although physical spectroscopic methods such as FT-Raman and Fourier Transform InfraRed (FTIR) are used to detect chemical changes in cancer tissues, their potential has not been investigated with respect to carcinogenesis. The study aimed to evaluate the usefulness of FT-Raman and FTIR spectroscopy as diagnostic methods of endometrial cancer carcinogenesis. The results indicated development of endometrial cancer was accompanied with chemical changes in nucleic acid, amide I and lipids in Raman spectra. FTIR spectra showed that tissues with development of carcinogenesis were characterized by changes in carbohydrates and amides vibrations. Principal component analysis and hierarchical cluster analysis of Raman spectra demonstrated similarity of tissues with cancer cells and lesions considered precursor of cancer (complex atypical hyperplasia), however they differed from the control samples. Pearson correlation test showed correlation between cancer and complex atypical hyperplasia tissues and between non-cancerous tissue samples. The results of the study indicate that Raman spectroscopy is more effective in assessing the development of carcinogenesis in endometrial cancer than FTIR.
Raman spectroscopy is a powerful bioanalytical method with high molecular specificity. The relatively low sensitivity of spontaneous Raman scattering can be overcome by signal enhancement effects such as surface enhanced Raman scattering and coherent Raman scattering, which is the collective term for coherent anti-Stokes Raman scattering and stimulated Raman scattering. Applications of Raman-based approaches for cancer diagnostics are summarized in the first part of the chapter. In the second part, complementary optical methods including autofluorescence, optical coherence tomography, second harmonic generation, and two-photon excited fluorescence were combined with Raman scattering as multi-contrast modalities. After training of multivariate chemometric classification models, Raman spectroscopy with or without other modalities has been used for pathological screening of cancer cells and tissues under ex vivo and in vivo conditions. Each modality is briefly introduced and typical examples are described.
The identification of muscle hemorrhage in a cadaver that is in an advanced stage of decomposition is an important but challenging task. Our study investigated whether Fourier transform infrared (FT-IR) microspectroscopy in conjunction with chemometrics could identify muscle hemorrhage using rat cadavers with advanced decomposition. In this study, an intramuscular blood injection method, instead of a mechanical injury method, was used to construct a muscle hemorrhage model, and the modeling idea of muscle hemorrhage identification was to discriminate and classify hemoglobin-leaking myofibrils from negative myofibrils. First, the optical images of hematoxylin/eosin (H&E) stained hemorrhagic muscle at different postmortem intervals (PMIs) were observed and showed that the morphological features of whole erythrocytes disappeared since the PMI of 4 d. Subsequently, principle component analysis (PCA) was performed and indicated that the biochemical differences in protein structures between fresh erythrocytes and myofibrils can be detected by the IR spectroscopic method. Ultimately, several classification models based on the partial least square discriminant analysis (PLS-DA) algorithm were successfully constructed for different PMIs and PMI ranges and achieved great prediction performances in external validations. This preliminary study demonstrates the feasibility of using FT-IR microspectroscopy combined with chemometrics as a potential approach for identifying muscle hemorrhage in cadavers with advanced decomposition for offering vital evidences in judicial process.
Lungs, due to their high oxygen availability and vascularization, are an ideal environment for cancer cell migration, metastasis and tumour formation. These processes are directly connected with extracellular matrix (ECM) remodelling, resulting from cancer cell infiltration and preparation of the environment suitable for tumour growth. Herein, we compare the potential of fast, label-free and non-destructive methods of Fourier-transform infrared spectroscopy (FT-IR) in standard and high definition (HD) modes with nonlinear coherent anti-Stokes Raman scattering (CARS), second harmonic generation (SHG), two-photon excited fluorescence (TPEF) and a fluorescence lifetime imaging (FLIM) technique for lung metastasis detection. We show their potential in the detection of lung macrometastasis, in which we already observed the ECM remodelling. The CARS image revealed a dense cell fraction typical of ECM remodeling and reduction of the TPEF signal together with an increase of fluorescence lifetime predominantly due to NAD(P)H suggesting metabolic changes in the metastatic foci. FT-IR spectroscopy allowed not only for macrometastasis detection but also their stage definition based mainly on the analysis of proteins, RNA and glycogen fractions. The multimodal approach additionally suggested partial enzymatic degradation of elastin in ECM and collagen remodelling.
Neurodegenerative protein-folding diseases involve the misfolding and aggregation of naturally occurring proteins—a process that plays a role in the progressive deterioration of neurons. Since the disease process involves protein structural changes, and associated compositional changes, vibrational spectroscopy has been used extensively to characterize and understand these processes in diseases including Alzheimer’s disease, Parkinson’s disease, amyotrophic lateral sclerosis, and the prion diseases. A wide range of vibrational spectroscopic and imaging approaches have been used to examine protein misfolding in vitro, in cell culture, and in biological tissue. Here, we describe a range of these techniques and detail how vibrational spectroscopy has led to a more detailed understanding of protein misfolding and subsequent neurodegeneration in the disease process.
The secondary structures (also termed conformations) of silk fibroin (SF) in animal silk fibers and regenerated SF materials are critical in determining mechanical performance and function of the materials. In order to understand the structure-mechanics-function relationships of silk materials, a variety of advanced infrared spectroscopic techniques, such as infrared microspectroscopies (micro-IR spectroscopies for short), synchrotron micro-IR spectroscopies, and infrared nanospectroscopies (nano-IR spectroscopies for short), have been used to disclose the conformations of SF in silk materials. These infrared spectroscopic methods provide a useful toolkit to understand conformations and conformational transitions of SF in various silk materials with spatial resolution from the nano- to micro-scale. In this review, we first summarize progress in understanding the structure and structure-mechanics relationships of silk materials. We then discuss the state-of-the-art micro- and nano-IR spectroscopic techniques used for silk materials characterization. We also provide a systematic discussion of the strategies to collect high-quality spectra and the methods to analyze these spectra. The use and misuse of IR spectroscopies in silk materials characterization are also highlighted. Finally, we demonstrate the challenges and directions for future exploration of silk-based materials with IR spectroscopies.
Results of studies of the use of FTIR spectroscopy in cancer diagnosis are reviewed. Materials studied include biological fluids, tissues, and model systems for studying experimental neoplasia. The methods for working with each system are described. A detailed description is given for the use of IR spectroscopy for cancer diagnosis and in the monitoring of cancer therapy for the evaluation of drug effectiveness and determination of the disease state. Statistical methods for processing the IR spectral data are presented. The main limitations to the use of IR spectroscopy for the diagnosis of oncological diseases and the potential for its introduction into clinical practice are described.
Current methods for cancer detection rely on tissue biopsy, chemical labeling/staining, and examination of the tissue by a pathologist. Though these methods continue to remain the gold standard, they are non-quantitative and susceptible to human error. Fourier transform infrared (FTIR) spectroscopic imaging has shown potential as a quantitative alternative to traditional histology. However, identification of histological components requires reliable classification based on molecular spectra, which are susceptible to artifacts introduced by noise and scattering. Several tissue types, particularly in heterogeneous tissue regions, tend to confound traditional classification methods. Convolutional neural networks (CNNs) are the current state-of-the-art in image classification, providing the ability to learn spatial characteristics of images. In this paper, we demonstrate that CNNs with architectures designed to process both spectral and spatial information can significantly improve classifier performance over per-pixel spectral classification. We report classification results after applying CNNs to data from tissue microarrays (TMAs) to identify six major cellular and acellular constituents of tissue, namely adipocytes, blood, collagen, epithelium, necrosis, and myofibroblasts. Experimental results show that the use of spatial information in addition to the spectral information brings significant improvements in the classifier performance and allows classification of cellular subtypes, such as adipocytes, that exhibit minimal chemical information but have distinct spatial characteristics. This work demonstrates the application and efficiency of deep learning algorithms in improving the diagnostic techniques in clinical and research activities related to cancer.
Using high‐ (HD) and ultra‐high‐definition (UHD) of Fourier Transform Infrared (FTIR) spectroscopic imaging, we characterised spectrally pulmonary metastases in a murine model of breast cancer comparing them with histopathological results (H&E staining). This comparison showed excellent agreement between the methods in case of localisation of metastases with size below 1 mm and revealed that label‐free HD and UHD IR spectral histopathology distinguish the type of neoplastic cells. We primary focused on differentiation between metastatic foci in the pleural cavity from cancer cells present in lung parenchyma and inflamed cells present in extracellular matrix of lungs due to growing of advanced metastases. In addition, a combination of unsupervised clustering and IR imaging indicated the high sensitivity of FTIR spectroscopy to identify chemical features of small macrometastases located under the pleural cavity and during epithelial–mesenchymal transition (EMT). FTIR based spectral histopathology was proved to detect not only phases of breast cancer metastasis to lungs but also to differentiate various origins of metastases seeded from breast cancer.
This article is protected by copyright. All rights reserved.
Oral injuries are currently diagnosed by histopathological analysis of biopsy, which is an invasive procedure and does not give immediate results. On the other hand, Raman spectroscopy technique is a real-time and minimally invasive analytical tool, with notable diagnostic capability. At the current stage, researchers are widely aware of the diagnostic potential of the technique and how it is considered promising for providing biochemical information in real-time and without damaging the tissue. The problem originates from the lack of relevant studies and clinical trials that could show the actual use of Raman spectroscopy to help patients. Our goal here is to narrow the relationship between physicists, chemists, engineers, computer scientists, and the medical community, and in fact discurss the potential of Raman spectroscopy as a novel clinical analysis method. In the present study, we focused in the use of Raman spectroscopy as a daily clinical practice. In this context, additional studies and in vivo tests should be performed with the same approach of a real application. We want to show to the scientific and industrial community what is really necessary for this, starting from a clinical point of view. Using our previous experience publishing in different oral pathologies and types of samples, we also aim to discuss about the current state, potential, and what is required to implement Raman spectroscopy for oral clinical applications.
The brain functions through chemical interactions between many different cell types, including neurons and glia. Acquiring comprehensive information on complex, heterogeneous systems requires multiple analytical tools, each of which have unique chemical specificity and spatial resolution. Multimodal imaging generates complementary chemical information via spatially localized molecular maps, ideally from the same sample, but requires method enhancements that span from data acquisition to interpretation. We devised a protocol for performing matrix-assisted laser desorp-tion/ionization (MALDI)-Fourier transform ion cyclotron resonance-mass spectrometry imaging (MSI), followed by infrared (IR) spectroscopic imaging on the same specimen. Multimodal measurements from the same tissue provide precise spatial alignment between modalities, enabling more advanced image processing such as image fusion and sharpening. Performing MSI first produces higher quality data from each technique compared to performing IR imaging before MSI. The difference is likely due to fixing the tissue section during MALDI matrix removal, thereby preventing analyte degradation occurring during IR imaging from unfixed specimen. Leveraging the unique capabilities of each modality, we utilized pan sharpening of MS (mass spectrometry) ion images with selected bands from IR spectroscopy and midlevel data fusion. In comparison to sharpening with histological images, pan sharpening can employ a plethora of IR bands, producing sharpened MS images while retaining the fidelity of the initial ion images. Using Laplacian pyramid sharpening, we determine the localization of several lipids present within the hippocampus with high mass accuracy at 5 µm pixel widths. Further, through midlevel data fusion of the imaging datasets combined with k-means clustering, the combined dataset discriminates between additional anatomical structures unrecognized by the individual imaging approaches. Significant differences between molecular ion abundances are detected between relevant structures within the hippocampus, such as the CA1 and CA3 regions. Our methodology provides high quality multiplex and multimodal chemical imaging of the same tissue sample, enabling more advanced data processing and analysis routines.
Tissue histology utilizing chemical and immunohistochemical labels plays an extremely important role in biomedicine and disease diagnosis. Recent research suggests that mid-infrared (IR) spectroscopic imaging may augment histology by providing quantitative molecular information. One of the major barriers to this approach is long acquisition time using Fourier-transform infrared (FTIR) spectroscopy. Recent advances in coherent sources, particularly quantum cascade lasers (QCLs), may mitigate this problem by allowing selective sampling of the absorbance spectrum. However, DFIR only provides a significant advantage when the number of spectral samples is minimized, requiring a priori knowledge of important spectral features. In this paper, we demonstrate the use of a GPU-based genetic algorithm (GA) using linear discriminant analysis (LDA) for DFIR feature selection. Our proposed method relies on pre-acquired broadband FTIR images for feature selection. Based on user-selected criteria for classification accuracy, our algorithm provides a minimal set of features that can be used with DFIR in a time-frame more practical for clinical diagnosis.
This paper presents an approach for label-free brain tumor tissue typing. For this application, our dual modality microspectroscopy system combines inelastic Raman scattering spectroscopy and Mie elastic light scattering spectroscopy. The system enables marker-free biomedical diagnostics and records both the chemical and morphologic changes of tissues on a cellular and subcellular level. The system setup is described and the suitability for measuring morphologic features is investigated. Graphical Abstract Bimodal approach for label-free brain tumor typing. Elastic and inelastic light scattering spectra are collected laterally resolved in one measurement setup. The spectra are investigated by multivariate data analysis for assigning the tissues to specific WHO grades according to their malignancy Bimodal approach for label-free brain tumor typing. Elastic and inelastic light scattering spectra are collected laterally resolved in one measurement setup. The spectra are investigated by multivariate data analysis for assigning the tissues to specific WHO grades according to their malignancy
Fourier Transform Infrared (FT-IR) microscopy, coupled with machine learning approaches, has been demonstrated to be a powerful technique for identifying abnormalities in human tissue. The ability to objectively identify the pre-diseased state, and diagnose cancer with high levels of accuracy, has the potential to revolutionise current histopathological practice. Despite recent technological advances in FT-IR microscopy, sample throughput and speed of acquisition are key barriers to clinical translation. Wide-field quantum cascade laser (QCL) infrared imaging systems with large focal plane array detectors and utilising discrete frequency imaging, have demonstrated that large tissue microarrays (TMA) can be imaged in a matter of minutes. However this ground breaking technology is still in its infancy and its applicability for routine disease diagnosis is as yet unproven. In light of this we report on a large study utilising a breast cancer TMA comprised of 207 different patients. We show that by using QCL imaging with continuous spectra acquired between 912-1800 cm-1, we can accurately differentiate between four different histological classes. We demonstrate that we can discriminate between malignant and non-malignant stroma spectra with high sensitivity (93.56%) and specificity (85.64%) for an independent test set. Finally we classify each core in the TMA and achieve high diagnostic accuracy on a patient basis with 100% sensitivity and 86.67% specificity. The absence of false negatives reported here opens up the possibility of utilising high throughput chemical imaging for cancer screening thereby reducing pathologist workload and improving patient care.
Delineating brain tumor margin is critical for maximizing tumor removal while sparing adjacent normal tissue for better clinical outcome. We describe the use of moxifloxacin-based two-photon (TP)/coherent anti-Stokes Raman scattering (CARS) combined microscopy for differentiating normal mouse brain tissue from metastatic brain tumor tissue based on histoarchitectural and biochemical differences. Moxifloxacin, an FDA-approved compound, was used to label cells in the brain, and moxifloxacin-based two-photon microscopy (TPM) revealed tumor lesions with significantly high cellular density and invading edges in a metastatic brain tumor model. Besides, label-free CARS microscopy showed diminishing of lipid signal due to the destruction of myelin at the tumor site compared to a normal brain tissue site resulting in a complementary contrast for tumor detection. This study demonstrates that moxifloxacin-based TP/CARS combined microscopy might be advantageous for tumor margin identification in the brain that has been a long-standing challenge in the operating room.
In biospectroscopy, suitably annotated and statistically independent samples (e.g. patients, batches, etc.) for classifier training and testing are scarce and costly. Learning curves show the model performance as function of the training sample size and can help to determine the sample size needed to train good classifiers. However, building a good model is actually not enough: the performance must also be proven. We discuss learning curves for typical small sample size situations with 5-25 independent samples per class. Although the classification models achieve acceptable performance, the learning curve can be completely masked by the random testing uncertainty due to the equally limited test sample size. In consequence, we determine test sample sizes necessary to achieve reasonable precision in the validation and find that 75-100 samples will usually be needed to test a good but not perfect classifier. Such a data set will then allow refined sample size planning on the basis of the achieved performance. We also demonstrate how to calculate necessary sample sizes in order to show the superiority of one classifier over another: this often requires hundreds of statistically independent test samples or is even theoretically impossible. We demonstrate our findings with a data set of ca. 2550 Raman spectra of single cells (five classes: erythrocytes, leukocytes and three tumour cell lines BT-20, MCF-7 and OCI-AML3) as well as by an extensive simulation that allows precise determination of the actual performance of the models in question.
Infrared spectra of single biological cells often exhibit the 'dispersion artefact' observed as a sharp decrease in intensity on the high wavenumber side of absorption bands, in particular the Amide I band at approximately 1655 cm(-1), causing a downward shift of the true peak position. The presence of this effect makes any biochemical interpretation of the spectra unreliable. Recent theory has shed light on the origins of the 'dispersion artefact' which has been attributed to resonant Mie scattering (RMieS). In this paper a preliminary algorithm for correcting RMieS is presented and evaluated using simulated data. Results show that the 'dispersion artefact' appears to be removed; however, the correction is not perfect. An iterative approach was subsequently implemented whereby the reference spectrum is improved after each iteration, resulting in a more accurate correction. Consequently the corrected spectra become increasingly more representative of the pure absorbance spectra. Using this correction method reliable peak positions can be obtained.
This study applies infrared (IR) spectroscopy to distinguish normal brain tissue from brain metastases and to determine the primary tumor of four frequent brain metastases such as lung cancer, colorectal cancer, breast cancer, and renal cell carcinoma. Standard methods sometimes fail to identify the origin of brain metastases. As metastatic cells contain the molecular information of the primary tissue cells and IR spectroscopy probes the molecular fingerprint of cells, IR spectroscopy based methods constitute a new approach to determine the primary tumor of a brain metastasis. IR spectroscopic images were recorded by a FTIR spectrometer equipped with a macro sample chamber and coupled to a focal plane array detector. Unsupervised cluster analysis of IR images revealed variances within each sample and between samples of the same tissue type. Cluster averaged IR spectra of tissue classes with known diagnoses were selected to develop a metric with eight variables. These data trained a supervised classification model based on linear discriminant analysis that was used to identify the origin of 20 cryosections including one brain metastasis with an unknown primary tumor.
LIBSVM is a library for Support Vector Machines (SVMs). We have been actively developing this package since the year 2000. The goal is to help users to easily apply SVM to their applications. LIBSVM has gained wide popularity in machine learning and many other areas. In this article, we present all implementation details of LIBSVM. Issues such as solving SVM optimization problems theoretical convergence multiclass classification probability estimates and parameter selection are discussed in detail.
Vibrational spectroscopic imaging techniques are new tools for visualizing chemical components in tissue without staining. The spectroscopic signature can be used as a molecular fingerprint of pathological tissues. Fourier transform infrared imaging which is more common than Raman imaging so far has already been applied to identify the primary tumor of brain metastases. The current study introduces a two level discrimination model for Raman microspectroscopic images to distinguish normal brain, necrosis and tumor tissue, and subsequently to determine the primary tumor. 22 Specimens of normal brain tissue and brain metastasis of bladder carcinoma, lung carcinoma, mamma carcinoma, colon carcinoma, prostate carcinoma and renal cell carcinoma were snap frozen, and thin tissue sections were prepared. Raman microscopic images were collected with 785 nm laser excitation at 10 μm step size. Cluster analysis, vertex component analysis and principal component analysis were applied for data preprocessing. Then, data of 17 specimens were used to train the discrimination model based on support vector machines with radial basis functions kernel. The training data were discriminated with accuracy better than 99%. Finally, the discrimination model correctly predicted independent specimens. The results were superior to discrimination by partial least squares discriminant analysis and support vector machines with linear basis function kernel that were applied for comparison.
Raman spectroscopy is a promising tool towards biopsy under vision as it provides label-free image contrast based on intrinsic vibrational spectroscopic fingerprints of the specimen. The current study applied the spectral unmixing algorithm vertex component analysis (VCA) to probe cell density and cell nuclei in Raman images of primary brain tumor tissue sections. Six Raman images were collected at 785 nm excitation that consisted of 61 by 61 spectra at a step size of 2 micrometers. After data acquisition the samples were stained with hematoxylin and eosin for comparison. VCA abundance plots coincided well with histopathological findings. Raman spectra of high grade tumor cells were found to contain more intense spectral contributions of nucleic acids than those of low grade tumor cells. Similarly, VCA endmember signatures of Raman images from high grade gliomas showed increased nucleic acid bands. Further abundance plots and endmember spectra were assigned to tissue containing proteins and lipids, and cholesterol microcrystals. Since no sample preparation is required, an important advantage of the Raman imaging methodology is that all tissue components can be observed - even those that may be lost in sample staining steps. The results demonstrate how morphology and chemical composition obtained by Raman imaging correlate with histopathology and provide complementary, diagnostically relevant information at the cellular level.
In this model study, we developed a method to distinguish between breast cancer cells and normal epithelial cells, which is in principal suitable for online diagnosis by Raman spectroscopy. Two cell lines were chosen as model systems for cancer and normal tissue. Both cell lines consist of epithelial cells, but the cells of the MCF-7 series are carcinogenic, where the MCF-10A cells are normal growing. An algorithm is presented for distinguishing cells of the MCF-7 and MCF-10A cell lines, which has an accuracy rate of above 99%. For this purpose, two classification steps are utilized. The first step, the so-called top-level classifier searches for Raman spectra, which are measured in the nuclei region. In the second step, a wide range of discriminant models are possible and these models are compared. The classification rates are always estimated using a cross-validation and a holdout-validation procedure to ensure the ability of the routine diagnosis to work in clinical environments. Copyright © 2009 John Wiley & Sons, Ltd.
This paper explores different phenomena that cause distortions of infrared absorption spectra by mixing of reflective and absorptive band shape components of infrared spectra, and the resulting distortion of observed band shapes. In the context of this paper, we refer to the line shape of the variations of the refractive index in spectral regions of an absorption maximum (i.e., in regions of "anomalous dispersion") as "dispersive" or "reflective" line shape contributions, in analogy to previous spectroscopic literature. These distortions usually result in asymmetric bands with a negative intensity contribution at the high wavenumber of the band, accompanied by a shift toward lower wavenumber, and confounded band intensities. In extreme cases of band distortions caused by the "resonance Mie" (RMie) mechanism, spectral peaks may be split into doublets of peaks, change from positive to negative peaks, or appear as derivative-shaped features.
A detailed comparison of six multivariate algorithms is presented to analyze and generate Raman microscopic images that consist
of a large number of individual spectra. This includes the segmentation algorithms for hierarchical cluster analysis, fuzzy
C-means cluster analysis, and k-means cluster analysis and the spectral unmixing techniques for principal component analysis and vertex component analysis
(VCA). All algorithms are reviewed and compared. Furthermore, comparisons are made to the new approach N-FINDR. In contrast
to the related VCA approach, the used implementation of N-FINDR searches for the original input spectrum from the non-dimension
reduced input matrix and sets it as the endmember signature. The algorithms were applied to hyperspectral data from a Raman
image of a single cell. This data set was acquired by collecting individual spectra in a raster pattern using a 0.5-μm step
size via a commercial Raman microspectrometer. The results were also compared with a fluorescence staining of the cell including
its mitochondrial distribution. The ability of each algorithm to extract chemical and spatial information of subcellular components
in the cell is discussed together with advantages and disadvantages.
KeywordsChemometrics–Raman spectroscopy–Image processing–Hyperspectral data
We report a computational method to remove or reduce dispersion artifacts from infrared microspectral data collected in transflection (reflection/absorption) mode. This artifact occurs along the edges of tissue samples, in particular if the tissue does not adhere well to the substrate. The method proposed for the removal of the artifact is similar to the phase correction used in standard Fourier transform infrared spectroscopy.
Raman microspectroscopic imaging provides molecular contrast in a label-free manner with subcellular spatial resolution. These properties might complement clinical tools for diagnosis of tissue and cells in the future. Eight Raman spectroscopic images were collected with 785 nm excitation from five non-dried brain specimens immersed in aqueous buffer. The specimens were assigned to molecular and granular layers of cerebellum, cerebrum with and without scattered tumor cells of astrocytoma WHO grade III, ependymoma WHO grade II, astrocytoma WHO grade III, and glioblastoma multiforme WHO grade IV with subnecrotic and necrotic regions. In contrast with dried tissue section, these samples were not affected by drying effects such as crystallization of lipids or denaturation of proteins and nucleic acids. The combined data sets were processed by use of the hyperspectral unmixing algorithms N-FINDR and VCA. Both unsupervised approaches calculated seven endmembers that reveal the abundance plots and spectral signatures of cholesterol, cholesterol ester, nucleic acids, carotene, proteins, lipids, and buffer. The endmembers were correlated with Raman spectra of reference materials. The focus of the single mode laser near 1 μm and the step size of 2 μm were sufficiently small to resolve morphological details, for example cholesterol ester islets and cell nuclei. The results are compared for both unmixing algorithms and with previously reported supervised spectral decomposition techniques.
Figure
Morphological details in tissue sections are resolved by Raman imaging and might contribute together with chemical information to improved diagnosis.
Metastases of various tumors to the brain account for the majority of brain cancers, and are associated with a poor prognosis. The most common primary sites are lung, breast, skin, kidney and colon; 10-40% of cancer patients develop brain metastases during the course of the disease. The incidence of brain metastasis appears to be rising; reasons may include better therapies for the systemic disease with longer survival of cancer patients but lower efficiency against brain metastases. In this article, we will discuss the conventional treatment with surgery, radiosurgery, radiotherapy and chemotherapy, but also new directions in the management of solid brain metastases. While general therapeutic nihilism should be avoided, it is important to recognize that the number of brain metastases, the extent of the systemic disease and also the tumor type have to be taken into account when choosing individual treatment regimens. Finally, special emphasis will be put on established and future approaches to prevent the disease. We thus aim to provide a framework for treating patients with different presentations of brain metastases, and to highlight important avenues for research.
We report for the first time a proof-of-concept experiment employing Raman spectroscopy to detect intracerebral tumors in vivo by brain surface mapping. Raman spectroscopy is a non-destructive biophotonic method which probes molecular vibrations. It provides a specific fingerprint of the biochemical composition and structure of tissue without using any labels. Here, the Raman system was coupled to a fiber-optic probe. Metastatic brain tumors were induced by injection of murine melanoma cells into the carotid artery of mice, which led to subcortical and cortical tumor growth within 14 days. Before data acquisition, the cortex was exposed by creating a bony window covered by a calcium fluoride window. Spectral contributions were assigned to proteins, lipids, blood, water, bone, and melanin. Based on the spectral information, Raman images enabled the localization of cortical and subcortical tumor cell aggregates with accuracy of roughly 250 μm. This study demonstrates the prospects of Raman spectroscopy as an intravital tool to detect cerebral pathologies and opens the field for biophotonic imaging of the living brain. Future investigations aim to reduce the exposure time from minutes to seconds and improve the lateral resolution.
In this manuscript, we report the application of EMSC to correct infrared micro-spectral data recorded from tissue that describe resonant Mie scattering contributions. Small breast micro-metastases previously undetectable using the raw measured spectra were provided clear contrast from the surrounding tissue after signal correction. The technique also proved transferrable, successfully correcting imaging data sets recorded from multiple patients. It is envisaged more robust methods of supervised analysis can now be constructed to automatically classify and diagnose tissue spectra.
Colon tissue constitutes a valid model for the comparative analysis of soft tissue by Raman and Fourier transform infrared (FTIR) imaging because it contains four major tissue types such as muscle tissue, connective tissue, epithelium and nerve cells. Raman microscopic images were recorded in the mapping mode using 785 nm laser excitation and a step size of 10 microm from three regions within a thin section that encompassed mucus, mucosa, submucosa, and longitudinal and circular muscle layers. FTIR microscopic images that were composed of 4, 8 and 9 individual images of 4096 spectra each were recorded from the same regions using a FTIR spectrometer coupled to a microscope with a focal plane array detector. Furthermore, Raman microscopic images were recorded at a step size of 2.5 microm from three ganglia that belong to the myenteric plexus. The results are discussed with respect to lateral resolution, spectral resolution, acquisition time and sensitivity of both modalities.
The purpose of the study was to investigate molecular changes associated with glioma tissues using FT-IR microspectroscopic imaging (FT-IRM). A multivariate statistical analysis allowed one to successfully discriminate between normal, tumoral, peri-tumoral, and necrotic tissue structures. Structural changes were mainly related to qualitative and quantitative changes in lipid content, proteins, and nucleic acids that can be used as spectroscopic markers for this pathology. We have developed a spectroscopic model of glioma to quantify these chemical changes. The model constructed includes individual FT-IR spectra of normal and glioma brain constituents such as lipids, DNA, and proteins (measured on delipidized tissue). Modeling of FT-IR spectra yielded fit coefficients reflecting the chemical changes associated with a tumor. Our results demonstrate the ability of FT-IRM to assess the importance and distribution of each individual constituent and its variation in normal brain structures as well as in the different pathological states of glioma. We demonstrated that (i) cholesterol and phosphatidylethanolamine contributions are highest in corpus callosum and anterior commissure but decrease gradually towards the cortex surface as well as in the tumor, (ii) phosphatidylcholine contribution is highest in the cortex and decreases in the tumor, (iii) galactocerebroside is localized only in white, but not in gray matter, and decreases in the vital tumor region while the necrosis area shows a higher concentration of this cerebroside, (iv) DNA and oleic acid increase in the tumor as compared to gray matter. This approach could, in the future, contribute to enhance diagnostic accuracy, improve the grading, prognosis, and play a vital role in therapeutic strategy and monitoring.
For high-perfomance status patients of unknown primary, non-toxic chemotherapeutic combinations may play an important role in palliation and even survival. A review of 220 patients with poorly differentiated neoplasms by Hainsworth et al. [87] demonstrated a major tumor response in 62% of patients treated with a combination of chemotherapy, including cisplatin, etoposide, and bleomycin. If no response is observed after one or two courses, therapy should be discontinued and alternate approaches considered. If complete response is obtained, radiotherapy of involved nodal basins should be also considered for further local control. The minority of patients who present with clinically favorable disease, such as women with primary peritoneal carcinomatosis or isolated axillary lymphadenopathy, or those whose tissue histology suggests responsiveness to chemotherapeutic or even surgical intervention, should be treated accordingly. Chemotherapy, radiotherapy, and surgery all play integral roles in the treatment of these patients. The promise of long-term survival or even curative resection demonstrates the  importance of a thorough and appropriate clinical evaluation. A few malignancies are as anxiety provoking as the cancer of unknown primary site. Often, the primary physician is bereft of a standard diagnostic and therapeutic protocol, as modern oncologic management is dependent on identification of a primary tumor. In addition, significant confusion and frustration occurs at the patient level over failure to identify the primary. The final goal in the treatment of patients with unknown primary cancers will be the development and utilization of new trials to determine the optimal combination of multimodal therapies in responsive subsets. This includes the use of newer chemotherapeutic agents, surgical techniques, and more advanced molecular characterization of the tumor.
This study uses infrared (IR) spectroscopic, point detection, mapping procedures to examine tissue samples from normal brain specimens and from astrocytic gliomas, the most frequent human brain tumors. Model systems were derived from cultured glioma cell lines. IR spectra of normal tissue sections distinguished white matter from gray matter by increased spectral contributions from lipids and cholesterol. Qualitatively the same differences were found in IR spectra of low and high grade glioma tissue sections pointing to a significant reduction of brain lipids with increasing malignancy. Whereas spectral contributions of proteins and lipids were similar in IR spectra of glioma cells and tissues, nucleic acid bands were more intense for cells suggesting higher proliferative activities. For statistical analyses of IR spectroscopic maps from 71 samples, a parameter for the lipid to protein ratio was introduced involving the CH(2) symmetric stretch band with lipids as main contributors and the amide I band of proteins. As this parameter correlated with the grade of gliomas obtained from standard histopathological examination, it was applied to classify brain tissue sections based on IR spectroscopic mapping.
The diagnosis of a brain metastasis is usually made during the routine follow up examinations of patients with known cancer, who are under the care of oncology departments. The involvement of the neurosurgeon depends on the philosophy and referral patterns of each oncology group. Patients with brain metastases of unknown primary (BMUP) are much more likely to seek the help of a neurosurgeon or a neurologist before contacting an oncologist, because the presenting clinical features originate from the brain. BMUPs are almost equal in numbers to brain primaries and differ from regular cerebral metastases regarding their site of origin, which will remain unknown in about 50% despite vigorous investigation. The clinical picture is similar to that of primary brain tumours but they seem to show different areas of predilection in the brain parenchyma. By reviewing the literature we are presenting the epidemiology, clinical presentation, diagnostic workup and treatment plan for this group of patients.
Brain metastases are secondary intracranial lesions which occur more frequently than primary brain tumors. The four most abundant types of brain metastasis originate from primary tumors of lung cancer, colorectal cancer, breast cancer and renal cell carcinoma. As metastatic cells contain the molecular information of the primary tissue cells and IR spectroscopy probes the molecular fingerprint of cells, IR spectroscopy based methods constitute a new approach to determine the origin of brain metastases. IR spectroscopic images of 4 by 4 mm2 tissue areas were recorded in transmission mode by a FTIR imaging spectrometer coupled to a focal plane array detector. Unsupervised cluster analysis revealed variances within each cryosection. Selected clusters of five IR images with known diagnoses trained a supervised classification model based on the algorithm soft independent modeling of class analogies (SIMCA). This model was applied to distinguish normal brain tissue from brain metastases and to identify the primary tumor of brain metastases in 15 independent IR images. All specimens were assigned to the correct tissue class. This proof-of-concept study demonstrates that IR spectroscopy can complement established methods such as histopathology or immunohistochemistry for diagnosis.
The objectives of this study were to optimize the preparation of pristine brain tissue to obtain reference information, to optimize the conditions for introducing a fiber-optic probe to acquire Raman maps, and to transfer previous results obtained from human brain tumors to an animal model. Brain metastases of malignant melanomas were induced by injecting tumor cells into the carotid artery of mice. The procedure mimicked hematogenous tumor spread in one brain hemisphere while the other hemisphere remained tumor free. Three series of sections were prepared consecutively from whole mouse brains: dried, thin sections for FTIR imaging, hematoxylin and eosin-stained thin sections for histopathological assessment, and pristine, 2-mm thick sections for Raman mapping. FTIR images were recorded using a spectrometer with a multi-channel detector. Raman maps were collected serially using a spectrometer coupled to a fiber-optic probe. The FTIR images and the Raman maps were segmented by cluster analysis. The color-coded cluster memberships coincided well with the morphology of mouse brains in stained tissue sections. More details in less time were resolved in FTIR images with a nominal resolution of 25 μm than in Raman maps collected with a laser focus 60 μm in diameter. The spectral contributions of melanin in tumor cells were resonance enhanced in Raman spectra on excitation at 785 nm which enabled their sensitive detection in Raman maps. Possible reasons why metastatic cells of malignant melanomas were not identified in FTIR images are discussed.
Figure
FTIR image of 310 × 398 pixels obtained from an unstained, dried 10 μm tissue section of normal mouse brain. Colors represent the class assignments by cluster analysis
The purpose of this study was to investigate molecular changes associated with glioma tissues by Raman microspectroscopy in order to develop its use in clinical practice. Spectroscopic markers obtained from C6 glioma tissues were compared to conventional histological and histochemical techniques. Cholesterol and phospholipid contents were highest in corpus callosum and decreased gradually towards the cortex surface as well as in the tumor. Two different necrotic areas have been identified: a fully necrotic zone characterized by the presence of plasma proteins and a peri-necrotic area with a high lipid content. This result was confirmed by Nile Red staining. Additionally, one structure was detected in the periphery of the tumor. Invisible with histopathological hematoxylin and eosin staining, it was revealed by immunohistochemical Ki-67 and MT1-MMP staining used to visualize the proliferative and invasive activities of glioma, respectively. Hierarchical cluster analysis on the only cluster averaged spectra showed a clear distinction between normal, tumoral, necrotic and edematous tissues. Raman microspectroscopy can discriminate between healthy and tumoral brain tissue and yield spectroscopic markers associated with the proliferative and invasive properties of glioblastoma. Development of in vivo Raman spectroscopy could thus accurately define tumor margins, identify tumor remnants, and help in the development of novel therapies for glioblastoma.
- Y Kienast
- F Winkler
Y. Kienast and F. Winkler, Expert Rev. Anticancer Ther., 2010,
10, 1763-1777.
- C C Chang
- C J Lin
C. C. Chang and C. J. Lin, ACM Transactions on Intelligent
Systems and Technology, 2011, 2(3), 27.
- B Bird
- M Miljkovic
- M Diem
B. Bird, M. Miljkovic and M. Diem, J. Biophotonics, 2010, 3,
597-608.
- M Miljkovic
- B Bird
- M Diem
M. Miljkovic, B. Bird and M. Diem, Analyst, 2012, 137, 3954-3964.
- T Bocklitz
- M Putsche
- C Stueber
- J Kaes
- A Niendorf
- P Roesch
- J Popp
T. Bocklitz, M. Putsche, C. Stueber, J. Kaes, A. Niendorf,
P. Roesch and J. Popp, J. Raman Spectrosc., 2009, 40, 1759-1765.
- A Beljebbar
- N Amharref
- A Leveques
- S Dukic
- L Venteo
- L Schneider
- M Pluot
- M Manfait
A. Beljebbar, N. Amharref, A. Leveques, S. Dukic, L. Venteo,
L. Schneider, M. Pluot and M. Manfait, Anal. Chem., 2008,
80, 8406-8415.
- P Bassan
- A Kohler
- H Martens
- J Lee
- H J Byrne
- P Dumas
- E Gazi
- M D Brown
- N W Clarke
- P Gardner
P. Bassan, A. Kohler, H. Martens, J. Lee, H. J. Byrne,
P. Dumas, E. Gazi, M. D. Brown, N. W. Clarke and
P. Gardner, Analyst, 2010, 135, 268-277.