Entity

Time filter

Source Type

Gloucester, United Kingdom

Isabelle M.,Biophotonics Research Group | Rogers K.,Cranfield University | Stone N.,Biophotonics Research Group
Journal of Biomedical Optics | Year: 2010

In this work, a novel technique for rapid image analysis of Fourier transform infrared (FTIR) data obtained from human lymph nodes is explored. It uses the mathematical principle of orthogonality as a method to quickly and efficiently obtain tissue and pathology information from a spectral image cube. It requires less computational power and time compared to most forms of cluster analysis. The values obtained from different tissue and pathology types allows for discrimination of noncancerous from cancerous lymph nodes. It involves the calculation of the dot product between reference spectra and individual spectra from across the tissue image. These provide a measure of the correlation between individual spectra and the reference spectra, and each spectrum or pixel in the image is given a color representing the reference most closely correlating with it. The correlation maps are validated with the tissue and pathology features identified by an expert pathologist from corresponding hematoxylin and eosin stained tissue sections. Although this novel technique requires further study to properly test and validate this tool, with inclusion of more lymph node hyperspectral datasets (containing a greater variety of tissue states), it demonstrates significant clinical potential for pathology diagnosis. © 2010 Society of Photo-Optical Instrumentation Engineers. Source


Lloyd G.R.,Biophotonics Research Group | Orr L.E.,Biophotonics Research Group | Orr L.E.,Cranfield University | Christie-Brown J.,Gloucestershire Royal Hospital | And 5 more authors.
Analyst | Year: 2013

Background: The potential use of Raman spectroscopy (RS) for the detection of malignancy within lymph nodes of the head and neck was evaluated. RS measures the presence of biomolecules by the inelastic scattering of light within cells and tissues. This can be performed in vivo in real-time. Methods: 103 lymph nodes were collected from 23 patients undergoing surgery for suspicious lymph nodes. Five pathologies, defined by consensus histopathology, were collected including reactive nodes (benign), Hodgkin's and non-Hodgkin's lymphomas, metastases from both squamous cell carcinomas and adenocarcinomas. Raman spectra were measured with 830 nm excitation from numerous positions on each biopsy. Spectral diagnostic models were constructed using principal component analysis followed by linear discriminant analysis (PCA-LDA), and by partial least squares discriminant analysis (PLS-DA) for comparison. Two-group models were constructed to distinguish between reactive and malignant nodes, and three-group models to distinguish between the benign, primary and secondary conditions. Results: Results were validated using a repeated subsampling procedure. Sensitivities and specificities of 90% and 86% were obtained using PCA-LDA, and 89% and 88% using PLS-DA, for the two-group models. Both PCA-LDA and PLS-DA models were also found to be very successful at discriminating between pathologies in the three-group models achieving sensitivities and specificities of over 78% and 89% for PCA-LDA, and over 81% and 89% for PLS-DA for all three pathology groups. Conclusion: Raman spectroscopy and chemometric techniques can be successfully utilised in combination for discriminating between different cancerous conditions of lymph nodes from the head and neck. © The Royal Society of Chemistry 2013. Source


Aneesh A.,University of Cardiff | Povazay B.,Medical University of Vienna | Hofer B.,Medical University of Vienna | Zhang E.Z.,University College London | And 9 more authors.
Progress in Biomedical Optics and Imaging - Proceedings of SPIE | Year: 2010

High speed, three-dimensional optical coherence tomography (3D OCT) at 800nm, 1060nm and 1300nm with approximately 4μm, 7μm and 6μm axial and less than 15μm transverse resolution is demonstrated to investigate the optimum wavelength region for in vivo human skin imaging in terms of contrast, dynamic range and penetration depth. 3D OCT at 1300nm provides deeper penetration, while images obtained at 800nm were better in terms of contrast and speckle noise. 1060nm region was a compromise between 800nm and 1300nm in terms of penetration depth and image contrast. Optimizing sensitivity, penetration and contrast enabled unprecedented visualization of micro-structural morphology underneath the glabrous skin, hairy skin and in scar tissue. Higher contrast obtained at 800 nm appears to be critical in the in vitro tumor study. A multimodal approach combining OCT and PA helped to obtain morphological as well as vascular information from deeper regions of skin. © 2010 Copyright SPIE - The International Society for Optical Engineering. Source


Sattlecker M.,Cranfield University | Stone N.,Biophotonics Research Group | Smith J.,Biophotonics Research Group | Bessant C.,Cranfield University
Journal of Raman Spectroscopy | Year: 2011

Over recent years, Raman spectroscopy has been demonstrated as a prospective tool for application in cancer diagnostics. The use of Raman spectroscopy for this purpose relies on pattern recognition methods that have been developed to perform well on data achieved under laboratory conditions. However, the application of Raman spectroscopy as a routine clinical tool is likely to result in imperfect data due to instrument-to-instrument variation. Such corruption to the pure tissue spectral data is expected to negatively impact the classification performance of the diagnostic model. In this paper, we present a thorough assessment of the robustness of the Raman approach. This was achieved by perturbing a set of spectra in different ways, including various linear shifts, nonlinear shifts and random noise and using previously optimised classification models to predict the class membership of each spectrum in a testing set. The loss of predictive power with increased corruption was used to calculate a score, which allows an easy comparison of the model robustness. For this approach, three different types of classification models, including linear discriminant analysis (LDA), partial least square discriminant analysis (PLS-DA) and support vector machine (SVM), built for lymph node diagnostics were the subject of the robustness testing. The results showed that a linear perturbation had the highest impact on the performance of all classification models. Among all linear corruption methods, a gradient y-shift resulted in the highest performance loss. Thus, the factor most likely to affect the predictive outcome of models when using different systems is a gradient y-shift. Copyright © 2011 John Wiley & Sons, Ltd. Source


Sattlecker M.,Cranfield University | Baker R.,Biophotonics Research Group | Stone N.,Biophotonics Research Group | Bessant C.,Cranfield University
Chemometrics and Intelligent Laboratory Systems | Year: 2011

Over the past years Fourier transform infrared (FTIR) spectroscopy has been demonstrated as a prospective tool for cancer diagnostics. In order to apply FTIR spectroscopy as a routine tool for biomedical diagnostics of tissue samples, strong and reliable classifiers are needed. Frequently, the number of available tissue samples is restricted and due to that data sets consist of a small number of samples, often less than 100. This can result in unstable classifiers, which perform poorly on unseen data. In this work we present a way to overcome this limitation by aggregating several support vector machines in to an ensemble. Different ensemble systems, including bagging, boosting and tree-based models, were investigated for a FTIR data set acquired from different types and stages of breast cancer. It was found that an ensemble system predicts 88.9% of the unseen multi-class test set correctly. In comparison a single classifier only achieved a predictive performance of 66.7%. As these results show, the application of SVM ensembles in biomedical diagnostics using FTIR spectroscopy can be highly beneficial. © 2011 Elsevier B.V. Source

Discover hidden collaborations