Search

Raman spectroscopy is a candidate technique for diagnosis applications in medicine due to its high molecular specificity. Optimizing the pre-treatment applied for Raman data is important for exploiting Raman signals and ensuring their relevance in medical diagnosis. One of the crucial steps in data pre-processing, normalization, can affect significantly the result interpretation. To select the appropriate normalization method, a strategy based on validity indices (VI) is proposed in this study. VI are based on measuring the quality of data partitioning without involving a full sequence of supervised classification. The approach was tested on Raman data acquired from control and in vitro glycated proteins (albumin and collagen). Protein glycation is a process involved in the molecular ageing of tissues that leads to the formation of products altering the functional and structural properties of proteins. Different methods of normalization were applied on the data sets: integrated intensity of the phenylalanine band, integrated intensity of the amide I band, standard normal variate (SNV), multiplicative signal correction (MSC), and extended multiplicative signal correction (EMSC) that performs simultaneously baseline correction and normalization. Following normalization, principal component analysis (PCA) was applied and VI were calculated from PCA scores resulting from each of the normalization methods mentioned. Based on VI quantitative values, our experiments permit to illustrate the effect of normalization on the data separability of control and glycated samples, and to determine the most appropriate normalization and simultaneously the most discriminant principal components to exploit vibrational information associated with glycation-induced modifications. In parallel, principal component analysis – linear discriminant analysis (PCA-LDA) was carried out for positioning the interest of VI in regard to a common chain of data processing. [ABSTRACT FROM AUTHOR]

Non-melanoma skin cancer includes basal cell carcinoma (BCC), squamous cell carcinoma (SCC) and Bowen's disease. The differential diagnosis of these lesions is sometimes difficult and relies on the histopathological examination of surgical specimens. However, a precise differential diagnosis is crucial for an accurate therapy and thus better patient care. FTIR spectral micro-imaging was applied directly on formalin-fixed paraffin-embedded samples of non-melanoma skin cancers. Chemometric and multivariate statistical analyses were developed to generate an automated IR-based histology without any chemical dewaxing. Different prediction models were developed using linear discriminant analysis combined with data reduction by Principal Component Analysis (PCA) or by wavenumber selection using statistical tests or genetic algorithms. Pseudo-colour maps were reconstructed and compared to conventional histology procedures. High correlation was obtained between the prediction maps and the histology which proves the great potential of FTIR spectroscopy for the differential diagnosis of skin carcinomas. [ABSTRACT FROM AUTHOR]

FTIR spectral imaging was applied on formalin-fixed paraffin-embedded biopsies from colon and skin cancerous lesions. These samples were deposited onto different substrates (zinc selenide and calcium fluoride respectively) and embedded using two types of paraffin. Formalin fixation followed by paraffin embedding is the gold standard in tissue storage. It can preserve molecular structures and it is compatible with immunohistochemistry. However, paraffin absorption bands are significant in the mid-infrared region and can mask some molecular vibrations of the tissue. Direct data processing was applied on spectral images without any chemical dewaxing of the tissues. Extended Multiplicative Signal Correction was used to correct the spectral contribution from paraffin. For this purpose, the signal of paraffin was modelled using Principal Component Analysis and paraffin spectra were removed from the raw images based on an outlier detection. Then, pseudo-colour images were computed by K-means clustering in order to highlight histological structures of interest. This robust chemometrics methodology was applied on the two samples. Tumour areas were successfully demarcated from the rest of the tissue in both colon and skin independently of the embedding material and of the substrate. [ABSTRACT FROM AUTHOR]