Search

Shape variations of red blood cells (RBCs) are known to occur upon exposure to various drugs or under diseased conditions. The commonly observed discocytic RBCs can be transformed to echinocytic or stomatocytic shape under such conditions. Raman spectra of the three major shape variations, namely discocyte, echinocyte, and stomatocyte, of RBCs were studied while subjecting the cells to oxygenated and deoxygenated conditions. Analysis of the recorded spectra suggests an increased level of hemoglobin (Hb)-oxygen affinity for the echinocytes. Also, some level of Hb degradation could be noticed for the deoxygenated echinocytes. The effects may arise from a reduced level of intracellular adenosine triphosphate in echinocytic cells and an increased fraction of submembrane Hb., ((2017) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE).)

For a given diagnostic problem, important considerations are the relative performances of the various optical biopsy techniques. A comparative evaluation of fluorescence, diffuse reflectance, combined fluorescence and diffuse reflectance, and Raman spectroscopy in discriminating different histopathologic categories of human breast tissues is reported. Optical spectra were acquired ex vivo from a total of 74 breast tissue samples belonging to 4 distinct histopathologic categories: invasive ductal carcinoma (IDC), ductal carcinoma in situ (DCIS), fibroadenoma (FA), and normal breast tissue. A probability-based multivariate statistical algorithm capable of direct multiclass classification was developed to analyze the diagnostic content of the spectra measured from the same set of breast tissue sites with these different techniques. The algorithm uses the theory of nonlinear maximum representation and discrimination feature for feature extraction, and the theory of sparse multinomial logistic regression for classification. The results reveal that the performance of Raman spectroscopy is superior to that of all others in classifying the breast tissues into respective histopathologic categories. The best classification accuracy was observed to be approximately 99%, 94%, 98%, and 100% for IDC, DCIS, FA, and normal breast tissues, respectively, on the basis of leave-one-sample-out cross-validation, with an overall accuracy of approximately 99%.

We investigate the potential of near-infrared Raman microspectroscopy to differentiate between normal and malignant skin lesions. Thirty-nine skin tissue samples consisting of normal, basal cell carcinoma (BCC), squamous cell carcinoma (SCC), and melanoma from 39 patients were investigated. Raman spectra were recorded at the surface and at 20-microm intervals below the surface for each sample, down to a depth of at least 100 microm. Data reduction algorithms based on the nonlinear maximum representation and discrimination feature (MRDF) and discriminant algorithms using sparse multinomial logistic regression (SMLR) were developed for classification of the Raman spectra relative to histopathology. The tissue Raman spectra were classified into pathological states with a maximal overall sensitivity and specificity for disease of 100%. These results indicate the potential of using Raman microspectroscopy for skin cancer detection and provide a clear rationale for future clinical studies.

We report the application of a support vector machine (SVM) for the development of diagnostic algorithms for optical diagnosis of cancer. Both linear and nonlinear SVMs have been investigated for this purpose. We develop a methodology that makes use of SVM for both feature extraction and classification jointly by integrating the newly developed recursive feature elimination (RFE) in the framework of SVM. This leads to significantly improved classification results compared to those obtained when an independent feature extractor such as principal component analysis (PCA) is used. The integrated SVM-RFE approach is also found to outperform the classification results yielded by traditional Fisher's linear discriminant (FLD)-based algorithms. All the algorithms are developed using spectral data acquired in a clinical in vivo laser-induced fluorescence (LIF) spectroscopic study conducted on patients being screened for cancer of the oral cavity and normal volunteers. The best sensitivity and specificity values provided by the nonlinear SVM-RFE algorithm over the data sets investigated are 95 and 96% toward cancer for the training set data based on leave-one-out cross validation and 93 and 97% toward cancer for the independent validation set data. When tested on the spectral data of the uninvolved oral cavity sites from the patients it yielded a specificity of 85%., (Copyright 2005 Society of Photo-Optical Instrumentation Engineers.)