19 February 2018 Automated pre-processing and multivariate vibrational spectra analysis software for rapid results in clinical settings
Author Affiliations +
Abstract
Vibrational spectroscopy, especially FTIR and Raman, has shown enormous potential in disease diagnosis, especially in cancers. Their potential for detecting varied pathological conditions are regularly reported. However, to prove their applicability in clinics, large multi-center multi-national studies need to be undertaken; and these will result in enormous amount of data. A parallel effort to develop analytical methods, including user-friendly software that can quickly pre-process data and subject them to required multivariate analysis is warranted in order to obtain results in real time. This study reports a MATLAB based script that can automatically import data, preprocess spectra— interpolation, derivatives, normalization, and then carry out Principal Component Analysis (PCA) followed by Linear Discriminant Analysis (LDA) of the first 10 PCs; all with a single click. The software has been verified on data obtained from cell lines, animal models, and in vivo patient datasets, and gives results comparable to Minitab 16 software. The software can be used to import variety of file extensions, asc, .txt., .xls, and many others. Options to ignore noisy data, plot all possible graphs with PCA factors 1 to 5, and save loading factors, confusion matrices and other parameters are also present. The software can provide results for a dataset of 300 spectra within 0.01 s. We believe that the software will be vital not only in clinical trials using vibrational spectroscopic data, but also to obtain rapid results when these tools get translated into clinics.
© (2018) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
T. Bhattacharjee, P. Kumar, L. Fillipe, "Automated pre-processing and multivariate vibrational spectra analysis software for rapid results in clinical settings", Proc. SPIE 10489, Optical Biopsy XVI: Toward Real-Time Spectroscopic Imaging and Diagnosis, 104890A (19 February 2018); doi: 10.1117/12.2290159; https://doi.org/10.1117/12.2290159
PROCEEDINGS
7 PAGES


SHARE
Back to Top