| Literature DB >> 26021717 |
Iftikhar Ahmad1, Manzoor Ahmad2, Karim Khan3, Sumara Ashraf3, Shakil Ahmad4, Masroor Ikram3.
Abstract
Mueller matrix polarimetry along with polar decomposition algorithm was employed for the characterization of ex vivo normal and adenocarcinoma human colon tissues by polarized light in the visible spectral range (425-725 nm). Six derived polarization metrics [total diattenuation (DT ), retardance (RT ), depolarization(ΔT ), linear diattenuation (DL), retardance (δ), and depolarization (ΔL)] were compared for normal and adenocarcinoma colon tissue samples. The results show that all six polarimetric properties for adenocarcinoma samples were significantly higher as compared to the normal samples for all wavelengths. The Wilcoxon rank sum test illustrated that total retardance is a good candidate for the discrimination of normal and adenocarcinoma colon samples. Support vector machine classification for normal and adenocarcinoma based on the four polarization properties spectra (ΔT , ΔL, RT ,and δ) yielded 100% accuracy, sensitivity, and specificity, while both DTa nd DL showed 66.6%, 33.3%, and 83.3% accuracy, sensitivity, and specificity, respectively. The combination of polarization analysis and given classification methods provides a framework to distinguish the normal and cancerous tissues.Entities:
Mesh:
Year: 2015 PMID: 26021717 DOI: 10.1117/1.JBO.20.5.056012
Source DB: PubMed Journal: J Biomed Opt ISSN: 1083-3668 Impact factor: 3.170