Lung cancer pathology discrimination techniques using time series analysis

Abstract

The goal of this study is to discover, analyze, compare, and interpret diffused reflectance polarimetric signatures from lung cancer cells through time series analysis techniques, by using recently invented efficient polarimetric backscattering detection techniques. Specifically, different time series analyses, relying on linear and generalized linear modeling, have been investigated, with special emphasis on the Granger test for the time series. The experimental results indicate that statistically enhanced discrimination between normal and different types of lung cancer cells and stages can be achieved based on the pairwise comparisons of the time series diffused reflectance signal intensities and depolarization properties of the cells.