White blood cells (WBC) have crucial roles in immune systems which defend the host against from disease conditions and harmful invaders. Various WBC subsets have been characterized and reported to be involved in many pathophysiologic conditions. It is crucial to isolate a specific WBC subset to study its pathophysiological roles in diseases. Identification methods for a specific WBC population are rely on invasive approaches, including Wright-Gimesa staining for observing cellular morphologies and fluorescence staining for specific protein markers. While these methods enable precise classification of WBC populations, they could disturb cellular viability or functions.
In order to classify WBC populations in a non-invasive manner, we exploited optical diffraction tomography (ODT). ODT is a three-dimensional (3-D) quantitative phase imaging technique that measures 3-D refractive index (RI) distributions of individual WBCs. To test feasibility of label-free classification of WBC populations using ODT, we measured four subtypes of WBCs, including B cell, CD4 T cell, CD8 T cell, and natural killer (NK) cell. From measured 3-D RI tomograms of WBCs, we obtain quantitative structural and biochemical information and classify each WBC population using a machine learning algorithm.
|