Three-dimensional nucleus-to-cytoplasm ratios provide better discrimination of normal and lung adenocarcinoma cells than in two dimensions

Abstract

We acquired multiphoton images of normal and lung adenocarcinoma cell lines in three dimensions. Image stacks of the cells were then processed to obtain nucleus-to-cytoplasm (N/C) ratios in two and three dimensions. While N/C ratios in three dimensions can be unambiguously determined from the volumetric ratios of the nucleus and cytoplasm, two-dimensional (2-D) N/C can vary depending on the axial plane selected for N/C ratio determination. We determined 2-D N/C ratios from three criteria: (1) axial position at which the nuclear area is the largest; (2) the largest 2-D N/C ratio value; and (3) axial position at the midpoint of nuclear axial position. We found that different definitions of 2-D N/C ratio will significantly affect its value. Furthermore, in general, larger variance was found in 2-D rather than three-dimensional (3-D) N/C ratios. Lack of ambiguity in definition and reduced variance suggest that 3-D N/C ratio is a better parameter for characterizing tumor cells in the clinical setting.