Optical projection imaging system and method for automatically detecting cells having nuclear and cytoplasmic densitometric features associated with disease

A method for identifying the at least one object of interest from a cell sample, comprising the steps of:

• (a) obtaining a cell sample and suspending the cells in solution;
• (b) if required, fixing the cells of the cell sample in solution;
• (c) marking the cells to generate optical densities within each cell of the sample;
• (d) illuminating the sample with at least one point source of light;
• (e) obtaining at least one projection image through the sample with a digital array detector;
• (f) compensating the at least one projection image for variations in background illumination;
• (g) analyzing the at least one projection image to detect at least one object of interest;
• (h) calculating a set of feature values having one-dimensional (1D) feature values and two-dimensional (2D) feature values for the at least one object of interest, wherein the set of feature values used to classify the at least one object of interest in the sample are selected from the group consisting of 1D nuclear densitometric features (NDFs), 2D NDFs, 1D cytoplasmic densitometric features (CDFs), 2D CDFs, area, mean radius, optical density (OD) variance, OD skewness, OD range, OD average, OD maximum, density of light spots, low DNA area, high DNA area, low DNA amount, high DNA amount, high average distance, mid/high average distance, correlation, homogeneity, entropy, fractal dimension, DNA index, texture, punctateness, connected components and harmonics in spatial density frequency space;
• (i) providing the set of feature values to at least one classifier;
• and
• (j) using the set of feature values and the at least one classifier for identifying the at least one object of interest from the cell sample.