Automated method for image analysis of residual protein
AssigneeChromaVision Medical Systems, Inc.
InventorsPresley Hays, Michal L. Peri, Douglas Harrington
Life of this patent
Estimating the term…
Abstract
An automated method for evaluating the amount of residual protein levels following treatment of cellular specimens. Particular methods include the measurement of maternal neutrophil alkaline phosphatase after treatment with or without urea or heat, measurement of leukocyte acid phosphatase after treatment with or without tartrate, and measurement of leukocyte esterase after treatment with α-naphthol butyrate with or without fluoride.
Claims16
An automated method for the measurement of residual protein in a treated cellular specimen, comprising:
- (a) providing a plurality of stained subsamples from a cellular specimen, wherein
- (1) at least one of the subsamples has been treated before being stained;
- (2) at least one of the samples has not been treated;
- and
- (3) the untreated subsample and the treated subsample had been stained together;
- (b) automatically selecting a Z position in each subsample for a candidate object of interest;
- (c) automatically obtaining a low magnification image of the candidate objects of interest in each subsample;
- (d) automatically filtering the candidate object of interest pixels in each subsample with a low pass filter;
- (e) automatically morphologically processing the candidate object of interest pixels in each subsample to identify artifact pixels;
- (f) automatically identifying the candidate objects of interest in each subsample from the remaining candidate object of interest pixels in the subsample not identified as artifact pixels;
- (g) adjusting the apparatus to a higher magnification;
- (h) automatically acquiring a higher magnification image of the subsample, at the location coordinates corresponding to the low magnification image, for each candidate object of interest;
- (i) automatically transforming pixels of the higher magnification image in the first color space to a second color space to differentiate higher magnification candidate objects of interest pixels from background pixels;
- (j) automatically identifying, at high magnification, objects of interest from the candidate object of interest pixels in the second color space;
- (k) scoring the protein level in the untreated subsamples;
- (l) scoring the protein level in the treated subsamples;
- and
- (m) determining
- Δ=[protein level in the treated subsamples]/[protein level in the untreated subsamples], wherein Δ is a measurement of the residual component of a cellular protein.
A computer program, residing on a computer-readable medium, for obtaining images of subsamples of a cellular specimen, the computer program comprising instructions for causing a computer to:
- (a) select a Z position in each treated and untreated subsample for candidate object of interest;
- (b) obtain a low magnification image of the candidate object of interest in each subsample;
- (c) filter the candidate object of interest pixels in each subsample with a low pass filter;
- (d) morphologically process the candidate object of interest pixels in each subsample to identify artifact pixels;
- (e) identify the candidate object of interest in each subsample from the remaining candidate object of interest pixels in the subsample not identified as artifact pixels;
- (f) adjust the apparatus to a higher magnification;
- (g) acquire a higher magnification image of the subsample, at the location coordinates corresponding to the low magnification image, for each candidate object of interest;
- (h) transform pixels of the higher magnification image in the first color space to a second color space to differentiate higher magnification candidate objects of interest pixels from background pixels;
- (i) identify, at higher magnification, object of interest from the candidate object of interest pixels in the second color space;
- (j) score a protein level in the untreated subsamples;
- (k) score a protein level in the treated subsamples;
- and
- (l) calculate a Δ=[protein level in the treated subsamples]/[protein level in the untreated subsamples], wherein Δ is a measurement of the residual component of a cellular protein.
Dependencies are inferred from the claim text.