Detection of circulating tumor cells using imaging flow cytometry

• US 8,885,913 B2
• Filed: 09/27/2013
• Issued: 11/11/2014
• Est. Priority Date: 01/25/1999
• Status: Expired due to Fees

• Current Assignee: Cytek Biosciences Inc
• Original Assignee: Amnis Corporation (DiaSorin S.p.A.)
• Inventors: Basiji, David A., Ortyn, William E., Morrissey, Philip J., Venkatachalam, Vidya
• Primary Examiner(s): BLACKMAN, ROCHELLE ANN J
• Application Number:
  US14/040,398
• Publication Number:
  US 20140030729A1
• Time in Patent Office: 410 Days
• Field of Search: 382/128, 382/133, 382/134, 600/410, 600/562, 356/451, 356/456, 356/326
• US Class Current: 382/133
• CPC Class Codes: G01N 15/147 the analysis being performed on a sample stream
• CPC Class Codes: G01N 2021/6441 with two or more labels
• CPC Class Codes: G01N 21/6428 Measuring fluorescence of fluorescent products of reactions or of fluorochrome labelled reactive substances, e.g. measuring quenching effects, using measuring "optrodes" in vivo A61B5/00; immunoassay G01N33/53
• CPC Class Codes: G01N 21/6458 Fluorescence microscopy fluorescence microscopes per se G02B21/0076 and G02B21/16
• CPC Class Codes: G01N 33/5091 for testing the pathological state of an organism
• CPC Class Codes: G01N 33/5094 for blood cell populations red blood cells G01N33/80
• CPC Class Codes: G01N 33/574 for cancer
• CPC Class Codes: G01N 33/57492 involving compounds localized on the membrane of tumor or cancer cells
• CPC Class Codes: G06T 2207/30024 Cell structures in vitro; Tissue sections in vitro
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06V 20/695 Preprocessing, e.g. image segmentation

Abstract

An automated identification of the types of white blood cells in a blood sample facilitates the manual identification of cancerous or other abnormal blood cells in the sample. Classifiers are predetermined for each type of white blood cell and subsequently used to automatically process images of cells in a sample stained with a nuclear dye or stain. The classifiers each comprise a linear weighted combination of morphometric and photometric features previously selected for white blood cells that were identified using monoclonal antibody stains. Red blood cells and excess fluid are removed from a sample being processed upstream of an imaging region of the imaging system. A plurality of different types of images are produced for each cell by the imaging system enabling automated identification of the white blood cells. Images of any cells not thus identified are manually reviewed to detect cancerous or abnormal cells.

Claims

1. A method for detecting cancerous or other types of abnormal cells in a blood sample, comprising:

(a) staining the nuclei of cells in the blood sample using a nuclear dye or stain, producing stained cells;

(b) removing red blood cells from the blood sample, leaving a residual sample comprising primarily white blood cells and a fluid;

(c) imaging the stained cells to simultaneously produce a plurality of different types of images of each stained cell;

(d) automatically evaluating the plurality of different types of images to detect features of the images, and based on classifiers previously defined as a function of features selected for distinguishing each of a plurality of different types of white blood cells, automatically identifying a first subset of the stained cells as specific types of white blood cells in the blood sample, where a second subset of cells includes cells that were not identified as being any type of white blood cell; and

(e) manually reviewing images of the second subset of cells to determine whether any of the cells in the second subset are cancerous or other type of abnormal cell.

2. The method of claim 1, further comprising using one or more other selected cell-specific stains to facilitate identifying or excluding images if the second subset as cancerous or other type of abnormal cell.

3. The method of claim 1, wherein staining the nuclei of cells in the residual sample comprises applying fixation and permeabilization of the cells before staining the nuclei with the nuclear dye or stain.

4. The method of claim 1, wherein removing the red blood cells comprises applying either a filtering process or using an acoustic technique to separate the red blood cells and excess fluid from the blood sample to produce the residual sample.

5. The method of claim 1, wherein the nuclear dye or stain comprises one of:

(a) 4′

,6-diamidino-2-phenylindole (DAPI);

(b) a cell-permeant cyanine nucleic acid stain;

(c) an A-T intercalating anthraquinone stain;

(d) 7-Aminoactinomycin D (7-AAD);

or

(e) propidium iodide.

6. The method of claim 1, wherein imaging the stained cells to simultaneously produce a plurality of different types of images comprises processing the stained cells with a flow cytometer that forms images of cells passing through an imaging region of the flow cytometer, using light in a plurality of different channels, light in each channel being used to produce a different one of the plurality of different types of images on separate portions of a light detector.

7. The method of claim 1, wherein the plurality of different types of images that are produced include a bright field image, a side scatter image, and a nuclear fluorescence image.

8. The method of claim 1, wherein automatically evaluating the plurality of different types of images to detect features of the images comprises automatically detecting morphometric parameters and photometric parameters evident in the plurality of different types of images, and wherein identifying the first subset of the stained cells as specific types of white blood cells in the blood sample comprises applying the classifiers previously determined to the morphometric and photometric parameters detected in the images of the cells, where each classifier is applied to detect a different type of white blood cell.

9. The method of claim 8, further comprising using linear discriminant analysis to determine the classifiers for each different type of white blood cell, to form a weighted linear combination of selected features for each classifier.

10. The method of claim 9, wherein the features were previously selected for determining each classifier for each different type of white blood cell by evaluating populations of donor blood cells that were stained with both the nuclear dye or stain, and with monoclonal antibodies appropriate for identifying each type of white blood cell, using features related to the monoclonal antibodies to identify each different type of white blood cells in the populations of donor cells, and then selecting specific features related to the nuclear dye or stain for each type of white blood cell thus identified that provided a greatest discrimination relative to other types of white blood cells, for use in the linear weighted combination of features used for a classifier for that type of white blood cell.

11. Apparatus for use in facilitating detection of cancerous or other types of abnormal cells in a blood sample, comprising:

(a) an image acquisition subsystem configured to simultaneously acquire a plurality of different types of images of individual cells in the blood sample, where the individual cells have been stained with a nuclear dye or stain, the plurality of different types of images exhibiting morphometric and photometric parameters characteristic of the type of cell being imaged; and

(b) a programmed image processing system for automatically identifying white blood cells in the blood sample, based on selected features derived from the morphometric and photometric parameters detected in the different types of images, using predefined classifiers that employ the selected features for each different type of white blood cell, the programmed image processing system also designating images for any remaining cells in the blood sample that were not automatically identified as any type of white cell, for subsequent manual review that can determine if any of the remaining cells are cancerous or abnormal.

12. The apparatus of claim 11, further comprising a red blood cell separator disposed upstream of an imaging region in the image acquisition subsystem, the red blood cell separator removing red blood cells from the blood sample that passes into the imaging region.

13. The apparatus of claim 12, wherein the red blood cell separator comprises either a filter with a filter screen sized to pass red blood cells but not white blood cells, or an acoustic separator that applies an acoustic force to shift the red blood cells and the white blood cells in a flow of fluid entering the imaging region apart, so that mostly white blood cells enter the imaging region and the red blood cells do not.

14. The apparatus of claim 12, wherein the red blood cell separator also reduces an amount of fluid included in the blood sample entering the imaging region, to increase a concentration of the white blood cells being imaged by the image acquisition subsystem.

15. The apparatus of claim 11, wherein the plurality of different types of images include:

(a) bright field images;

(b) side scatter images; and

(c) nuclear fluorescence images.

16. The apparatus of claim 11, wherein the image acquisition subsystem includes a beam splitter and a plurality of optical filters for creating different light paths to create the plurality of different types of images.

17. The apparatus of claim 11, wherein the predefined classifiers employed by the programmed image processing system for identifying each type of white blood cell are based on a linear discriminant analysis and are a linear weighted combination of the selected features for each type of white blood cell, the selected features having been previously identified as providing greater discrimination between one type of white blood cells and other types of white blood cells.

18. A method for determining classifiers for identifying each type of white blood cell in a sample of blood, comprising:

(a) providing a donor blood sample;

(b) labeling white blood cells in the donor sample with a nuclear dye or stain, and with monoclonal antibodies selected for identification of each type of white blood cell, producing a stained sample;

(c) substantially reducing a number of red blood cells in the stained sample, producing a residual sample;

(d) processing cells in the residual sample using an imaging system that simultaneously produces a set of different types of images for each white blood cell that is imaged by the imaging system, the set including a bright field image, a side scatter image, an immunofluorescence image, and a nuclear fluorescence image;

(e) using the immunofluorescence images to determine truth in regard to identification of the white blood cells included in the residual sample;

(f) analyzing the bright field, side scatter, and nuclear fluorescence images to detect photometric and morphometric parameters comprising features for each type of white blood cells that was identified; and

(f) based on selected features detected, defining a classifier for each type of white blood cell, for use in automated identification of white blood cells labeled with the nuclear dye or stain that are included in subsequent samples processed with the imaging system.

19. The method of claim 18, wherein defining the classifier for each type of white blood cell comprises using a linear discriminant analysis to produce a linear weighted combination of selected features identified in the bright field, side scatter, and nuclear fluorescence images.

20. The method of claim 19, wherein the features are selected for each type of white blood cell based on their ability to discriminate a type of white blood cell from other types of white blood cells.

21. The method of claim 18, wherein the monoclonal antibodies used for labeling the white blood cells in the residual sample comprise:

(a) an anti-CD45 monoclonal antibody;

(b) an anti-CD14 monoclonal antibody;

(c) an anti-CD123 monoclonal antibody; and

(d) an anti-CD193 monoclonal antibody.

22. The method of claim 18, wherein the nuclear dye or stain comprises one of:

(a) 4′

,6-diamidino-2-phenylindole (DAPI);

(b) a cell-permeant cyanine nucleic acid stain;

(c) an A-T intercalating anthraquinone stain;

(d) 7-Aminoactinomycin D (7-AAD);

or

(e) propidium iodide.

23. The method of claim 18, wherein reducing the number of red blood cells in the donor blood sample to produce the residual sample comprises at least one selected from the group consisting of:

(a) subjecting the stained sample to an acoustic force that shifts the red blood cells out of a flow stream entering an imaging region of the imaging system;

(b) filtering the stained sample with a filter screen sized to allow the red blood cells that are smaller than the white blood cells to pass through the filter screen, while blocking the white blood cells, so that the white blood cells are conveyed by the flow stream into the imaging region of the imaging system, but the red blood cells are not; and

(c) lysing the red blood cells in the donor blood sample and separating lysed red blood cells from the white blood cells using centrifugal force.

24. The method of claim 18, further comprising applying fixation and permeabilization of the white blood cells before labeling the white blood cells with the nuclear dye or stain and with the monoclonal antibodies.

25. The method of claim 18, further comprising collecting additional donor blood samples and repeating the method of claim 18 for the additional donor blood samples to improve a sensitivity and specificity of the classifier used for identifying each type of white blood cell in blood samples subsequently processed by the imaging system.