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

• US 20020150285A1
• Filed: 01/22/2002
• Published: 10/17/2002
• Est. Priority Date: 03/28/2001
• Status: Active Grant

• Current Assignee: VisionGate, Inc.
• Original Assignee: VisionGate, Inc.
• Inventors: Nelson, Alan C.
• Granted Patent: US 6,519,355 B2
• Time in Patent Office: Days
• US Class Current: 382/131
• CPC Class Codes: G01N 15/1433 using image recognition
• CPC Class Codes: G01N 15/147 the analysis being performed on a sample stream
• CPC Class Codes: G01N 21/05 Flow-through cuvettes G01N21/09 takes precedence; handling fluid samples G01N1/10
• CPC Class Codes: G01N 2223/419 computed tomograph
• CPC Class Codes: G01N 2223/612 biological material
• CPC Class Codes: G01N 23/046 using tomography, e.g. computed tomography [CT]
• CPC Class Codes: G06T 2207/10101 Optical tomography; Optical coherence tomography [OCT]
• 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 20/693 Acquisition

Abstract

A system and method for rapidly detecting cells of interest using multi-dimensional, highly quantitative, nuclear and cytoplasmic densitometric features (NDFs and CDFs) includes a flow optical tomography (FOT) instrument capable of generating various optical projection images (or shadowgrams) containing accurate density information from a cell, a computer and software to analyze and reconstruct the projection images into a multi-dimensional data set, and automated feature collection and object classifiers. The system and method are particularly useful in the early detection of cancers such as lung cancer using a bronchial specimen from sputum or cheek scrapings and cervical/ovarian cancer using a cervical scraping, and the system can be used to detect rare cells in specimens including blood.

Claims

1. A method for detecting cells of interest in 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;

(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.

2. The method of claim 1, 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 and 2D nuclear densitometric features (NDFs), cytoplasmic densitometric features (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.

3. The method of claim 1, wherein the step of using the at least one classifier comprises using a classifier having at least one discriminant function.

4. The method of claim 3 wherein the at least one discriminant function uses a set of discriminant features selected from the group consisting of 1D and 2D nuclear densitometric features (NDFs), cytoplasmic densitometric features (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.

5. The method of claim 1 wherein the cell sample comprises a human lung specimen.

6. The method of claim 1 wherein the cell sample comprises a human cervical specimen.

7. The method of claim 1 wherein the cell sample comprises a human blood sample.

8. The method of claim 1 wherein the cell sample comprises rare cells.

9. The method of claim 1 wherein the step of marking the sample comprises staining with a stain selected from the group consisting of stoichiometric DNA stain, stoichiometric RNA stain, proportional DNA stain, and proportional RNA stain.

10. The method of claim 9 wherein the step of marking the sample comprises staining with a stain selected from the group consisting of Feulgen stain, Romanowski stain, May-Grunwald-Giemsa stain, Methyl Green and thionin.

11. The method of claim 1 wherein the step of marking the sample comprises staining with an antibody marker.

12. The method of claim 1 wherein the step of marking the sample comprises staining with a nucleic acid sequence marker.

13. The method of claim 1 further comprising the step of selecting the at least one object of interest to include cancer diagnostics.

14. The method of claim 1 further comprising the step of selecting the at least one object of interest to include a preinvasive cancer cell.

15. The method of claim 1 further comprising the step of selecting the at least one object of interest to include an invasive cancer cell.

16. The method of claim 1 further comprising the step of using the at least one object of interest for screening a patient for cancer.

17. The method of claim 1 further comprising the step of using the at least one object of interest for determining whether a patient will develop invasive cancer.

18. The method of claim 14 wherein the preinvasive cancer cell is derived from an epithelial cancer.

19. The method of claim 18 wherein the epithelial cancer is selected from the group consisting of lung cancer, throat cancer, cervical cancer, ovarian cancer, breast cancer, prostate cancer, skin cancer and cancer of the gastrointestinal tract.

20. The method of claim 15 wherein the invasive cancer cell is derived from an epithelial cancer.

21. The method of claim 20 wherein the epithelial cancer is selected from the group consisting of lung cancer, throat cancer, cervical cancer, ovarian cancer, breast cancer, prostate cancer, skin cancer, cancer of the gastrointestinal tract, lymphatic cancer and bone cancer.

22. The method of claim 14 wherein the preinvasive cancer cell is derived from a neuroendocrine cancer.

23. The method of claim 22 wherein the neuroendocrine cancer is selected from the group consisting of lung and throat cancer, cervical cancer, breast cancer, and cancer of the gastrointestinal tract, lymphatic cancer and bone cancer.

24. The method of claim 15 wherein the invasive cancer cell is derived from a neuroendocrine cancer.

25. The method of claim 24 wherein the neuroendocrine cancer is selected from the group consisting of lung cancer, throat cancer, cervical cancer, breast cancer, and cancer of the gastrointestinal tract, lymphatic cancer and bone cancer.

26. The method of claim 1 further including the step of combining at least two projection images using methods for computerized image reconstruction to generate multi-dimensional images.

27. The method of claim 26 wherein the computerized image reconstruction methods include methods selected from the group consisting of fan beam projection geometries and cone beam projection geometries.

28. The method of claim 26 further comprising steps wherein the multi-dimensional images, are processed as follows:

(a) compensating the multi-dimensional images for variations in background illumination;

(b) analyzing the multi-dimensional images to detect at least one object of interest;

(c) calculating a surface that bounds each object of interest;

(d) calculating a set of multi-dimensional feature values for each object of interest; and

(e) providing the set of multi-dimensional feature values to at least one classifier that identifies and characterizes cells of interest in the cell sample.

29. The method of claim 28 wherein the multi-dimensional feature values used to classify cells of interest in the sample are selected from the group consisting of 1D nuclear densitometric features (NDFs), 2D NDFs, 3D NDFs and 3D+ NDFs, 1D cytoplasmic densitometric features (CDFs), 2D CDFs, 3D CDFs and 3D+ CDFs, area, mean radius, volume, mean volume, optical density (OD) variance, OD skewness, OD range, OD average, OD maximum, density of light spots, low DNA volume, high DNA volume, 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.

30. The method of claim 28 wherein the multi-dimensional feature values include 1D, 2D, 3D and 3D+ features that are utilized by a classification program to characterize cells of interest with regard to type, maturation, disease status, relative abundance and presence of quantitative markers.

31. The method of claim 28 wherein the classifier detects and characterizes cells of interest including rare cells.

32. The method of claim 28 further including the step of diagnosing the cell sample based upon the characterization of cells.

34. The method of claim 1 wherein the at least one projection image comprises a shadowgram.

33. 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.