Method and apparatus for computer controlled cell based diagnosis Number:7,522,757 from the United States Patent and Trademark Office (PTO) owispatent

Title: Method and apparatus for computer controlled cell based diagnosis

Abstract: A computer controlled method for detecting and diagnosing a rare cell type in a tissue sample is provided, said method comprising treating the tissue sample such that it generates a first signal indicative of the presence at a location of a rare cell, detecting the first signal, treating the location at which the first signal is detected to generate a second signal indicative of a diagnostically useful cellular characteristic and detecting the second signal. The first signal can be morphological or a color present in a sought cell either before or after staining. The second signal can be generated by in situ PCR or PCR in situ hybridization. In one preferred embodiment, the rare cell type is a fetal cell in a maternal blood tissue sample, said sample consisting of a smear of unenriched maternal blood. In another embodiment, the method is used to diagnose or genotype cancer cells in a blood or tissue biopsy sample.

Patent Number: 7,522,757 Issued on 04/21/2009 to Tsipouras,   et al.

---

Inventors:	Tsipouras; Petros (Madison, CT), Tafas; Triantafyllos (Rocky Hill, CT)
Assignee:	Ikonisys, Inc. (New Haven, CT)
Appl. No.:	11/264,273
Filed:	November 1, 2005

---

Related U.S. Patent Documents

---

	Application Number	Filing Date	Patent Number	Issue Date
	10130559
	PCT/US99/27608	Nov., 1999

---

Current U.S. Class:	382/133 ; 382/128; 382/134
Current International Class:	G06K 9/00 (20060101)
Field of Search:	382/128-134

---

References Cited [Referenced By]

---

U.S. Patent Documents

4513438	April 1985	Graham et al.
4983044	January 1991	Schweber
5352613	October 1994	Tafas et al.
5740269	April 1998	Oh et al.
5764792	June 1998	Kennealy
5889881	March 1999	MacAulay et al.
6136540	October 2000	Tsipouras et al.
6221596	April 2001	Yemini et al.

Foreign Patent Documents

	0 595 506		May., 1994		EP
	0 713 086		Apr., 1999		EP
	WO 93/18186		Sep., 1993		WO
	WO 94/02646		Feb., 1994		WO
	WO 97/20198		Jun., 1997		WO
	WO 97/43732		Nov., 1997		WO
	WO 99/02960		Jan., 1999		WO
	WO 99 08091		Feb., 1999		WO
	WO 01/37192		Nov., 1999		WO
	WO 99/58972		Nov., 1999		WO

Other References

Baxes, 1994, "Digital Image Processing, Passage," U.S., New York, Wiley, pp. 127-137. cited by other . Lizardi, et al., 1998, "Mutation Detection and Single-Molecule Counting Using Isothermal Rooling-Circle Amplification," Nature Genetics, 19(3):225-232. cited by other . Mesker, et al., 1994, "Detection of Immunocytochemically Stained Rare Events Using Image Analysis," Ctyochemistry 17:209-215. cited by other . Oosterwijk, et al., 1998, "Strategies for Rare-Event Detection: An Approach for Automated Fetal Cell Detection in Maternal Blood," Am. J. Hum. Genet. 63:1783-1792. cited by other . Oosterwijk, et al., 1998, "Development of a Preparation and Staining Method for Fetal Erythroblasts in Maternal Blood: Simultaneous Immunocytochemical Staining and FISH Analysis," Cytometry 32:170-177. cited by other . Oosterwijk, et al., 1998, "Fetal Cell Detection in Maternal Blood: A Study in 236 Samples Using Erythroblast Morphology, DAB and HbF Staining, and FISH Analysis," Cytometry 32:178-185. cited by other . Tanke, et al., 1996, "Detection of "Rare Event" Fetal Erythroblasts in Maternal Blood Using Automated Microscopy," Early Hum. Devel. 47 Suppl.:S89-S93. cited by other . Verwoerd, et al., 1987, "Somatic Cell Mutations in Humans Detected by Image Analysis of Immunofluorescently Stained Erythrocytes," in: Clinical Cytometry and Histometry, Burger et al., eds., Academic Press, pp. 465-469. cited by other . "XL Vision Announces Advanced Imaging Technology For Early Detection of Metastatic Cancer," Press Release dated Dec. 18, 1995, XL Vision, Sebastian, FL, pp. 7-11. cited by other.

Primary Examiner: Tucker; Wesley
Attorney, Agent or Firm: Kelley Drye & Warren LLP

---

Parent Case Text

---

This application is a continuation of U.S. application Ser. No. 10/130,559, filed on May 17, 2002, which is a national phase application of PCT/US99/27608 (WO 01/37192), filed on Nov. 18, 1999, which disclosures are herein incorporated by reference in their entirety.

---

Claims

---

What is claimed is:

1. A method in a computer system comprising the steps of: detecting on a microscope slide of a cell sample monolayer containing one or more rare cells; employing detectors, determining in an automated sequence of steps a first signal indicative of the position of rare cell(s) on said microscope slide; using said determined position of said rare cell(s) to automatically position a dispensing system over said determined position; dispensing from said dispensing system a diagnostic label having diagpositic significance to said rare cell(s) in situ; determining if there is the presence of a second signal indicative of said diagnostic label being bound to said rare cell(s).

2. The method of claim 1, wherein the sample of cells are derived from maternal blood and rare cell(s) is a fetal cell.

3. A method of obtaining an image signal of a rare cell found on a substrate surface, in accordance with claim 1, wherein the step of determining the position of said rare cell based upon detection of pre-determined parameters associated with said rare cell(s) is performed by a computerized automated microscopy system.

4. The method of claim 3, wherein said computerized automated microscopy system used has two or more objectives.

5. A method of obtaining an image signal of a rare cell found on a substrate surface, in accordance with claim 1, wherein the step of treating said cell sample with a diagnostic label by a label or stain dispensing system is performed by a computerized label or stain dispensing system.

6. A method of obtaining an image signal of a rare cell found on a substrate surface. in accordance with claim 1, wherein the step of detecting whether a signal indicative of said diagnostic label is associated with said determined position of said rare cell(s) is performed by a computerized automated microscopy system.

7. The method of claim 6, wherein said computerized automated microscopy system has two or more objectives.

8. A method of obtaining an image signal of a rare cell found on a substrate surface, in accordance with claim 1, wherein the step o providing a diagnosis based on detection of said diagnostic label being associated with the position of one or more detected rare cell(s), further comprises the step of recording said rare cell(s) position in a computer memory.

9. A method of obtaining an image signal of a rare cell found on a substrate surface, in accordance with claim 1, wherein said cell sample is enriched by an increased rare cell concentration.

10. A method in a computer system, in accordance with claim 1, further comprising providing a putative diagnosis if a second signal, indicative of said second diagnostic label being bound to said rare cell(s), is detected.

11. A computer implemented method of automatically obtaining from a sample of maternal blood containing an unnaturally present concentration of fetal cells, a signal having diagnostic significance relative to the fetal cells, said method comprising an automatically implemented sequence of the steps of: preparing a smear of the sample of maternal blood on a substrate; acquiring/observing the smear image using a computerized microscopic vision system operatively configured to read record said substrate and to obtain a first signal indicative of the presence of a fetal cell; contacting the fetal cell with an agent to generate a second signal, the second signal having said diagnostic significance; acquiring/observing the fetal cell using the computerized microscopic vision system to record/obtain the second signal; and counting occurrences of the second signal in a plurality of fetal cells emitting said first signal.

12. The method of claim 11, wherein the first signal is further processed to represent morphological measurements of the fetal cells.

13. The method of claim 11, wherein the first signal and the second signal do not mask one another when both are present.

14. The method of claim 11, further comprising the step of: calibrating a coordinate system to said substrate so that coordinates of said fetal cell(s) identified at one point in time can be returned to at a later point of time.

15. The method of claim 11, wherein said computerized microscopic vision system has two or more objectives.

16. A computer implemented method, in accordance with claim 11, wherein said sample of maternal blood is enriched by an increased rare cell concentration.

---

Description

---

1. FIELD OF THE INVENTION

The present invention relates to computer controlled methods and apparatus for obtaining and preparing cell samples and for identifying a rare cell of interest from a field of cells and making a diagnosis based on a characteristic of a rare cell selected in the field. In one important embodiment, the invention relates to obtaining and preparing a maternal blood sample for fetal cell based prenatal diagnosis. Diagnosis performed by embodiments of the invention includes employing a computer to identify, characterize, and count objects in the optical field which are tagged using fluorescence in situ hybridization to visualize selected objects in the optical field.

2. BACKGROUND OF THE INVENTION

The advent of DNA based prenatal diagnosis for human genetic disorders has led to the development of a number of new diagnostic methods. These diagnostic methods permit early detection and consequently informed decisions and intervention with respect to fetus having a genetic disorder. These methods, however, have a number of disadvantages. Each of the new diagnostic methods with which this discussion is concerned requires that a sample of isolated fetal cells be obtained, so that the DNA of the fetus may be examined or tested for signs of specific genetic disorders.

The disadvantages of these modern methods are at least two-fold. First, there is a need to obtain a sample of fetal cells. Currently, fetal cells are obtained by invasive procedures requiring obstetric intervention by amniocentesis or by chorionic villus sampling. These highly specialized procedures carry a small, but significant, risk to the fetus. Early in pregnancy, the level of risk to the fetus is high and the number of cells obtained is low. Therefore, results of these procedures often are not obtained until 18-20 weeks of pregnancy. Second, accurately assessing, quantifying, and assigning a significance to images of cells is difficult, time-consuming and unacceptably subjective.

One modem procedure for obtaining fetal cells relies on leakage of fetal cells into the maternal circulation. By simply drawing a sample of maternal blood, it is theoretically possible to obtain fetal cell material in a sufficient quantity for prenatal diagnosis by DNA based methods. Obtaining fetal cells from the maternal blood circulation avoids any risk to the fetus and can be undertaken as early as 10-12 weeks of pregnancy.

Fetal cells which have been detected in the maternal blood circulation include trophoblasts, lymphocytes and nucleated erythrocytes. Trophoblasts were the first fetal cells to be identified in the maternal blood circulation, due to their large size. However, nucleated erythrocytes have generated the greatest degree of interest as sources of genetic material for prenatal diagnosis due to their rarity in the adult blood circulation, their abundance in fetal blood and their limited life span. These factors combine to reduce errors in distinguishing fetal cell material from maternal cell material. Fetal cells circulating in the maternal blood have a life span ranging from a few weeks (for the nucleated erythrocytes) to a few years (for the lymphocytes).

Although they are consistently present in the maternal blood circulation, fetal cells are very rare, severely limiting their diagnostic utility. Estimates of the concentration of fetal cells within the maternal blood circulation vary widely, from a high level of 1 fetal cell in 10.sup.5 maternal cells, to a low level of one fetal cell in 10.sup.9 maternal cells. Thus, a 10 ml sample of maternal blood will ordinarily contain between about 10 and 100 fetal cells. Throughout this description, the concentration of fetal cells found in a freshly drawn maternal blood sample, prior to any further treatment, is referred to as the "naturally present concentration" of fetal cells, typically, but not necessarily, within the above ranges. Also throughout this description, the term "unenriched maternal blood" shall refer to a sample of maternal blood which contains only a naturally present concentration of fetal cells.

Since the naturally present concentration of fetal cells in unenriched maternal blood is so low, in order to obtain a diagnostically significant sample of fetal cells modern techniques include methods of physically isolating the fetal cells from the maternal cells in the sample. In essence, modern techniques are methods of concentrating the fetal cells within a sample, i.e., enriching the sample, for example by removing excess maternal cells, without removing fetal cells. These methods are extremely difficult to perform, often fail to isolate a sufficient number of fetal cells to be diagnostically significant and sometimes fail to provide a sample of a sufficient number of undamaged fetal cells of adequate purity for reliable subsequent diagnosis.

The normal human complement of chromosomes consists of the sex chromosomes (designated X and Y) and 22 autosomes (numbered 1-22). It has been estimated that a minimum of 1 in 10 human conceptions has a chromosome abnormality. As a general rule, an abnormal number of sex chromosomes is not lethal, although infertility can result. In contrast, an abnormal number of autosomes typically results in early death. Of the three autosomal trisomies found in live-born babies (trisomy 21, 18 and 13), only individuals with trisomy 21 (more commonly known as Down's syndrome), survive past infancy.

Although Down's syndrome is easily diagnosed after birth, prenatal diagnosis is problematic. To date, karyotyping of fetal cells remains the established method for the diagnosis of Down's syndrome and other genetic abnormalities associated with an aberration in chromosomal number and/or arrangement. Such genetic abnormalities include, for example, chromosomal additions, deletions, amplifications, translocations and rearrangements. The assessment of such abnormalities is made with respect to the chromosomes of a healthy individual, i.e., an individual having the above-described normal complement of human chromosomes.

Genetic abnormalities include the above-noted trisomies, such as Down's syndrome, as well as monosomies and disomies. Genetic abnormalities also include additions and/or deletions of whole chromosomes and/or chromosome segments. Alterations such as these have been reported to be present in many malignant tumors. Thus, aberrations in chromosome number and/or distribution (e.g., rearrangements, translocations) represent a major cause of mental retardation and malformation syndromes (du Manoir et al., et al., Human Genetics 90 (6): 590-610 (1993)) and possibly, oncogenesis. See also, e.g., (Harrison's Principles of Internal Medicine, 12th edition, ed. Wilson et al., McGraw Hill, N.Y., N.Y., pp. 24-46 (1991)), for a partial list of human genetic diseases that have been mapped to specific chromosomes, and in particular, for a list of X chromosome linked disorders. In view of the growing number of genetic disorders associated with chromosomal aberrations, various attempts have been reported in connection with developing simple, accurate, automated assays for genetic abnormality assessment.

In general, karyotyping is used to diagnose genetic abnormalities that are based upon additions, deletions, amplifications, translocations and rearrangements of an individual's nucleic acid. The "karyotype" refers to the number and structure of the chromosomes of an individual. Typically, the individual's karyotype is obtained by, for example, culturing the individual's peripheral blood lymphocytes until active cell proliferation occurs, preparing single, proliferating (e.g. metaphase, and possibly, interphase) cells for chromosome visualization, fixing the cells to a solid support and subjecting the fixed cells to in situ hybridization to specifically visualize discrete portions of the individual's chromosomes.

The rapid development of non-isotopic in situ hybridization techniques and the general availability of an ever-expanding repertoire of chromosome-specific DNA probes have extended the number of genetic disorders for which karyotyping is feasible. See, e.g., Lichter et al., "Analysis of Genes and Chromosomes by Non-isotopic in situ Hybridization", GATA 8(1):24-35 (1991). Such methods include the use of probe sets directed to chromosome painting for visualizing one or more preselected chromosomal subregions in a targeted fashion. Methods such as these require at least a modicum of knowledge regarding the types of aberration (s) expected in order to select useful DNA probes complementary to target nucleic acids present in a clinical or tumor cell sample.

Nucleic acid hybridization techniques are based upon the ability of a single stranded oligonucleotide probe to base-pair, i.e., hybridize, with a complementary nucleic acid strand. Exemplary in situ hybridization procedures are disclosed in U.S. Pat. No. 5,225,326 and copending U.S. patent application Ser. No. 07/668,751, the entire contents of which are incorporated herein by reference. Fluorescence in situ hybridization ("FISH") techniques, in which the nucleic acid probes are labeled with a fluorophor (i.e., a fluorescent tag or label that fluoresces when excited with light of a particular wavelength), represents a powerful tool for the analysis of numerical, as well as structural aberrations chromosomal aberrations. See, e.g., PCT Application WO 94/02646, inventors M. Asgari et al., published Feb. 3, 1994, hereinafter, "Asgari") co-pending U.S. patent application Ser. No. 07/915,965; P. Lichter, et al., Genet. Anal. Tech. Appl. 8: 24-35 (1991); and S. Du Manoir, et al., Human Genetics 90 (6): 590-610 (1993), the entire contents of which publications are incorporated herein by reference.

Asgari reports in situ hybridization assays for determining the sex of a fetus, genetic characteristics or abnormalities, infectious agents and the like, by nucleic acid hybridization of fetal cells such as those circulating in material blood. The fetal cells are distinguished from maternal cells present in the fixed sample by staining with an antibody which specifically recognizes the maternal or fetal cell or by in situ hybridization to detect one or more fetal mRNAs. The method reportedly is useful for detecting chromosomal abnormalities in fetal cells. However, the fetal cells must be enriched prior to analysis.

PCT Application WO 94/02830, inventors M. Greaves, et al., published Feb. 3, 1994, (hereinafter, "Greaves") report a method for phenotyping and genotyping a cell sample. The method involves contacting a fixed cell with an antibody labeled with a first fluorophor for phenotyping the cell via histochemical staining, followed by contacting the fixed cell with a DNA probe labeled with a second fluorophor for genotyping the cell. The first and second fluorophors fluoresce at different wavelengths from one another, thereby allowing the phenotypic and genetic analysis on the identical fixed sample.

Despite the above-described advances in the development of fluorescent in situ hybridization methods for the diagnosis of genetic abnormalities, the analysis of the fluorophor-labeled sample remains labor-intensive and involves a significant level of subjectivity. This is particularly true in connection with the prenatal diagnosis of genetic abnormalities in which fetal cells must either be isolated from maternal cells or visually distinguished therefrom prior to assessment for genetic abnormalities. Thus, for example, a laboratory technician must manually prepare and sequentially stain the sample (first, with a histochemical stain to phenotype the cells, second, with a hybridization probe to genotype the cell); visually select fetal cells from other cells in the optical field (using, for example, the above-mentioned histochemical staining procedure); assess the relative distribution of fluorescent color that is attributable to hybridization of the fluorophor-tagged probe; and compare the visually-perceived distribution to that observed in control samples containing a normal human chromosome complement. As will be readily apparent, the above-described procedure is quite time-consuming. Moreover, because the results are visually-perceived, the frequency of erroneous results can vary from one experiment to the next, as well as from one observer to the next.

The discussion thus far has focused on a particular type of rare cell, fetal cells circulating in the mother's peripheral blood, and a particular diagnostic setting, detecting trisomy 21 (Down's syndrome). Many other diagnostic settings are known in which a signal is to be detected in other rare cells. For example, a particular enzyme level or genetic characteristic found in cells of a particular morphology may indicate a condition of medical significance, such as a precancerous condition, a cancerous tumor, a metastasis of a tumor, infection by a virus, and various other genetic conditions, for example.

3. SUMMARY OF THE INVENTION

It is desired to provide a computer controlled method and apparatus for detecting and diagnosing a rare cell type in a tissue sample, said diagnosis being based upon a characteristic of that rare cell. It is further desired to provide a computer controlled method and apparatus for detecting fetal cells in a blood preparation and performing a fetal cell based prenatal diagnosis that solves the above-identified problems, which overcomes such other problems and meets such other goals as will be apparent to the person skilled in this art after reading a description of the invention. It is also desired to provide a computer implemented diagnostic screen for genetic disorders.

The invention in one aspect involves a method of operating a computer system to detect whether a genetic condition defined by at least one target nucleic acid is present in a sample. The method involves the use of probes and digitized images of the probes hybridized to a sample, together with counting objects and analysis of a statistical expectation to detect whether the genetic condition is present. The counting may involve, for example, counting the number of times a genetic abnormality is detected and comparing that count to a statistical expectation of the abnormality in a particular tissue type, cell type or sample. The counting may involve counting the number of times a genetic abnormality occurs and comparing that count to the number of times a cell type occurs in the same sample or to the number of times a normal nucleic acid occurs in the same sample. The counting may involve counting the number of times more than one different genetic abnormality occurs in a single cell. The computer system also may be used to identify cell type, count cells, examine cell morphology, etc. and compare or correlate this information with the count of the genetic abnormality. Various diagnostic circumstances are described below and are known to those of ordinary skill in the art.

One method of counting involves fluorescence in situ hybridization. This aspect of the invention, which is preferred, is exemplary of the methods of the invention.

Generally, the invention provides a method of operating a computer system to detect whether a genetic condition defined by at least one target nucleic acid is present in a fixed sample, the method comprising: receiving a digitized image, preferably a color image, of the fixed sample, which has been subjected to fluorescence in situ hybridization under conditions to specifically hybridize a fluorophor-labeled probe to the target nucleic acid; processing the color image in a computer to separate objects of interest from background; measuring size and color parameters of the objects of interest; identifying first objects having specific predetermined characteristics associated with the target nucleic acid; counting first objects identified; and analyzing the count of first objects with respect to a statistical expectation to detect whether the genetic condition is present. This method is applicable to many genetic conditions, including wherein the genetic condition is human trisomy 21. In addition to the foregoing, it will be understood that the statistical expectation can be based on a tissue type, for example. The computer can be used to identify the tissue type of a cell being examined, but the tissue type also can be known.

In some embodiments, the step of receiving further includes a step of producing an image file of red, green and blue pixels representative of red, green and blue intensities at respective pixel locations within the color image received. In some embodiments, the step of processing further includes steps of manually selecting a plurality of pixels within the background; determining color intensity value ranges corresponding to the portion of the background; and identifying as the background those areas of the image having color intensity values within the ranges determined. In some embodiments, before the step of measuring, there may be processing in the computer to filter the color image to make color intensity values of dark pixels in the color image lighter and to make color intensity values of light pixels in the color image darker. The step of filtering may further comprise passing the color image through a hole filling filter; passing the filled color image through an erosion filter; performing a separate operation on the eroded filled color image, to define outlines around areas; selecting pixels within the outlines by performing a logical NOT operation; and performing a logical AND operation between the selected pixels and the filled color image.

In some embodiments, the genetic condition is further defined by a ratio of the target nucleic acid to a second nucleic acid. Then, the method further includes identifying second objects having specific predetermined characteristics associated with the second nucleic acid; and counting second objects identified; wherein analyzing the count of first objects includes finding a ratio of the count of first objects to the count of second objects. In some embodiments, the target nucleic acid defines a dominant trait and the second nucleic acid defines a corresponding recessive trait. The method in those embodiments may include indicating the genetic condition as possessing the dominant trait, possessing the recessive trait, or possessing the dominant trait and carrying the recessive trait depending on the ratio found. When the target nucleic acid is a rearrangement of the second nucleic acid, the method may further include selecting the probe to hybridize with a break region between rearranged and non-rearranged nucleic acids. Finally, the method may include indicating the genetic condition as a severity level related to the ratio found.

According to the invention, instructions for carrying out the above methods may be recorded on a computer medium. The invention may then provide a computer software product comprising: a computer readable storage medium having fixed therein a sequence of computer instructions directing a computer system to detect whether a genetic abnormality is present in a fixed sample containing at least one target nucleic acid, the instructions directing steps of: receiving a digitized color image of the fixed sample, which has been subjected to fluorescence in situ hybridization under conditions to specifically hybridize a fluorophor-labeled probe to the target nucleic acid; processing the color image in a computer to separate objects of interest from background; measuring size and color parameters of the objects of interest so as to identify and enumerate objects having specific predetermined characteristics associated with the target nucleic acid; and analyzing the enumeration of objects with respect to a statistically expected enumeration to detect whether the genetic abnormality is present. The steps defined by the instructions recorded on the medium may be varied as described above in connection with the method.

A method of operating a computer according to yet another aspect of the invention to count occurrences of a target substance in a cell-containing sample which has been labeled with a target-specific fluorophor may comprise: receiving a digitized color image of the fluorophor-labeled sample; obtaining a color image of the fluorophor-labeled sample; separating objects of interest from background in the color image; measuring parameters of the objects of interest so as to enumerate object having specific characteristics; and analyzing the enumeration of objects with respect to a statistically expected enumeration to determine the genetic abnormality. Again, the steps of the method may be varied as above.

Similarly, there is, according to other aspects of the invention, a computer software product comprising: a computer readable storage medium having fixed therein a sequence of computer instructions directing a computer system to count occurrences of a target substance in a cell-containing sample which has been labeled with a target-specific fluorophor, the instructions directing steps of: receiving a digitized color image of the fluorophor-labeled sample; obtaining a color image of the fluorophor-labeled sample; separating objects of interest from background in the color image; measuring parameters of the objects of interest so as to enumerate object having specific characteristics; and analyzing the enumeration of objects with respect to a statistically expected enumeration to determine the genetic abnormality. The instructions can be made to implement all of the variations on the methods described above.

According to yet another aspect of the invention, there is provided an apparatus for analyzing an image of a cell-containing sample which has been labeled with a target-specific fluorophor, comprising: a computer system on which image processing software executes; and a storage medium in which is fixed a sequence of image processing instructions including receiving a digitized color image of the fluorophor-labeled sample, obtaining a color image of the fluorophor-labeled sample, separating objects of interest from background in the color image, measuring parameters of the objects of interest so as to enumerate object having specific characteristics, and analyzing the enumeration of objects with respect to a statistically expected enumeration to determine the genetic abnormality. Again, the instructions can be varied to implement all the variations described above.

In yet another aspect, the invention provides a computer-implemented method of processing body fluid or tissue sample image data, the method comprising creating a subset of a first image data set representing an image of a body fluid or tissue sample taken at a first magnification, the subset representing a candidate blob which may contain a rare cell creating a subset of a second image data set representing an image of the candidate blob taken at a second magnification, the subset of the second data set representing the rare cell, storing the subset of the second data set in a computer memory, measuring size and color parameters of the objects of interest so as to identify objects having specific predetermined characteristics associated with the target nucleic acid, counting the objects identified in the step of measuring, and analyzing the count of objects with respect to a statistically expected count to detect whether the genetic abnormality is present.

In general, a subset of a first image data set can be created by observing an optical field of a monolayer of cells from a body fluid or tissue sample using a computerized microscopic vision system to detect a signal indicative of the presence of a rare cell.

The method can further produce an image file of red, green and blue pixels representative of red, green and blue intensities at respective pixel locations within the color image received. According to some aspects of the invention, the processing further includes manually selecting a plurality of pixels within the background; determining color intensity value ranges corresponding to the portion of the background; and identifying as the background those areas of the image having color intensity values within the ranges determined. The method may include before the step of measuring, processing in the computer to filter the color image to make color intensity values of dark pixels in the color image lighter and to make color intensity values of light pixels in the color image darker. Moreover, the filtering may include passing the color image through a hole filling filter; passing the filled color image through an erosion filter; performing a separate operation on the eroded filled color image, to define outlines around areas; selecting pixels within the outlines by performing a logical NOT operation, and performing a logical AND operation between the selected pixels and the filled color image.

The method further comprises contacting a body fluid or tissue sample at a location corresponding to each candidate blob represented in the subset of the first image data set, with a reagent to generate a medically significant signal. This method provides the advantage of being able to remove from further processing a body fluid or tissue sample for which no subset of the first data set representing a candidate blob is created. The signal can be measured to determine whether it is a significant signal level. The first and/or the second image data subsets can be transformed into a representation that is more suitable for control and processing by a computer as described herein. In a preferred embodiment, the image data is transformed from an RGB (Red Green Blue) signal into an HLS (Hue Luminescence Saturation) signal. Filters and/or masks are utilized to distinguish those cells that meet preselected criteria and eliminate those that do not, and thus identify rare cells.

In another aspect of the invention, there is provided a method of operating a laboratory service, the method comprising steps of receiving a body fluid or tissue sample, creating a body fluid or tissue sample smear, operating a computerized microscope so that a software program automatically identifies a rare cell in the smear and detecting a medically significant signal in the rare cell.

In yet another aspect of the invention, there is provided computer software product including a computer-readable storage medium having fixed therein a sequence of instructions which when executed by a computer direct performance of steps of detecting and diagnosing a rare cell type. The steps encompass: creating a subset of a first image data set representing an image of a body fluid or tissue sample taken at a first magnification, the subset representing a candidate blob which may contain a rare cell creating a subset of a second image data set representing an image of the candidate blob taken at a second magnification, the subset of the second data set representing the rare cell, storing the subset of the second data set in a computer memory, measuring size and color parameters of the objects of interest so as to identify objects having specific predetermined characteristics associated with the target nucleic acid; counting the objects identified in the step of measuring; and analyzing the count of objects with respect to a statistically expected count to detect whether the genetic abnormality is present.

In general, a subset of a first image data set can be created as described above. The method can further produce an image file of red, green and blue pixels representative of red, green and blue intensities at respective pixel locations within the color image received. According to some aspects of the invention, the processing further includes manually selecting a plurality of pixels within the background; determining color intensity value ranges corresponding to the portion of the background; and identifying as the background those areas of the image having color intensity values within the ranges determined. The method may include before the step of measuring, processing in the computer to filter the color image to make color intensity values of dark pixels in the color image lighter and to make color intensity values of light pixels in the color image darker. Moreover, the filtering may include passing the color image through a hole filling filter; passing the filled color image through an erosion filter; performing a separate operation on the eroded filled color image, to define outlines around areas; selecting pixels within the outlines by performing a logical NOT operation, and performing a logical AND operation between the selected pixels and the filled color image. The steps further encompass contacting a body fluid or tissue sample at a location corresponding to each candidate blob represented in the subset of the first image data set, with a reagent to generate a medically significant signal. This provides the advantage of being able to remove from further processing a body fluid or tissue sample for which no subset of the first data set representing a candidate blob is created. There is an optional step by which the signal can be measured to determine whether it is of a significant level. Another optional step encompasses transformation of one or both of the first and the second image data subsets into a representation that is more suitable for control and processing by a computer as described herein. In a preferred embodiment, the image data is transformed from an RGB (Red Green Blue) signal into an HLS (Hue Luminescence Saturation) signal. Filters and/or masks are utilized to distinguish those cells that meet pre-selected criteria and eliminate those that do not.

According to one aspect of the invention, there is provided a method of preparing a sample of cells for a diagnostic procedure. The sample of cells is obtained and fixed as a monolayer on a substrate, the sample of cells including a rare cell which is present in the sample at no greater than one in every 10,000 cells (i.e. no greater than 0.01%). An optical field covering at least a portion of the sample of cells is observed using a computerized microscopic vision system for a signal indicative of the presence of a rare cell. The signal is detected, and coordinates where the signal is detected are identified, for the diagnostic procedure. In one embodiment the rare cell is present at no greater than 0.001% of the cells. In other embodiments the rare cell is present at no greater than 0.0001%, 0.00001% or even 0.000001%.

In one particularly important embodiment, the rare cell is a fetal cell in a sample of cells from maternal blood. In a preferred embodiment, the sample contains only a naturally present concentration of fetal cells which can be no greater than 0.001%, 0.0001%, 0.00001%, 0.000001% or even 0.0000001%.

In another specific embodiment of the invention, the rare cell type to be detected and diagnosed is a cancer cell found in a sample of cells or tissue from an animal or patient. The sample can be blood or other body fluid containing cells or a tissue biopsy. As an illustration of this embodiment, cancer cell markers described in Section 5, infia, e.g, GM4 protein, telomerase protein or nucleic acids, and p53 proteins or nucleic acids, may be used in the generation of the first or second signal, in a manner to be determined by the specific application of the invention.

In one embodiment of the invention, when the rare cell type is present in the sample, the method of the invention detects the rare cell type at a frequency of no less than 80%. In other embodiments, the detection frequencies are no less than 85%, 90%, 95% and 99%.

According to one particularly important embodiment of the invention, there is provided a method of preparing a sample of blood for a diagnostic procedure, which includes: preparing a smear of a sample of unenriched maternal blood containing a naturally present concentration of fetal cells; observing an optical field covering a portion of the smear using a computerized microscopic vision system for a signal indicative of the presence of a fetal cell; detecting said signal; and identifying, for the diagnostic procedure, coordinates within the smear at which the signal is detected.

In one embodiment, the signal is further processed to represent morphological measurements of the rare cell. In another embodiment, the cells are treated with a label to enhance the optical distinction of rare cells from other cells. In this embodiment, the signal can be, for example, from a label which selectively binds to the rare cells. In another embodiment, the diagnostic procedure involves moving to the coordinates identified and magnifying the optical field until the image is of an isolated rare cell.

In some embodiments, the optical field is stepped over a sequence of portions of the cells covering substantially all of the cells. This may be achieved, for example, by moving the cells on the substrate under control of the computerized microscopic vision system relative to a lens of the computerized microscopic vision system. In another embodiment, the coordinates at which the first signal was obtained are identified, and then the rare cell at those coordinates specifically is contacted after the coordinates have been identified.

According to another aspect of the invention, there is provided a method of obtaining from a sample of cells a signal having diagnostic significance relative to a rare cell in the sample of cells. The rare cell is present in the sample at no greater than one in every 10,000 cells. The method includes preparing a monolayer of the sample of cells fixed on a substrate. The rare cell is contacted with an agent to generate a diagnostic signal, the diagnostic signal having the diagnostic significance. The monolayer is observed using a computerized microscopic vision system to obtain the diagnostic signal. In some embodiments, the diagnostic signal can be used to identify the rare cell. In other embodiments, a locating signal can be used to identify the rare cell, and the diagnostic signal is obtained after the cell is located.

In one embodiment, the rare cell is present in the sample at no greater than one in every 10,000 cells (i.e., no greater than 0.01% of the cells). In other embodiments, the rare cell is present at no greater than 0.001%, 0.00001% or even 0.000001%. In one particularly important embodiment, the rare cell is a fetal cell in a sample of cells from maternal blood. Preferably the sample contains only a naturally present concentration of fetal cells which can be no greater than 0.001%, 0.0001%, 0.00001%, 0.000001% or even 0.0000001%.

According to an important embodiment of the invention, there is provided a method of obtaining from a sample of unenriched maternal blood, containing a naturally present concentration of fetal cells, a signal having diagnostic significance relative to the fetal cells. The method includes: preparing a smear of the sample of unenriched maternal blood; observing the smear using a computerized microscopic vision system to obtain a first signal indicative of the presence of a fetal cell; contacting the fetal cell with an agent to generate a second signal, the second signal having the diagnostic significance; and observing the fetal cell using the computerized microscopic vision system to obtain the second signal. In one embodiment, the smear can comprise at least 250 .mu.l of the unenriched maternal blood and even can comprise at least 500 .mu.l of the unenriched blood.

As described above, the first signal can be further processed to represent morphological measurements of the rare cell. Likewise, the cells can be treated with a label to enhance optical distinctions of rare cells from other cells, such as maternal cells. To achieve this, the first signal can be from a label which selectively binds to the rare cell, such as a fetal cell. Likewise, as above, the step of observing can involve stepping an optical field over a sequence of portions of the cells, which can be accomplished, for example, by moving the cells or the substrate under control of the computerized microscopic vision system relative to a lens of the computerized microscopic vision system.

In any of the foregoing embodiments, the cells can be prepared on a substrate, and a coordinate system can be calibrated to the substrate so that coordinates of the rare cell identified in one step can be returned to later in another step. Likewise, the substrate in certain important embodiments has a length that is 10 times its width, the substrate being substantially elongated in one direction. The length can even be 20 times the width. The substrate can be a flexible film, and in one important embodiment, is an elongated flexible film that can carry a relatively large volume of cells, such as would be provided from a relatively large volume of smeared maternal blood.

In any of the foregoing embodiments, the first signal and the second signal can be selected whereby they do not mask one another when both are present. Likewise, in any of the foregoing embodiments, the second signal can be generated by in situ PCR or PCR in situ or fluorescence in situ hybridization (FISH). Alternatively, in a particularly preferred embodiment, in any of the foregoing embodiments, the second signal can be generated by rolling-circle amplification (RCA). RCA generates a single stranded DNA (ssDNA) comprising tandem repeats of a target gene sequence. In a specific mode of signal generation using RCA, one or a combination of detectable labels, such as fluorophores useful for multi-parametric are coupled to nucleic acid probes complementary to the ssDNA generated by RCA color coding are used in an RCA format designated "condensation of amplified circles after hybridization of encoding tags" (RCA-CACHET). In another mode of signal generation using RCA, a hapten is incorporated into the ssDNA and detected by means of immunocytochemistry.

In an alternative embodiment, the first step of the process is performed using fluorescence microscopy, which enables identification of the possible rare cell positions at even lower magnification and higher processing speed compared with the method described above. The rare cells of interest are stained with a fluorescent label or fluorophore.

In one important embodiment, the substrate is a plurality of substrates on which the sample of cells is prepared, such as a plurality of smears of maternal blood, each of the plurality including a total of at least 5 .mu.l of the.sample. A rare cell-containing substrate (in which the first signal is obtained) is identified. Then, only the rare cell-containing substrate/substrates which has/have been identified is/are treated to generate the second signal.

According to yet another aspect of the invention, there is provided a method of performing a diagnosis for a fetus, using an unenriched sample of maternal blood containing naturally present fetal cells. This method includes: preparing a smear of at least 250 .mu.l of the sample of unenriched maternal blood; identifying a fetal cell within the smear; contacting the fetal cell with an agent that produces a diagnostic signal; and observing the diagnostic signal. In one important embodiment, the step of identifying can comprise observing cells within the smear, using a computerized microscopic vision system, measuring a signal produced by the observed cells indicative of the presence of fetal cells, and defining coordinates at which the measured signal indicates the presence of the fetal cell. Important embodiments directed to volumes of the maternal blood, substrate configurations and so forth are as described above.

According to yet another aspect of the invention, there is provided a method of obtaining from a sample of unenriched maternal blood containing a naturally present concentration of fetal cells, an image of a substantially isolated fetal cell. This method includes: preparing a smear of at least 250 .mu.l of the sample of unenriched maternal blood; observing the smear with a computerized microscopic vision system for a signal indicative of the presence of a fetal cell; identifying coordinates at which the signal is observed; moving to the coordinates identified, an optical field including an image of the fetal cell; and magnifying the optical field until the image is of an isolated fetal cell. Important embodiments directed to volumes of the maternal blood sample, substrate configurations and so forth are as described above.

According to yet another aspect of the invention, there is provided a device for screening fetal cells contained within a smear of an unenriched sample of maternal blood containing a naturally present concentration of fetal cells, comprising: a flexible film having thereon a smear of at least 250 .mu.l of maternal blood. In one embodiment, the flexible film has thereon a smear of at least 500 .mu.l of maternal blood. In one important embodiment, the flexible film is an elongated film, the length being at least 10 times the width. It is particularly preferred that the flexible film include marking coordinates, whereby the computerized microscopic vision system described herein can locate a cell relative to a point on the film, permitting that cell to be returned to at a later time, if desired.

According to yet another aspect of the invention, there is provided a device for screening rare cells contained within a sample of cells at a concentration of no greater than one rare cell for every 10,000 cells in the sample of cells. The device is a flexible film having fixed thereon the sample of cells, wherein the flexible film is at least five inches long. In one preferred embodiment the flexible film has a length at least 10 times its width. In another important embodiment, the flexible film includes marking coordinates, whereby the computerized microscopic vision system described herein can locate a cell relative to a point on the film, permitting the cell to be returned to at a later time, if desired.

According to another aspect of the invention, there is provided a device for dispensing materials to a specific location on a slide. The device includes a microscopic vision system for detecting a signal indicative of the presence of a rare cell in a sample of cells. The device also includes means for identifying the coordinates of the rare cell in an optical field. The device further has attached to it a dispenser for dispensing a volume of material and means for moving the dispenser to the coordinates whereby the volume of material may be dispensed upon the rare cell. The material dispensed can be reagents such as a label, PCR, primers, and the like. According to another important embodiment of the invention, the need for scanning large areas of microscopic preparations in the minimum possible amount of time is met by the use of an apparatus or system that provides a "composed" image. It is based on the simultaneous use of an array of computer controlled objective lenses, arranged on a support system and having the capacity to focus on a microscopic preparation. Each of the objective lenses is connected to a charge coupled device camera, herein referred to as a CCD camera, being connected to image acquisition hardware installed in a host computer. The images are stored in the computer memory and they are combined in an appropriate side to side fashion, so that a "composed" image is formed in the computer memory. The "composed" image can be further processed as a unity, using any kind of imaging procedures to detect specific features that are in question. The significant advantage of the described system consists in its capacity to acquire images simultaneously from a number of objective lenses, thus minimizing the time needed to process large areas of the sample in a manner that is inversely proportional to the number of objectives used.

The "composing" system can process any kind of microscopic preparation using either transmitted or reflected light. It is particularly useful where large numbers of samples need to be processed imposing significant time constraints, for example, for processing large numbers of microscopic biological preparations for screening and/or diagnostic purposes, etc.

4. BRIEF DESCRIPTION OF THE DRAWINGS

In the accompanying drawings, in which like reference designations indicate like elements:

FIG. 1 is a flow chart summarizing the method of one aspect of the invention;

FIG. 2 is a block diagram of an analysis system used in one embodiment of one aspect of the invention;

FIG. 3 is a flow chart of stage I leading to detecting the first signal;

FIGS. 4A and 4B taken together are a flow chart of stage II leading to detecting the first signal;

FIG. 5 is a flow chart of detection of the second signal;

FIG. 6 is a schematic representation of a variation of an apparatus embodying aspects of the invention, using a continuous smear technique;

FIG. 7 is a block diagram of an analysis and reagent dispensing system used in one embodiment of one aspect of the invention;

FIG. 8 is an outline of a multiple objective microscopy system;

FIG. 9 is an image "composition" method;

FIG. 10 is a flowchart of the calibration steps of one embodiment of the invention;

FIG. 11 is a flowchart of the preprocessing steps of one embodiment of the invention; and

FIGS. 12A and 12B are a flowchart of the main processing steps of one embodiment of the invention.

5. DETAILED DESCRIPTION OF THE INVENTION

The invention will be better understood upon reading the following detailed description of the invention and of various exemplary embodiments of the invention, in connection with the accompanying drawings. While the detailed description explains the invention with respect to fetal cells, a rare cell type, and blood as the body fluid or tissue sample, it will be clear to those skilled in the art that the invention can be applied to and, in fact, encompasses diagnosis based on any cell type and any body fluid or tissue sample for which it is possible to create a monolayer of cells on a substrate.

Body fluids and tissue samples that fall within the scope of the invention include but are not limited to blood, tissue biopsies, spinal fluid, meningeal fluid, urine, alveolar fluid, etc. For those tissue samples in which the cells do not naturally exist in a monolayer, the cells can be dissociated by standard techniques known to those skilled in the art. These techniques include but are not limited to trypsin, collagenase or dispase treatment of the tissue.

In an important specific embodiment, the invention is used to detect and diagnose fetal cells. Our approach is directly opposite to that taken by others seeking a non-invasive method for performing fetal cell based prenatal diagnosis. Rather than attempting substantially to enrich the concentration of fetal cells within a maternal blood sample, our approach involves identifying fetal cells within an unenriched maternal blood sample and subsequently performing diagnostic procedures on the fetal cells so identified, in situ.

A summary of this new approach, shown in the flow chart of FIG. 1, is as follows: Prepare one or more blood smears from a sample of unenriched maternal blood 101; Screen the one or more blood smears until a predetermined number of fetal cells (e.g., nucleated erythrocytes) have been identified and their coordinates defined 103; and Process those smears or coordinates of a smear at which fetal cells have been identified, diagnosing the presence or absence of a particular genetic feature in the fetal cells 105.

In this approach, two signals are defined, referred to hereinafter as the first signal and the second signal. As used herein, "signal" should be taken in its broadest sense, as a physical manifestation which can be detected and identified, thus carrying information. One simple and useful signal is the light emitted by a fluorescent dye selectively bound to a structure of interest. That signal indicates the presence of the structure, which might be difficult to detect absent the fluorescent dye. An alternating signal is the presence (or absence) of detectable structures having predetermined morphological characteristics, such as shape and size.

Screening 103 is based on the first signal. The first signal, which in this exemplary embodiment indicates cell identity, may be generated by a fluorescent dye bound to an antibody against the hemoglobin .epsilon.-chain, i.e., embryonal hemoglobin, for example. Alternatively, for example, a metric of each cell's similarity to the characteristic morphology of nucleated erythrocytes, discerned using cell recognition algorithms may serve as the first signal. In yet another example, the first signal may be a measure of the presence of the characteristic color of fetal hemoglobin after staining with eosin and acid hematoxylin. It should now be evident that in embodiments which diagnose fetal conditions any detectable indicator of the presence of fetal cells may serve as the first signal, subject to certain constraints noted below.

Diagnosing 105 can be based on the second signal (or on a combination of a first and second signal). The second signal, which in this exemplary embodiment indicates the presence of a particular genetic characteristic being tested for, may be generated, for example, by in situ PCR-amplification or PCR in situ hybridization or FISH. In another embodiment, the second signal which indicates the presence of a particular genetic characteristic being tested for may be generated using RCA technology as described by Lizardi et al., 1998, Nature Genetics 19: 225-232, entitled "Mutation Detection and Single-Molecule Counting Using Isothermal Rolling-Circle Amplification", the entire disclosure of which is incorporated by reference in its entirety, see, id., FIGS. 1, 4, and 6, incorporated specifically herein by reference; see also, Nilsson et al., 1994, Science 265: 2085-2088; Nilsson et al., 1997, Nature Genet. 16: 252-255; Fire et al., 1995, Proc. Nat'1. Acad. Sci. USA, 92: 4631-4645; and Liu et al., 1996, J. Am. Chem. Soc. 118:1587-1594, each of which is incorporated herein by reference in its entirety. Cells that emit both signals, i.e., the cell is a fet