This application is related to Japanese Patent Application No. 2001-226382 filed on Jul. 26, 2001, whose priority is claimed under 35 USC § 119, the disclosure of which is incorporated by reference in its entirety.
1. Field of the Invention
The present invention relates to a method of classifying particles on a two-dimensional frequency distribution map and a blood analyzer utilizing the same. In particular, it relates to a method and an apparatus for classifying mature blood cells and immature blood cells contained in blood on the two-dimensional frequency distribution map.
2. Description of Related Art
As the above-mentioned method, conventionally known is a method of classifying mature erythrocytes and reticulocytes contained in blood on a one-dimensional frequency distribution map into two particle clusters by obtaining a distribution peak on the one-dimensional frequency distribution map, setting a threshold value of the frequency in accordance with the peak and determining the two clusters by the threshold value (for example, see Japanese Examined Patent Publication No. 2674705).
An object of the present invention is to improve accuracy in classifying particles indicated on a two-dimensional distribution map.
The present invention provides a method of classifying particles indicated on a two-dimensional frequency distribution map into particle clusters, comprising the steps of: dividing the particles into a first cluster and a second cluster by a line containing a mode coordinate of the particles, presuming a third cluster so that the first and third clusters are symmetrical with respect to the mode coordinate, calculating variance and covariance of a cluster including the first and third clusters to obtain an ellipse region surrounding the first and third clusters based on the calculated variance and covariance, and determining the particles in the obtained ellipse region as a particle cluster.
These and other objects of the present application will become more readily apparent from the detailed description given hereinafter. However, it should be understood that the detailed description and specific examples, while indicating preferred embodiments of the invention, are given by way of illustration only, since various changes and modifications within the spirit and scope of the invention will become apparent to those skilled in the art from this detailed description.
A method of the present invention for classifying particles indicated on a two-dimensional frequency distribution map into particle clusters, includes the steps of: dividing the particles into a first cluster and a second cluster by a line containing a mode coordinate of the particles, presuming a third cluster so that the first and third clusters are symmetrical with respect to the mode coordinate, calculating variance and covariance of a cluster including the first and third clusters to obtain an ellipse region surrounding the first and third clusters based on the calculated variance and covariance, and determining the particles in the obtained ellipse region as a particle cluster.
According to the present invention, the particles may include blood cells and the two-dimensional distribution map may indicate a particle cluster including mature granulocytes and immature granulocytes of leukocytes or a particle cluster including mature erythrocytes and reticulocytes. In this cases, a cluster of mature granulocytes or a cluster of mature erythrocytes can be classified by the ellipse region calculated by the method of the invention.
In the case of classifying the mature granulocytes and the immature granulocytes of the leukocytes, the two-dimensional distribution map may be a scattergram based on a side fluorescent light intensity and a side scattered light intensity detected by using a flow cytometer from a sample containing leukocytes which have been subjected to nucleic acid staining.
Further, in the case of classifying the mature erythrocytes and the reticulocytes, may be used as the distribution map a scattergram based on the side fluorescent light intensity and the forward scattered light intensity detected by using the flow cytometer from a sample containing erythrocytes which have been subjected to fluorescent staining.
The present invention further provides a blood analyzer which analyzes blood by utilizing the above-described fractioning method.
In another aspect, the present invention provides a blood analyzer comprising: a detecting section for detecting characteristic parameters of particles contained in blood, a distribution map preparing section for preparing a two-dimensional frequency distribution map of the particles based on the detected parameters, and a classifying section for classifying the particles on the distribution map, wherein the classifying section carries out the steps of: dividing the particles into a first cluster and a second cluster by a line containing a mode coordinate of the particles; presuming a third cluster so that the first and third clusters are symmetrical with respect to the mode coordinate; calculating variance and covariance of a cluster including the first and third clusters to obtain an ellipse region surrounding the first and third clusters based on the calculated variance and covariance; and determining the particles in the obtained ellipse region as a particle cluster.
Hereinafter, the present invention is detailed by way of an embodiment with reference to
Structure of a Blood Analyzer
On the other hand, side scattered light emitted from the blood cells passing through the orifice 13 enters a photomultiplier tube 29 via a condenser lens 27 and a dichroic mirror 28. Further, side fluorescent light emitted from the blood cells passing through the orifice 13 enters a photomultiplier tube 31 via the condenser lens 27, the dichroic mirror 28, a filter 36 and a pinhole plate 30.
A forward scattered light signal output from the photodiode 26, a side scattered light signal output from the photomultiplier tube 29 and a side fluorescent light signal output from the photomultiplier tube 31 are amplified by amplifiers 32, 33 and 34, respectively, and input to an analysis section 35.
In a measurement process, valves 46 and 47 are opened to suck a blood-containing sample liquid under a negative pressure applied by a suction device 49 out of a reaction chamber 48 in which the sample liquid is reacted with a reagent. When the path between the valve 46 and the nozzle 6 is filled with the sample liquid, the valves 46 and 47 are closed. Then, the valve 50 is opened, thereby the sheath liquid is fed from the sheath liquid chamber 42 to the cell 1 under the positive pressure applied by the pressurizing device 43 and drained into the drain chamber 45.
When the valve 41 is opened, the pressure applied by the pressurizing device 43 is transmitted to the tip of the nozzle 6 via the quantifying syringe 44. Thereby, the pressure of the sheath liquid outside the nozzle and that of the sheath liquid inside the nozzle are balanced at the tip of the nozzle 6. When a piston 44b of the quantifying syringe 44 is driven by a motor 44a in this state, the sample liquid existing between the valve 46 and the nozzle 6 is easily discharged from the nozzle 6 to the orifice 13 and narrowed by the sheath liquid to pass through the orifice 13. The sample liquid is then drained into the drain chamber 45 together with the sheath liquid.
Then, the piston 44b of the quantifying syringe 44 is stopped to finish the measurement process.
Subsequently, the motor 44a is driven in a reverse direction to put the piston 44b back, thereby the quantifying syringe 44 returns to an initial state. During this procedure, the valves 41 and 50 are opened so that the above-mentioned washing process is carried out to get ready for the next measurement process.
The sample liquids contained in the other reaction chambers 51, 52 and 53, respectively, are also measured in sequence by opening and closing valves 54, 55 and 56 in the same manner as the above-described process.
A valve 57 functions to empty the drain chamber 45, so that it is opened and closed as needed.
A condition storing section 62 stores the given conditions and a data storing section 63 stores optical data obtained from the signals output from the photodiode 26 and the photomultiplier tubes 29 and 31. A distribution map preparing section 64 prepares a two-dimensional frequency distribution map (scattergram) based on the optical data stored in the data storing section 63, i.e., two parameters out of a forward scattered light intensity (Fsc), a side scattered light intensity (Ssc) and a side fluorescent light intensity (Sf1). An extracting section 65 extracts coordinates and regions from the distribution map prepared by the distribution map preparing section 64.
A classifying section 66 determines classification regions of particles on the distribution map prepared by the distribution map preparing section 64. A calculating section 67 counts the number of the particles in the classification regions. The calculation results obtained by the calculating section 67 are displayed in a display section 68 together with the distribution map prepared by the distribution map preparing section 64. Further, a fluid system driving section 69 drives the valves 41, 46, 47, 50, 54, 55, 56 and 57 and the motor 44a shown in FIG. 2. The analysis section 35 is constituted by consisted of a personal computer.
Preparation of Two-dimensional Frequency Distribution Maps Depending on a sample, one measurement mode is selected at the input section 61 out of four measurement modes: a nucleated erythrocyte measurement mode; a leukocyte/basophil measurement mode; a leukocyte 4-part differential measurement mode; and a reticulocyte measurement mode. In accordance with the selected mode, blood quantified by a blood quantifying section (not shown) and reagents such as a diluent, a stain solution and a hemolytic agent are contained in the corresponding one of reaction chambers 48, 51, 52 and 53 and the blood is subjected to a predetermined treatment. Blood samples are thus prepared and measured in the sheath row cell 1 in sequence.
In the nucleated erythrocyte measurement mode, blood of 18 μl and Stromatolyzer NR hemolytic agent (manufactured by Sysmex Corporation) of 882 μl are introduced in the reaction chamber 48. Then, Stromatolyzer NR fluorescent stain solution (manufactured by Sysmex Corporation) of 18 μl is added. The reaction is continued in this state for about 7 seconds to hemolyze erythrocytes and stain leukocytes and nucleated erythrocytes.
The thus treated sample is discharged from the nozzle 6 by the quantifying syringe 44. Among data obtained by the optical measurement, a side fluorescent light intensity (Sf1) and a forward scattered light intensity (Fsc) are used to prepare a two-dimensional frequency distribution map of FIG. 4. In
In the leukocyte/basophil measurement mode, blood of 18 μl and Stromatolyzer FB (II) (manufactured by Sysmex Corporation) of 882 μl are introduced in the reaction chamber 51. The reaction is continued in this state for about 14 seconds, thereby the erythrocytes are hemolyzed and the nuclei of the leukocytes other than the basophils are exposed and shrunk.
The thus treated sample is discharged from the nozzle 6 by the quantifying syringe 44. Among data obtained by the optical measurement, a side scattered light intensity (Ssc) and a forward scattered light intensity (Fsc) are used to prepare a two-dimensional frequency distribution map of FIG. 5. In
In the leukocyte 4-part differential measurement mode, blood of 18 μl and Stromatolyzer 4DL hemolytic agent (manufactured by Sysmex Corporation) of 882 μl are introduced in the reaction chamber 52. Then, Stromatolyzer 4DS fluorescent stain solution (manufactured by Sysmex Corporation) of 18 μl is added. The reaction is continued in this state for about 22 seconds to hemolyze the erythrocytes and stain the leukocytes.
The thus treated blood sample is discharged from the nozzle 6 by the quantifying syringe 44. Among data obtained by the optical measurement, a side scattered light intensity (Ssc) and a side fluorescent light intensity (Sf1) are used to prepare a two-dimensional frequency distribution map of FIG. 6. In
In the reticulocyte measurement mode, blood of 4.5 μl and Retsearch (II) diluent (manufactured by Sysmex Corporation) of 895.5 μl are introduced in the reaction chamber 53. Then, Retsearch (II) fluorescent stain solution (manufactured by Sysmex Corporation) of 18 μl is added. The reaction is continued in this state for 31 seconds to stain the reticulocytes and the like.
The thus treated blood sample is discharged from the nozzle 6 by the quantifying syringe 44. Among data obtained by the optical measurement, a side fluorescent light intensity (Sf1) and a forward scattered light intensity (Fsc) are used to prepare a two-dimensional frequency distribution map of FIG. 7. In
Classification Process
Leukocytes can be classified into lymphocytes, monocytes, neutrophils, basophils and eosinophils. In the leukocyte 4-part differential measurement mode, they are scattered on the distribution map as shown in FIG. 6. If the blood sample is an abnormal sample containing immature granulocytes such as immature neutrophils, a two-dimensional frequency distribution map corresponding to
Then, the classifying section 66 (
First, as shown in
Then, as shown in
Subsequently, on the assumption that a cluster including the first and third particle clusters G1 and G3 is normally distributed, variance and covariance thereof are obtained (step S4).
Then, as shown in
Then, as shown in
The method of determining the ellipse Q shown in
Provided that the point P is on the coordinates (P1, P2), the variance in the direction of the major axis of the ellipse Q is VL, the variance in the direction of the minor axis of the ellipse Q is VS and optional coordinates on the distribution map are (X, Y), the ellipse Q satisfies the following equation:
(X−P1)2/VL+(Y−P2)2/VS=C (C is a fixed value)
The left side of the equation defines a ratio between the major axis and the minor axis and an inclination angle of the ellipse Q. A size of the ellipse Q is suitably determined by the fixed value C.
In the above-described classification process, a cluster existing in a position symmetrical about the point P to the first particle cluster G1 is presumed as the third particle cluster G3. Therefore, regardless of the angle of the particle cluster G0 with respect to the line L, the cluster G3 can be always presumed as a cluster equivalent to the cluster G1. As a result, the particle cluster including the clusters G1 and G3 can be assumed to be normally distributed, which allows accurate classification.
Further, the above-described classifying method can also be adopted to classify the mature erythrocytes and the immature erythrocytes shown in FIG. 7.
According to the present invention, two particle clusters which are not clearly distinguished physiologically or morphologically, such as mature granulocytes and immature granulocytes contained in blood, can accurately be classified on a two-dimensional distribution map. Even in the case where the distribution of such particle clusters is variable depending on the sample, for example, the inclination of the clusters easily fluctuates, the classification can be carried out accurately.
Number | Date | Country | Kind |
---|---|---|---|
2001-226382 | Jul 2001 | JP | national |
Number | Name | Date | Kind |
---|---|---|---|
4661913 | Wu et al. | Apr 1987 | A |
5325169 | Nakamoto et al. | Jun 1994 | A |
5690105 | Shibata et al. | Nov 1997 | A |
5721433 | Kosaka | Feb 1998 | A |
5795727 | Bierre et al. | Aug 1998 | A |
6091843 | Horesh et al. | Jul 2000 | A |
6246786 | Nishikiori et al. | Jun 2001 | B1 |
6662117 | Naito | Dec 2003 | B2 |
Number | Date | Country |
---|---|---|
2674705 | Jul 1997 | JP |
Number | Date | Country | |
---|---|---|---|
20030030784 A1 | Feb 2003 | US |