U.S. patent application number 17/691739 was filed with the patent office on 2022-09-15 for analysis method and analyzer.
This patent application is currently assigned to SYSMEX CORPORATION. The applicant listed for this patent is SYSMEX CORPORATION. Invention is credited to Shoichiro Asada, Konobu Kimura, Kenichiro Suzuki, Masamichi Tanaka.
Application Number | 20220291199 17/691739 |
Document ID | / |
Family ID | 1000006253410 |
Filed Date | 2022-09-15 |
United States Patent
Application |
20220291199 |
Kind Code |
A1 |
Asada; Shoichiro ; et
al. |
September 15, 2022 |
ANALYSIS METHOD AND ANALYZER
Abstract
Disclosed is an analysis method for analyzing a specimen
containing cells, the analysis method including: applying light to
a measurement sample prepared from the specimen and detecting light
generated from cells; obtaining, with respect to each of a
plurality of cells contained in the specimen, feature data of the
cell on the basis of the detected light; analyzing the feature data
with use of an artificial intelligence algorithm, thereby
classifying each of the cells into a plurality of cell types; and
displaying information based on a result of the classifying.
Inventors: |
Asada; Shoichiro; (Kobe-shi,
JP) ; Kimura; Konobu; (Kobe-shi, JP) ; Tanaka;
Masamichi; (Kobe-shi, JP) ; Suzuki; Kenichiro;
(Kobe-shi, JP) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
SYSMEX CORPORATION |
Kobe-shi |
|
JP |
|
|
Assignee: |
SYSMEX CORPORATION
Kobe-shi
JP
|
Family ID: |
1000006253410 |
Appl. No.: |
17/691739 |
Filed: |
March 10, 2022 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G01N 15/1429 20130101;
G01N 33/4915 20130101; G01N 15/1404 20130101; G01N 15/1434
20130101 |
International
Class: |
G01N 33/49 20060101
G01N033/49; G01N 15/14 20060101 G01N015/14 |
Foreign Application Data
Date |
Code |
Application Number |
Mar 12, 2021 |
JP |
2021-040829 |
Mar 12, 2021 |
JP |
2021-040830 |
Mar 12, 2021 |
JP |
2021-040831 |
Claims
1. An analysis method for analyzing a specimen containing cells,
the analysis method comprising: applying light to a measurement
sample prepared from the specimen and detecting light generated
from cells; obtaining, with respect to each of a plurality of cells
contained in the specimen, feature data of the cell on the basis of
the detected light; analyzing the feature data with use of an
artificial intelligence algorithm, thereby classifying each of the
cells into a plurality of cell types; and displaying information
based on a result of the classifying.
2. The analysis method of claim 1, wherein the classifying includes
specifying, on the basis of the feature data, a plurality of cell
types to each of which each of the cells has a possibility of
morphologically belonging, among a plurality of cell types
determined in advance.
3. The analysis method of claim 2, wherein the plurality of cell
types include at least one normal cell and at least one abnormal
cell, and the classifying includes obtaining a probability that
each of the cells belongs to the abnormal cell.
4. The analysis method of claim 1, wherein the plurality of cell
types include lymphocyte, monocyte, eosinophil, neutrophil,
basophil, and abnormal blood cell.
5. The analysis method of claim 3, wherein the abnormal blood cell
includes at least one of immature granulocyte, blast, and abnormal
lymphocyte.
6. The analysis method of claim 1, further comprising counting
cells on the basis of the result of the classifying and displaying
a result of the counting of the cells.
7. The analysis method of claim 1, wherein the analyzing of the
data includes inputting the feature data of one cell into a deep
learning algorithm, and obtaining, as an output from the deep
learning algorithm, a result of classifying the cell into a
plurality of cell types.
8. The analysis method of claim 7, wherein the deep learning
algorithm has been trained using, as teaching data, feature data of
cells and information regarding types of the cells.
9. The analysis method of claim 1, wherein the detecting includes
detecting light generated as a result of a cell passing through a
flow cell to which light is applied, and the obtaining of the
feature data includes obtaining a waveform signal that changes over
time in accordance with the detected light.
10. The analysis method of claim 9, wherein the feature data is
waveform data generated by sampling the waveform signal at a
plurality of time points.
11. The analysis method of claim 1, wherein the analyzing of the
data is performed by using a processor and a parallel-processing
processor that operates under an order of the processor.
12. The analysis method of claim 1, further comprising: displaying
a graph on which a plurality of cells contained in the specimen are
plotted on the basis of the feature data; and displaying, in
accordance with one or a plurality of cells having been selected on
the graph, a result of classifying the selected cell.
13. The analysis method of claim 1, wherein the analyzing of the
data includes obtaining a probability that one cell corresponds to
each of a plurality of the cell types.
14. The analysis method of claim 13, wherein the displaying of the
result of the classifying includes displaying the probability.
15. The analysis method of claim 14, wherein the displaying of the
result of the classifying includes displaying a cell type that has
a highest probability.
16. The analysis method of claim 15, wherein displaying of
information based on the result of the classifying includes
displaying statistic information generated on the basis of the
probability.
17. The analysis method of claim 1, further comprising: displaying
a graph on which a plurality of cells contained in the specimen are
plotted on the basis of the feature data; and displaying, in
accordance with one or a plurality of cells having been selected on
the graph, a result of classifying the selected one or plurality of
cells.
18. The analysis method of claim 1, further comprising: displaying
a graph on which a plurality of cells contained in the specimen are
plotted on the basis of the feature data; and displaying, in
accordance with a plurality of cells having been selected on the
graph, statistic information based on a result of classifying the
selected plurality of cells.
19. The analysis method of claim 1, further comprising: displaying
a graph on which a plurality of cells contained in the specimen are
plotted on the basis of the feature data; receiving an input of an
extraction condition of a cell; extracting a cell that satisfies
the extraction condition on the basis of classification
information; and displaying, among the cells displayed on the
graph, the extracted cell so as to be distinguished from other
cells.
20. An analyzer configured to analyze a specimen containing cells,
the analyzer comprising: a sample preparation part configured to
prepare a measurement sample from the specimen; a detection part
configured to apply light to the measurement sample prepared by the
sample preparation part, and to detect light generated from cells;
a signal processing part configured to obtain, with respect to each
of a plurality of cells contained in the specimen, feature data of
the cell on the basis of the detected light; a display part; and at
least one processor configured to analyze the feature data with use
of an artificial intelligence algorithm, thereby classifying each
of the cells into a plurality of cell types, the at least one
processor being configured to cause the display part to display
information based on a result of the classifying.
Description
RELATED APPLICATIONS
[0001] This application claims priority to Japanese Patent
Application No. 2021-040831, filed on Mar. 12, 2021, No.
2021-040830, filed on Mar. 12, 2021, and No. 2021-040829, filed on
Mar. 12, 2021, the entire contents of which are incorporated herein
by reference.
BACKGROUND OF THE INVENTION
1. Field of the Invention
[0002] The present invention relates to an analysis method and an
analyzer for analyzing cells in a specimen.
2. Description of the Related Art
[0003] Japanese Laid-open Patent Publication No. 2016-514267
(translation of PCT International Application) describes a method
for classifying white blood cells into subpopulations by using
parameters based on fluorescence and light scattering obtained by
causing blood cells to flow in a flow cell. In this method, white
blood cell subpopulations are classified on the basis of a
plurality of parameters including axial light loss (ALL),
intermediate angle scatters (IAS), fluorescence (FL1), and
depolarized side scatter (PSS), which are generated as a result of
applying light to each blood cell flowing in the flow cell. The
classified white blood cell subpopulations are displayed on a
scattergram.
SUMMARY OF THE INVENTION
[0004] The scope of the present invention is defined solely by the
appended claims, and is not affected to any degree by the
statements within this summary.
[0005] In such a conventional cell classification method as
disclosed in Japanese Laid-open Patent Publication No. 2016-514267
(translation of PCT International Application), one cell is
classified into one classification in a uniform manner. However,
for example, neutrophils that are the same in terms of type may
have different morphological features due to the difference in
maturity, and thus may include some neutrophils that have
morphological features similar to those of immature granulocytes.
In the conventional cell classification method, even when a cell
that has a feature of a normal cell such as a neutrophil and a
feature of an abnormal cell such as an immature granulocyte is
included, such a cell is classified into one type in a uniform
manner, which does not lead to more detailed analysis in some
cases.
[0006] An analysis method of the present invention is for analyzing
a specimen containing cells. The analysis method includes: applying
light to a measurement sample prepared from the specimen and
detecting light generated from cells; obtaining, with respect to
each of a plurality of cells contained in the specimen, feature
data of the cell on the basis of the detected light; analyzing the
feature data with use of an artificial intelligence algorithm,
thereby classifying each of the cells into a plurality of cell
types; and displaying information based on a result of the
classifying.
[0007] An analyzer of the present invention is configured to
analyze a specimen containing cells. The analyzer includes: a
sample preparation part configured to prepare a measurement sample
from the specimen; a detection part configured to apply light to
the measurement sample prepared by the sample preparation part, and
to detect light generated from cells; a signal processing part
configured to obtain, with respect to each of a plurality of cells
contained in the specimen, feature data of the cell on the basis of
the detected light; a display part; and at least one processor
configured to analyze the feature data with use of an artificial
intelligence algorithm, thereby classifying each of the cells into
a plurality of cell types, the at least one processor being
configured to cause the display part to display information based
on a result of the classifying.
[0008] According to the present invention, more detailed analysis
can be performed for a cell that has features of a plurality of
cell types.
BRIEF DESCRIPTION OF THE DRAWINGS
[0009] FIG. 1 shows a cell analyzer and a host computer;
[0010] FIG. 2 is a block diagram of a measurement unit;
[0011] FIG. 3 is a block diagram of a sample preparation part;
[0012] FIG. 4 is a schematic diagram of an FCM detection part;
[0013] FIG. 5 illustrates a process of generating a digital signal
performed by an A/D converter;
[0014] FIG. 6 illustrates a structure of feature parameter
data;
[0015] FIG. 7 is a schematic diagram describing a process of cell
classification based on a deep learning algorithm;
[0016] FIG. 8 shows cell types to be analyzed by the deep learning
algorithm;
[0017] FIG. 9 is a block diagram of a processing unit;
[0018] FIG. 10 is a flow chart describing operation of the cell
analyzer;
[0019] FIG. 11 is a flow chart describing a test result data
generation process;
[0020] FIG. 12 schematically shows a first example of a data
structure of test result data;
[0021] FIG. 13 schematically shows a second example of the data
structure of test result data;
[0022] FIG. 14 schematically shows a third example of the data
structure of test result data;
[0023] FIG. 15 schematically shows a fourth example of the data
structure of test result data;
[0024] FIG. 16 schematically shows a fifth example of the data
structure of test result data;
[0025] FIG. 17 is a flow chart of a first example of a test result
displaying process;
[0026] FIG. 18 shows an example of a specimen list screen;
[0027] FIG. 19 shows an example of a detail screen;
[0028] FIG. 20 shows an example of an analysis screen;
[0029] FIG. 21 shows an example of the analysis screen;
[0030] FIG. 22 shows an example of the analysis screen;
[0031] FIG. 23 shows an example of the analysis screen;
[0032] FIG. 24 is a flow chart of a second example of the test
result displaying process;
[0033] FIG. 25 shows an example of the analysis screen;
[0034] FIG. 26 shows an example of the analysis screen;
[0035] FIG. 27 shows an example of the analysis screen;
[0036] FIG. 28 shows an example of the analysis screen;
[0037] FIG. 29 is a flow chart of a third example of the test
result displaying process;
[0038] FIG. 30 is a flow chart of a fourth example of the test
result displaying process;
[0039] FIG. 31 shows an example of the analysis screen;
[0040] FIG. 32 shows an example of the analysis screen;
[0041] FIG. 33 shows an example of the analysis screen;
[0042] FIG. 34 shows an example of the analysis screen;
[0043] FIG. 35 shows an example of the analysis screen;
[0044] FIG. 36 shows an example of the analysis screen;
[0045] FIG. 37 shows an example of the analysis screen;
[0046] FIG. 38 shows an example of the analysis screen;
[0047] FIG. 39A shows an example of a scattergram;
[0048] FIG. 39B shows another example of the scattergram;
[0049] FIG. 40 is a flow chart of a fifth example of the test
result displaying process;
[0050] FIG. 41 shows an example of the analysis screen;
[0051] FIG. 42 shows an example of the analysis screen;
[0052] FIG. 43 is a flow chart of a transmission process to a host
computer;
[0053] FIG. 44A shows an example of a structure of output data;
[0054] FIG. 44B shows an example of the structure of output
data;
[0055] FIG. 45A shows an example of the structure of output
data;
[0056] FIG. 45B shows an example of the structure of output
data;
[0057] FIG. 46A shows an example of a screen for receiving settings
regarding output data;
[0058] FIG. 46B shows an example of the screen for receiving
settings regarding output data;
[0059] FIG. 47 shows a configuration example of a
parallel-processing processor;
[0060] FIG. 48 shows an installation example of the
parallel-processing processor to the measurement unit;
[0061] FIG. 49 shows an installation example of the
parallel-processing processor to the measurement unit;
[0062] FIG. 50 shows an installation example of the
parallel-processing processor to the measurement unit;
[0063] FIG. 51 shows an installation example of the
parallel-processing processor to the measurement unit;
[0064] FIG. 52 shows an outline of arithmetic processes executed by
a processor and the parallel-processing processor;
[0065] FIG. 53A shows an outline of a matrix operation executed by
the parallel-processing processor;
[0066] FIG. 53B shows an outline of a matrix operation executed by
the parallel-processing processor;
[0067] FIG. 54 shows that a plurality of arithmetic processes are
executed by the parallel-processing processor;
[0068] FIG. 55A shows an outline of an arithmetic process regarding
a convolution layer;
[0069] FIG. 55B shows an outline of an arithmetic process regarding
a convolution layer;
[0070] FIG. 56 is a flow chart of cell classification using a deep
learning algorithm performed by the processor and the
parallel-processing processor;
[0071] FIG. 57 is a flow chart of parallel processing
execution;
[0072] FIG. 58 is a block diagram showing another configuration
example of the cell analyzer;
[0073] FIG. 59 is a block diagram showing another configuration
example of the cell analyzer;
[0074] FIG. 60 is a schematic diagram showing control of the
parallel-processing processor by the processor in another
configuration example of the cell analyzer;
[0075] FIG. 61 is a block diagram showing another configuration
example of the cell analyzer;
[0076] FIG. 62 is a block diagram showing another configuration
example of the cell analyzer;
[0077] FIG. 63 is a block diagram showing another configuration
example of the cell analyzer;
[0078] FIG. 64 is a block diagram showing another configuration
example of the cell analyzer;
[0079] FIG. 65 is a block diagram showing another configuration
example of the cell analyzer;
[0080] FIG. 66 is a schematic diagram showing control of the
parallel-processing processor by the processor in another
configuration example of the cell analyzer;
[0081] FIG. 67 is a block diagram showing another configuration
example of the cell analyzer;
[0082] FIG. 68 is a block diagram showing another configuration
example of the cell analyzer;
[0083] FIG. 69 is a block diagram showing another configuration
example of the cell analyzer;
[0084] FIG. 70A illustrates a structure of a neural network;
[0085] FIG. 70B illustrates a structure of the neural network;
[0086] FIG. 70C illustrates a structure of the neural network;
and
[0087] FIG. 71 illustrates training of the neural network.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0088] Hereinafter, the outline and an embodiment of the present
invention will be described in detail with reference to the
accompanying drawings. In the following description and drawings,
the same reference character denotes the same or similar
components, and description regarding the same or similar
components is omitted.
[0089] The analysis method described below is a method in which:
light is applied to a measurement sample prepared from a specimen
containing cells, and light generated from cells is detected; with
respect to each of a plurality of cells contained in the specimen,
feature data of the cell is obtained on the basis of the detected
light; and the feature data is analyzed by using an artificial
intelligence algorithm, whereby each of the cells is classified
into a plurality of cell types.
[0090] The specimen may be a biological sample collected from a
subject. For example, the biological sample can include peripheral
blood such as venous blood or arterial blood, urine, and body fluid
other than blood and urine. The body fluid other than blood and
urine can include bone marrow fluid, ascites, pleural fluid,
cerebrospinal fluid, and the like. Hereinafter, body fluid other
than blood and urine may be simply referred to as "body fluid". A
blood sample is a sample of which the number of cells can be
counted and the cell types of the cells can be determined, i.e., a
sample in a state of containing cells. Preferably, the blood sample
is whole blood. Blood is preferably peripheral blood. An example of
blood includes peripheral blood collected by using an anticoagulant
agent such as ethylenediaminetetraacetate (sodium salt or potassium
salt), heparin sodium, or the like. Peripheral blood may be
collected from an artery or may be collected from a vein.
[0091] Cell types that are classified according to this analysis
method are based on cell types according to morphological
classification, and are different in accordance with the type of
the biological sample. When the biological sample is blood and when
the blood has been collected from a healthy individual, cell types
to be determined in the present embodiment include red blood cell,
nucleated cell such as white blood cell, platelet, and the like,
for example. Nucleated cell includes neutrophil, lymphocyte,
monocyte, eosinophil, and basophil, for example. Neutrophil
includes segmented neutrophil and band neutrophil, for example.
When blood has been collected from a non-healthy individual,
nucleated cell includes abnormal cell.
[0092] Abnormal cells mean cells that are not usually observed in
peripheral blood of a healthy individual. Abnormal cells may
include lymphocytic abnormal cells. Lymphocytic abnormal cells may
include atypical lymphocytes (reactive lymphocytes), abnormal
lymphocytes including mature lymphoma, and plasma cells, for
example. Abnormal cells may include blasts. Blasts include
myeloblasts, lymphoblasts, proerythroblasts, basophilic
erythroblasts, polychromatic erythroblasts, orthochromatic
erythroblasts, promegaloblasts, basophilic megaloblasts,
polychromatic megaloblasts, orthochromatic megaloblasts, and the
like. Abnormal cells may include megakaryocytes. Abnormal cells may
include immature granulocytes. Immature granulocytes include
promyelocytes, myelocytes, metamyelocyte, and the like, for
example.
[0093] Abnormal cells may include other abnormal cells that are not
contained in peripheral blood of a healthy individual. Examples of
abnormal cells are cells that appear when a person has a disease,
and are tumor cells, for example. In the case of the hematopoietic
system, examples of diseases are: myelodysplastic syndrome;
leukemia such as acute myeloblastic leukemia, acute promyelocytic
leukemia, acute myelomonocytic leukemia, acute monocytic leukemia,
erythroleukemia, acute megakaryoblastic leukemia, acute myeloid
leukemia, acute lymphoblastic leukemia, lymphoblastic leukemia,
chronic myelogenous leukemia, chronic lymphocytic leukemia, or the
like; malignant lymphoma such as Hodgkin's lymphoma, non-Hodgkin's
lymphoma, or the like; or multiple myeloma.
[0094] When the biological sample is urine, cell types include, for
example, red blood cell, white blood cell, epithelial cell of
transitional epithelium, squamous epithelium, etc., and the like.
Abnormal cells can include, for example, bacteria, fungi such as
filamentous fungi and yeast, tumor cells, and the like.
[0095] When the biological sample is body fluid that usually does
not contain blood components, such as ascites, pleural fluid, or
spinal fluid, cell types can include red blood cell, white blood
cell, and large cell, for example. The "large cell" here means a
cell that is separated from an inner membrane of a body cavity or a
peritoneum of a viscus, and that is larger than white blood cells.
For example, mesothelial cells, histiocytes, tumor cells, and the
like correspond to the "large cell".
[0096] When the biological sample is bone marrow fluid, cell types
can include, as normal cells, mature blood cell and immature
hematopoietic cell. The mature blood cells include, for example,
red blood cells, nucleated cells such as white blood cells,
platelets, and the like. Nucleated cells such as white blood cells
include, for example, neutrophils, lymphocytes, plasma cells,
monocytes, eosinophils, and basophils. Neutrophils include
segmented neutrophils and band neutrophils, for example. Immature
hematopoietic cells include, for example, hematopoietic stem cells,
immature granulocytic cells, immature lymphoid cells, immature
monocytic cells, immature erythroid cells, megakaryocytic cells,
mesenchymal cells, and the like. Immature granulocytes can include,
for example, cells such as metamyelocytes, myelocytes,
promyelocytes, and myeloblasts. Immature lymphoid cells include,
for example, lymphoblasts and the like. Immature monocytic cells
include monoblasts and the like. Immature erythroid cells include,
for example, nucleated erythrocytes such as proerythroblasts,
basophilic erythroblasts, polychromatic erythroblasts,
orthochromatic erythroblasts, promegaloblasts, basophilic
megaloblasts, polychromatic megaloblasts, and orthochromatic
megaloblasts. Megakaryocytic cells include, for example,
megakaryoblasts and the like.
[0097] Examples of abnormal cells that can be contained in bone
marrow include hematopoietic tumor cells of: myelodysplastic
syndrome; leukemia such as acute myeloblastic leukemia, acute
promyelocytic leukemia, acute myelomonocytic leukemia, acute
monocytic leukemia, erythroleukemia, acute megakaryoblastic
leukemia, acute myeloid leukemia, acute lymphoblastic leukemia,
lymphoblastic leukemia, chronic myelogenous leukemia, or chronic
lymphocytic leukemia; malignant lymphoma such as Hodgkin's lymphoma
or non-Hodgkin's lymphoma; or multiple myeloma, which have been
described above, and metastasized tumor cells of a malignant tumor
developed in an organ other than bone marrow.
[0098] This analysis method is suitably performed by using a cell
analyzer which analyzes a specimen containing cells. The cell
analyzer may include: a sample preparation part which prepares a
measurement sample from the specimen; a detection part which
applies light to the measurement sample prepared by the sample
preparation part and detects light generated from cells; a signal
processing part which obtains, with respect to each of a plurality
of cells contained in the specimen, feature data of the cell on the
basis of the detected light; and a controller which analyzes the
feature data with use of an artificial intelligence algorithm,
thereby classifying each of the cells into a plurality of cell
types. An example of such a cell analyzer will be described
below.
1. BASIC CONFIGURATION
[0099] With reference to FIG. 1, a basic configuration of a cell
analyzer is described. FIG. 1 is a schematic diagram showing an
appearance of a cell analyzer 100. The cell analyzer 100 is an
apparatus that analyzes a specimen derived from an organism in
accordance with a test order transmitted from a host computer 500.
The cell analyzer 100 includes a measurement unit 400 and a
processing unit 300. The host computer 500 and the cell analyzer
100 are collectively referred to as a test system 1000.
2. CONFIGURATION OF MEASUREMENT UNIT
[0100] With reference to FIG. 2, a configuration of the measurement
unit 400 is described. FIG. 2 shows an example of a block diagram
of the measurement unit 400. As shown in FIG. 2, the measurement
unit 400 includes: a specimen suction part 450 which suctions a
specimen; a sample preparation part 440 which prepares a
measurement sample from the suctioned specimen; an FCM detection
part 410 which detects each blood cell in the measurement sample;
an analog processing part 420 which processes an analog signal
outputted from the FCM detection part 410; a measurement unit
controller 480 which converts the signal processed by the analog
processing part 420 into a digital signal and analyzes the digital
signal; and an apparatus mechanism part 430.
[0101] FIG. 3 is a schematic diagram for describing the specimen
suction part 450 and the sample preparation part 440. The specimen
suction part 450 includes: a nozzle 451 for suctioning a blood
specimen from a blood collection tube T; and a pump 452 for
providing a negative pressure/positive pressure to the nozzle 451.
The nozzle 451 is moved in upwardly and downwardly by the apparatus
mechanism part 430, to be inserted into the blood collection tube
T. When the pump 452 provides a negative pressure in a state where
the nozzle 451 is inserted in the blood collection tube T, the
blood specimen is suctioned via the nozzle 451.
[0102] The sample preparation part 440 includes five reaction
chambers 440a to 440e. The reaction chambers 440a to 440e are used
in measurement channels of DIFF, RET, WPC, PLT-F, and WNR,
respectively. Each reaction chamber has connected thereto, via flow
paths, a hemolytic agent container containing a hemolytic agent and
a staining liquid container containing a staining liquid, which
serve as reagents for the corresponding measurement channel. One
reaction chamber and reagents (a hemolytic agent and a staining
liquid) connected thereto form a measurement channel. For example,
the DIFF measurement channel is configured by a DIFF hemolytic
agent and a DIFF staining liquid which serve as DIFF measurement
reagents; and the reaction chamber 440a. The other measurement
channels are configured in similar manners. Here, an example of a
configuration in which one measurement channel includes one
hemolytic agent and one staining liquid is shown. However, one
measurement channel may not necessarily include both of a hemolytic
agent and a staining liquid, and a plurality of measurement
channels may share one reagent.
[0103] Through horizontal and up-down movement by the apparatus
mechanism part 430, the nozzle 451 having suctioned a blood
specimen accesses, from above, a reaction chamber, among the
reaction chambers 440a to 440e, that corresponds to a measurement
channel that corresponds to an order, and the nozzle 451 discharges
the suctioned blood specimen. The sample preparation part 440
supplies a corresponding hemolytic agent and a corresponding
staining liquid to the reaction chamber having the blood specimen
discharged therein, to mix the blood specimen, the hemolytic agent,
and the staining liquid in the reaction chamber, thereby preparing
a measurement sample. The prepared measurement sample is supplied
from the reaction chamber to the FCM detection part 410 via a flow
path, to be subjected to measurement of cells according to flow
cytometry.
[0104] FIG. 4 shows a configuration example of an optical system of
the FCM detection part 410. As shown in FIG. 4, in measurement
performed by a flow cytometer, when each cell contained in a
measurement sample passes through a flow cell (sheath flow cell)
4113 provided in the flow cytometer, a light source 4111 applies
light to a flow cell 4113, and scattered light and fluorescence
emitted from the cell in the flow cell 4113 due to this light are
detected by light-receiving elements 4116, 4121, 4122.
[0105] In FIG. 4, light emitted from a laser diode being the light
source 4111 is applied via a light application lens system 4112 to
each cell passing through the flow cell 4113.
[0106] As shown in FIG. 4, forward scattered light emitted from a
particle passing through the flow cell 4113 is received by the
light-receiving element 4116 via a condenser lens 4114 and a pin
hole part 4115. The light-receiving element 4116 is a photodiode,
for example. Side scattered light is received by a light-receiving
element 4121 via a condenser lens 4117, a dichroic mirror 4118, a
bandpass filter 4119, and a pin hole part 4120. The light-receiving
element 4121 is a photomultiplier, for example. Side fluorescence
is received by a light-receiving element 4122 via the condenser
lens 4117 and the dichroic mirror 4118. The light-receiving element
4122 is a photomultiplier, for example. As the light-receiving
element 4116, an avalanche photodiode or a photomultiplier may be
used, instead of a photodiode. As the light-receiving element 4121
and the light-receiving element 4122, a photodiode or an avalanche
photodiode may be used.
[0107] Reception light signals outputted from the respective
light-receiving elements 4116, 4121, 4122 are inputted to the
analog processing part 420 via amplifiers 4151, 4152, 4153,
respectively. The analog processing part 420 is connected via
signal transmission paths 421 to an A/D converter 482 of the
measurement unit controller 480 described later.
[0108] With reference back to FIG. 2, the analog processing part
420 performs processes such as noise removal and smoothing onto
each analog signal inputted from the FCM detection part 410, and
outputs the processed analog signal to the measurement unit
controller 480.
[0109] The measurement unit controller 480 includes the A/D
converter 482, a processor 4831, a RAM 4834, a storage 4835, a bus
controller 4850, a parallel-processing processor 4833, a bus 485,
and an interface part 489 connected to the processing unit 300. The
measurement unit controller 480 includes an interface part 484
provided between the bus 485 and the A/D converter 482. The
measurement unit controller 480 further includes an interface part
488 provided between various types of hardware (i.e., the specimen
suction part 450, the apparatus mechanism part 430, the sample
preparation part 440, the FCM detection part 410) and the bus
485.
[0110] The A/D converter 482 samples the analog signal outputted
from the analog processing part 420 at a predetermined sampling
rate (e.g., sampling at 1024 points at a 10 nanosecond interval,
sampling at 128 points at an 80 nanosecond interval, sampling at 64
points at a 160 nanosecond interval, or the like), and converts the
analog signal into a digital signal. The A/D converter 482 converts
the analog signal that is from the measurement start of the
specimen to the measurement end, into a digital signal. When a
plurality of types of analog signals (e.g., analog signals
respectively corresponding to forward scattered light intensity,
side scattered light intensity, and fluorescence intensity) are
generated through measurement in a certain measurement channel, the
A/D converter 482 converts each analog signal that is from the
measurement start to the measurement end, into a digital signal. As
described with reference to FIG. 4, three types of analog signals
(i.e., forward scattered light signal, side scattered light signal,
and fluorescence signal) are inputted via a plurality of
corresponding signal transmission paths 421 to the A/D converter
482. The A/D converter 482 converts each of the analog signals
inputted from the plurality of signal transmission paths 421, into
a digital signal.
[0111] FIG. 5 is a schematic diagram for describing a process of
sampling of an analog signal performed by the A/D converter 482.
When a measurement sample containing a cell C is caused to flow in
the flow cell 4113, and light is applied to the flow cell 4113,
forward scattered light is generated in a forward direction with
respect to the advancing direction of light. Similarly, side
scattered light and side fluorescence are generated to a side
direction with respect to the advancing direction of light. The
forward scattered light is received by the light-receiving element
4116, and a signal corresponding to the amount of the received
light is outputted. The side scattered light is received by the
light-receiving element 4121, and a signal corresponding to the
amount of the received light is outputted. The side fluorescence is
received by the light-receiving element 4122, and a signal
corresponding to the amount of the received light is outputted. In
association with passage through the flow cell 4113 of a plurality
of cells contained in the measurement sample, an analog signal
which represents change in the signal associated with a lapse of
time is outputted from each of the light-receiving elements 4116,
4121, 4122. The analog signal corresponding to forward scattered
light is referred to as a "forward scattered light signal", an
analog signal corresponding to side scattered light is referred to
as a "side scattered light signal", and an analog signal
corresponding to side fluorescence is referred to as a
"fluorescence signal". One pulse of each analog signal corresponds
to one cell.
[0112] The analog signals are inputted to the A/D converter 482.
From a start point, which is the time point when the level of the
forward scattered light signal, among the analog signals inputted
from the light-receiving elements 4116, 4121, 4122, has reached a
level set as a predetermined threshold, the A/D converter 482
samples the forward scattered light signal, the side scattered
light signal, and the fluorescence signal. The A/D converter 482
samples the respective analog signals at a predetermined sampling
rate (e.g., sampling at 1024 points at a 10 nanosecond interval,
sampling at 128 points at an 80 nanosecond interval, sampling at 64
points at a 160 nanosecond interval, or the like). The sampling
time is fixed irrespective of the size of the pulse. The sampling
time is set so as to be longer than the time between the rise and
fall of the level of the analog signal during passage of one cell
through the beam spot of the flow cell 4113. Accordingly, matrix
data that has, as elements, values indicating the analog signal
level at a plurality of time points is obtained as a digital signal
that corresponds to one cell. In this manner, the A/D converter 482
generates a digital signal of forward scattered light, a digital
signal of side scattered light, and a digital signal of side
fluorescence that correspond to one cell. The A/D conversion is
repeated until the number of cells for which the digital signals
have been obtained reaches a predetermined number, or until a
predetermined time has lapsed from the start of causing the
measurement sample to flow in the flow cell 4113. Accordingly, as
shown in FIG. 5, with respect to N cells in the measurement sample,
digital signals obtained by digitizing the waveforms of the analog
signals of each cell is obtained. In the present specification, a
set of sampling data of each cell included in each digital signal
(in the example in FIG. 5, the set of 1024 digital values from t=0
ns to t=10240 ns) is referred to as waveform data.
[0113] Each piece of waveform data generated by the A/D converter
482 is provided with an index for identifying the corresponding
cell. As the indexes, for example, integers of 1 to N are provided
in the order of the generated pieces of waveform data, and the
waveform data of forward scattered light, the waveform data of side
scattered light, and the waveform data of side fluorescence are
each provided with the same index. The index corresponds to a cell
ID included in a cell data structure described later.
[0114] Since one piece of waveform data corresponds to one cell,
the index corresponds to the cell that has been measured. Since an
identical index is provided to the pieces of the waveform data that
correspond to the same cell, a deep learning algorithm described
later can analyze the waveform data of forward scattered light, the
waveform data of side scattered light, and the waveform data of
fluorescence that correspond to an individual cell; and classify
the type of the cell.
[0115] In addition to generation of waveform data that corresponds
to each cell, the A/D converter 482 calculates peak values of
pulses of each signal, and generates feature parameter data. FIG. 6
is a schematic diagram showing an example of feature parameter
data. In parallel with the process of converting each analog signal
of each cell into digital values in order to generate waveform
data, the A/D converter 482 sequentially stores the maximum value
of digital values included in the waveform data of each cell from
the head column of feature parameter data. The column N stores the
value of the cell for which the N-th waveform data has been
generated. The position of the column matches the index provided at
the head of the waveform data. That is, the digital value
corresponding to the waveform data having an index "N" is stored in
the column N. The maximum value of the digital values corresponds
to the peak height of the corresponding pulse of the analog signal
of the cell. Therefore, with respect to each of the forward
scattered light signal, the side scattered light signal, and the
side fluorescence signal, through extraction of the maximum value
of the digital values included in the waveform data of each cell,
the peak value (referred to as FSCP) of the pulse of the forward
scattered light signal, the peak value (referred to as SSCP) of the
pulse of the side scattered light signal, and the peak value
(referred to as SFLP) of the pulse of the side fluorescence signal
can be obtained for each cell.
[0116] With reference back to FIG. 2, the A/D converter 482 inputs
the generated digital signals and feature parameter data to the bus
485. The bus controller 4850 transmits the digital signals
outputted from the A/D converter 482, to the RAM 4834, through DMA
(Direct Memory Access) transfer, for example. The RAM 4834 stores
the digital signals and the feature parameter data.
[0117] The processor 4831 is connected to the interface part 489,
the interface part 488, the interface part 484, the RAM 4834, and
the storage 4835 via the bus 485. The processor 4831 is connected
to the parallel-processing processor 4833 via the bus 485. The
processing unit 300 is connected to components of the measurement
unit 400 via the interface part 489 and the bus 485. The bus 485 is
a transmission line having a data transfer rate of not less than
several hundred MB/s. The bus 485 may be a transmission line having
a data transfer rate of not less than 1 GB/s, for example. The bus
485 performs data transfer in accordance with the PCI-Express or
PCI-X standard, for example.
[0118] Using the parallel-processing processor 4833, the processor
4831 analyzes each digital signal by executing analysis software
4835a stored in the storage 4835.
[0119] The processor 4831 is a CPU (Central Processing Unit), for
example. The parallel-processing processor 4833 executes, in
parallel, a plurality of arithmetic processes which are at least a
part of the process regarding analysis of waveform data. The
parallel-processing processor 4833 is a GPU (Graphics Processing
Unit), an FPGA (Field Programmable Gate Array), or an ASIC
(Application Specific Integrated Circuit), for example. When the
parallel-processing processor 4833 is an FPGA, the
parallel-processing processor 4833 may have programed therein in
advance an arithmetic process regarding a deep learning algorithm
60 realized by the analysis software 4835a, for example. When the
parallel-processing processor 4833 is an ASIC, the
parallel-processing processor 4833 may have incorporated therein in
advance a circuit for executing the arithmetic process regarding
the deep learning algorithm 60, or may have a programmable module
built therein in addition to such an incorporated circuit, for
example. As the parallel-processing processor 4833, GeForce,
Quadro, TITAN, Jetson, or the like of NVIDIA Corporation is
suitably used, for example.
[0120] FIG. 7 illustrates an arithmetic process of the deep
learning algorithm 60 of the analysis software 4835a.
[0121] The deep learning algorithm 60 is implemented as a neural
network that includes a multi-layered middle layer. Preferably, the
neural network is a convolutional neural network (CNN) having a
convolution layer. The number of nodes of an input layer 60a in the
neural network corresponds to the number of sequences included in
the waveform data of one cell to be inputted. In the example in
FIG. 6, the number of nodes of the input layer 60a corresponds to
the number of sequences, i.e., 1024.times.3, of the waveform data
86a, 86b, 86c of one cell.
[0122] An output layer 60b of the neural network includes the
number of nodes that corresponds to the cell types to be analyzed.
In the example in FIG. 8, the cell types to be analyzed are nine
types, i.e., "neutrophil (NEUT)", "lymphocyte (LYMPH)", "monocyte
(MONO)", "eosinophil (EO)", "basophil (BASO)", "immature
granulocyte (IG)", "blast (Blast)", "abnormal lymphocyte (Abn
LYMPH)", and "not applicable (NONE)". The number of nodes of the
output layer 60b in this case is nine.
[0123] When waveform data has been inputted to the input layer 60a
of the neural network forming the deep learning algorithm 60, a
probability that a cell corresponding to the waveform data
corresponds to each cell type is outputted from the output layer
60b, as classification information 82 in which the cell is
classified into a plurality of cell types. The data outputted from
the output layer 60b includes a cell ID, identification information
(the label value in FIG. 8) of the cell type, and a numerical value
of the probability corresponding to each cell type.
[0124] The processor 4831 stores into the RAM 4834 the
classification information 82 of the cell obtained by the deep
learning algorithm 60. The classification information 82 and the
feature parameter data of each cell is transmitted to the
processing unit 300 via the interface part 489.
3. CONFIGURATION OF PROCESSING UNIT 300
[0125] A configuration of the processing unit 300 is described with
reference to FIG. 9. The processing unit 300 is connected to the
processor 4831 of the measurement unit 400 via an interface part
3006 and a bus 3003, and can receive the classification information
82 and the feature parameter data generated by the measurement unit
400. The interface part 3006 is a USB interface, for example.
[0126] The processing unit 300 includes a processor 3001, the bus
3003, a storage 3004, the interface part 3006, a display part 3015,
and an operation part 3016. The processing unit 300 as hardware is
implemented by a general personal computer, and functions as a
processing unit of the cell analyzer 100, by executing a dedicated
program stored in the storage 3004.
[0127] The processor 3001 is a CPU, and can execute a program
stored in the storage 3004.
[0128] The storage 3004 includes a hard disk device. The storage
3004 stores at least a program for processing the classification
information 82 of each cell transmitted from the measurement unit
400 and for generating a test result of the specimen.
[0129] The display part 3015 includes a computer screen. The
display part 3015 is connected to the processor 3001 via the
interface part 3006 and the bus 3003. The display part 3015 can
receive an image signal inputted from the processor 3001 and
display a test result.
[0130] The operation part 3016 includes a pointing device including
a keyboard, a mouse, or a touch panel. A user such as a doctor or a
laboratory technician operates the operation part 3016 to input a
measurement order to the cell analyzer 100, thereby being able to
input a measurement instruction in accordance with the measurement
order. The operation part 3016 can also receive an instruction of
displaying a test result from the user. By operating the operation
part 3016, the user can view various types of information regarding
the test result, such as a graph, a chart, or flag information
provided to the specimen.
4. OPERATION OF CELL ANALYZER 100
[0131] With reference to FIG. 10, a specimen analysis operation
performed by the cell analyzer 100 is described.
[0132] When the processor 3001 of the processing unit 300 has
received a measurement order and a measurement instruction from the
user via the operation part 3016, the processor 3001 transmits a
measurement command to the measurement unit 400 (step S1).
[0133] Upon receiving the measurement command, the processor 4831
of the measurement unit 400 starts measurement of a specimen. The
processor 4831 causes the specimen suction part 450 to suction a
specimen from a blood collection tube T (step S10). Next, the
processor 4831 causes the specimen suction part 450 to dispense the
suctioned specimen into one of the reaction chambers 440a to 440e
of the sample preparation part 440. The measurement command
transmitted from the processing unit 300 in step S1 includes
information of a measurement channel for which measurement is
requested by the measurement order. On the basis of the information
of the measurement channel included in the measurement command, the
processor 4831 controls the specimen suction part 450 so as to
discharge the specimen into the reaction chamber of the
corresponding measurement channel.
[0134] The processor 4831 causes the sample preparation part 440 to
prepare a measurement sample (step S11). In step S11, upon
receiving an order from the processor 4831, the sample preparation
part 440 supplies the reagents (hemolytic agent and staining
liquid) into the reaction chamber having the specimen discharged
therein, to mix the specimen with the reagents. Accordingly, a
measurement sample in which red blood cells are hemolyzed by the
hemolytic agent and in which cells, such as white blood cells or
reticulocytes, serving as the target of the measurement channel are
stained by the staining liquid, is prepared in the reaction
chamber.
[0135] The processor 4831 causes the FCM detection part 410 to
measure the prepared measurement sample (step S12). In step S12,
the processor 4831 controls the apparatus mechanism part 430 to
send the measurement sample in the reaction chamber of the sample
preparation part 440, to the FCM detection part 410. The reaction
chamber and the FCM detection part 410 are connected to each other
by a flow path, and the measurement sample sent from the reaction
chamber flows in the flow cell 4113, and is irradiated with laser
light by the light source 4111 (see FIG. 4).
[0136] As a result of the measurement sample being supplied to the
flow cell 4113, a forward scattered light signal, a side scattered
light signal, and a side fluorescence signal are inputted to the
A/D converter 482. The A/D converter 482 generates digital signals
each being a set of pieces of waveform data that correspond to
individual cells, and feature parameter data storing the maximum
value included in the waveform data of each cell. The methods for
generating the waveform data and the feature parameters have been
described above.
[0137] The processor 4831 controls the bus controller 4850 to cause
the waveform data and the feature parameter data generated by the
A/D converter 482 to be taken into the RAM 4834 through DMA
transfer. Through the DMA transfer, the waveform data is directly
transferred to the RAM 4834, not via the processor 4831. The
waveform data is stored into the RAM 4834.
[0138] By using the deep learning algorithm 60, the processor 4831
executes cell classification on the basis of the generated waveform
data, and generates classification information (step S13).
[0139] The processor 4831 transmits, to the processing unit 300,
the classification information 82 and the feature parameter data of
each individual cell which have been obtained as a result of step
S13 (step S14).
[0140] Upon receiving the classification information 82 and the
feature parameter data from the measurement unit (step S2), the
processor 3001 of the processing unit 300 analyzes the
classification information 82 by using a program stored in the
storage 3004, and generates test result data of the specimen (step
S3). The test result data is stored into the storage 3004. The
process of step S3 will be described later.
[0141] The processor 3001 displays the test result on the display
part 3015 (step S4). The process of step S4 will be described
later.
[0142] The processor 3001 transmits the test result to the host
computer 500 (step S5). The process of step S5 will be described
later.
[0143] The host computer 500 receives the test result transmitted
from the processor 3001 of the processing unit 300 (step S21).
Accordingly, a series of processes ends.
5. TEST RESULT DATA GENERATION PROCESS
[0144] FIG. 11 is a flow chart showing details of step S3 (test
result data generation process) executed by the processor 3001 of
the processing unit 300.
[0145] On the basis of the classification information 82 received
from the measurement unit 400, the processor 3001 determines a main
cell type of the cell (step S31). Specifically, on the basis of the
probability for each cell type included in the classification
information 82, the processor 3001 determines, as the main cell
type, a cell type that has the highest probability. For example, in
a case where the probability data obtained with respect to a
certain cell is as indicated below, the cell type that has the
highest probability is neutrophil, and thus, the main cell type of
the cell is neutrophil.
TABLE-US-00001 TABLE 1 Cell type Probability neutrophil 90%
lymphocyte 10% monocyte 0% eosinophil 0% basophil 0% immature
granulocyte 0% blast 0% abnormal lymphocyte 0%
[0146] Next, on the basis of the determined main cell type, the
processor 4831 counts the number of cells contained in the
specimen, for each cell type (step S32). In the process of step
S32, on the basis of the information of cell types of each
individual cell, the number of cells is counted for each cell type.
For example, when WDF is set as the measurement channel, the number
of cell types to be classified is nine as shown as an example in
FIG. 8. For example, when the number of cells of which the main
cell type is neutrophil is M, the processor 3001 generates a
counting result in which the number of neutrophils is M. The
processor 3001 performs a similar process on lymphocyte, monocyte,
eosinophil, basophil, immature granulocyte, blast, and abnormal
lymphocyte, and generates a counting result for each cell type.
[0147] The processor 3001 may further determine whether the
specimen is normal/abnormal on the basis of the count value of each
cell type. For example, since blasts are not contained in
peripheral blood of a healthy individual, when not less than a
predetermined number of cells of which the main cell type has been
determined as blast are contained, the specimen is suspected to
have some abnormality. In addition, since immature granulocytes are
not usually contained in blood of a healthy individual, when not
less than a predetermined number of cells of which the main cell
type has been determined as immature granulocyte are contained, the
specimen is suspected to have some abnormality. The processor 3001
compares the count value (absolute value) of each cell type with a
predetermined numerical range or threshold, and when the count
value is outside the numerical range or exceeds the threshold, the
processor 3001 may add a flag indicating that abnormality is
suspected, to the analysis result.
[0148] The processor 3001 may further obtain the content ratio of
cells on the basis of the count value of each cell type. For
example, when WDF is set as the measurement channel, the processor
3001 obtains the ratio of five white blood cell subclasses of
neutrophil, lymphocyte, monocyte, eosinophil, and basophil. Blood
of a healthy individual contains these five white blood cell
subclasses at predetermined proportions. The processor 3001 may
determine whether the specimen is normal/abnormal by comparing
these proportions of white blood cell subclasses with predetermined
numerical ranges or thresholds. When a specific white blood cell
ratio is outside a predetermined numerical range or exceeds a
threshold, the processor 3001 may add a flag suggesting that the
specimen is abnormal, to the analysis result.
[0149] The processor 3001 generates the above-described test result
data in association with the cell ID (step S33). FIG. 12
schematically shows a first example of a data structure of test
result data generated by the processor 3001.
[0150] The test result data is configured as a relational database,
and includes a plurality of data items (columns). In the example
shown in FIG. 12, subject ID, specimen ID, measurement result,
identification code of measurement channel, cell ID, identification
information of cell type, and probability for each cell type are
provided as data items.
[0151] The data of "subject ID" stores a subject ID being the
information for identifying the subject from whom the specimen has
been collected. The subject ID is a multiple-character string
composed of a combination of numerals and alphabets, for
example.
[0152] The data of "specimen ID" stores a specimen ID being the
information for identifying the specimen. The specimen ID is a
multiple-character string composed of a combination of numerals and
alphabets, for example.
[0153] The data of "measurement result" stores a measurement result
obtained by the processor 3001 analyzing measurement data on the
basis of the main cell type in step S32. The measurement result
includes a counting result of cells based on the cell type (main
cell type) that has the highest probability, as described above.
When the measurement channel is WDF, the measurement result
includes the number and proportion of each of the white blood cell
subclasses (monocyte, neutrophil, lymphocyte, eosinophil,
basophil). When the measurement channel is RET, the measurement
result includes the number of reticulocytes. When the measurement
channel is WPC, the measurement result includes the number of
hematopoietic progenitor cells. When the measurement channel is
WNR, the measurement result includes the number of white blood
cells and the number of nucleated erythrocytes.
[0154] The data of "measurement channel" stores information that
indicates the measurement channel of the measurement having been
performed on the specimen. In the example of the measurement unit
400 of the present embodiment, DIFF, RET, WPC, PLT-F, and WNR exist
as the measurement channels. In the data items of the measurement
channel, a character string of "DIFF", "RET", "WPC", "PLT-F", or
WNR" is stored in accordance with the measurement channel in which
the measurement has been performed. The information indicating the
measurement channel may be a character string indicating the
measurement channel, or may be a numeral or alphabet assigned to
each measurement channel. For example, numerals may be assigned
such that DIFF=1, RET=2, WPC=3, PLT-F=4, and WNR=5, and the numeral
that corresponds to the used measurement channel may be stored. The
data of "measurement channel" stores one or a plurality of values
with respect to one specimen, in accordance with the number of
measurement channels in which measurement has been performed.
[0155] As for the data of "cell ID", a plurality of pieces of data
are created in accordance with the number of cells detected in the
measurement in one measurement channel. The cell ID is a numerical
value of a natural number such as 1, 2, 3, . . . , or N. The cell
ID matches the index (see FIG. 5) added to the head of waveform
data and to the sequential order (see FIG. 6) of the column of the
feature parameter data that correspond to each individual cell
described above.
[0156] The data of "feature parameter" stores the values of the
forward scattered light peak value (FSCP), the side scattered light
peak value (SSCP), and the side fluorescence peak value (SFLP) of
the corresponding cell. In the example in FIG. 12, as the data of
"feature parameter", three feature parameters of FSCP, SSCP, and
SFLP are created for one cell ID. As described with reference to
FIG. 6, in the feature parameter data, the sequential orders of
columns correspond to cell IDs. For example, the head column (first
column) of the feature parameter data stores a feature parameter
that corresponds to the cell having a cell ID "1". Therefore, the
data of FSCP corresponding to cell ID=1 stores a value "59" in the
head column of the FSCP matrix data in the feature parameter data.
Similarly, data of SSCP and data of SFLP store a value "30" in the
head column of the SSCP matrix data, and a value "134" in the head
column of the SFLP matrix data, respectively.
[0157] The data of "cell type" stores information indicating a cell
type to be analyzed. The information indicating a cell type is, for
example, a label value that can specify a cell type as shown in
FIG. 8. As for the data of "cell type", a plurality of pieces of
data are created in accordance with the number of cell types to be
analyzed in the measurement channel. For example, when the number
of cell types to be analyzed in a measurement channel is nine
including "none" as shown in FIG. 8, nine pieces of data of "cell
type" are created. The identification information may be a
character string such as "neutrophil" or "NEUT", instead of a label
value.
[0158] As for the data of "probability", one piece of data is
created for data of one cell type. With respect to the
corresponding cell ID and identification information, the data of
"probability" stores, in the form of numerical data, a value of a
probability outputted by the deep learning algorithm 60. For
example, when the classification information 82 of a target cell is
as shown in Table 1 above, "90%" is stored as a variable
corresponding to neutrophil, "10%" is stored as a variable
corresponding to lymphocyte, and "0%" is stored as variables
corresponding to the other pieces of identification
information.
[0159] The processor 3001 generates test result data shown in FIG.
12, and stores the test result data into the storage 3004.
[0160] FIG. 13 schematically shows a second example of test result
data. In the data structure of test result data shown in FIG. 13,
when compared with that in FIG. 12, data storing a main flag is
added to each piece of identification information. In the case of
the data structure in FIG. 13, the classification information is
composed of identification information, probability, and main
flag.
[0161] The data of "main flag" stores a value of 0 or 1. The
processor 3001 stores 1 into the main flag that corresponds to the
cell type (main cell type) that has the highest probability, and
stores 0 into the main flags that correspond to the other cell
types. With this configuration, the processor 3001 need not specify
a main cell type for each individual cell on the basis of the
probabilities every time using in an arithmetic operation or
outputting the information of the main cell type. Accordingly,
speed up of processing can be realized. For example, in a
scattergram described later, a plot of each cell is displayed in a
color that is different for each cell type. In such a case, the
processor 3001 can, with reference to the main flag, determine the
color of the plot of each cell, and need not specify the main cell
type on the basis of the probabilities of each individual cell. In
the case of the second example, when the deep learning algorithm 60
generates classification information 82, the deep learning
algorithm 60 may execute a process of setting a numerical value in
each main flag together with the classification information 82.
[0162] FIG. 14 schematically shows a third example of test result
data. In the test result data shown in FIG. 14, when compared with
that in FIG. 12, one piece of data of "main cell type" is added to
one cell ID. In the case of the data structure in FIG. 14, the
classification information is composed of identification
information, probability, and identification information indicating
the main cell type.
[0163] The data of "main cell type" stores identification
information of the cell type (main cell type) that has the highest
probability. In the second example in FIG. 13, in order to specify
the main cell type of each individual cell, it is necessary to
search the columns storing the main flags. However, in the third
example in FIG. 14, with reference to the data of "main cell type",
the main cell type can be specified. Accordingly, further speed up
of processing can be realized.
[0164] FIG. 15 schematically shows a fourth example of test result
data. In the test result data shown in FIG. 15, when compared with
that in FIG. 12, data of "research flag" is added so as to
correspond to each piece of identification information. In the case
of the data structure in FIG. 15, the classification information is
composed of identification information, probability, and research
flag.
[0165] The data of "research flag" stores a value of 0 or 1. The
research flag is a flag for distinguishing the cell type to be
displayed as a test result of a reportable item, from the cell type
to be displayed in an auxiliary manner as a research item, on a
result display screen described later. The processor 3001 stores
"1" into a research flag that corresponds to a cell type having a
probability not higher than a predetermined threshold, and stores
"0" into a research flag that corresponds to a cell type having a
probability higher than this threshold. When 1 has been set in the
research flag, in a case where test result data is to be displayed,
it is possible to provide the probability, the counting result, or
the like regarding this cell type, with information indicating that
these pieces of information are preferably used for research.
[0166] FIG. 16 schematically shows a fifth example of test result
data. In the test result data shown in FIG. 16, when compared with
that in FIG. 12, data of "rank" is added to each piece of
identification information. In the case of the data structure in
FIG. 16, the classification information is composed of
identification information, probability, and rank.
[0167] The data of "rank" stores one of numerals of 1 to m (m: the
number of cell types to be analyzed), which are provided in a
descending order of probabilities to all of the cell types
associated with a cell ID. The processor 4831 calculates ranks in a
descending order of probabilities, and stores each calculated rank
into the data of "rank" of the corresponding cell type. In this
case, with reference to the data of "rank", the processor 3001 can
smoothly understand not only the cell type that has the highest
probability, but also the cell type that has the second highest
probability, for example.
6. TEST RESULT DISPLAYING PROCESS
[0168] FIG. 17 is a flow chart illustrating a first example of a
test result displaying process (step S4) executed by the processor
3001.
[0169] The processor 3001 displays a specimen list screen 700
(sample explorer screen) (step S410). The processor 3001 displays a
detail screen 800 (browser screen) in accordance with an operation
by the user (step S411). In accordance with a display instruction
from the user, the processor 3001 displays an analysis screen 900
including a scattergram (step S412). The processor 3001 reads out
classification information of a cell that corresponds to a dot
selected in a scattergram 901 displayed on the analysis screen 900,
and displays the read out classification information (step
S413).
[0170] FIG. 18 shows an example of the specimen list screen 700.
The specimen list screen 700 includes a tool bar 710 and a data
display region 720. A plurality of icons for performing
predetermined operations on the screen are displayed in the tool
bar 710. Specifically, the tool bar 710 includes: a specimen list
screen icon 710a for calling the specimen list screen 700 (SE
screen); a detail icon 710b for calling the detail screen 800; a
measurement registration icon 710c for calling a measurement order
input screen for inputting a measurement order; and a validation
icon 710d for validating a test result.
[0171] The data display region 720 in which data of test results is
displayed is provided in a part below the tool bar 710. The tool
bar 710 is always displayed in an upper part of the screen,
irrespective of the content displayed in the data display region
720. A specimen list display region 931 and a measurement result
display region 932 are displayed in the data display region 720 of
the specimen list screen 700.
[0172] The specimen list display region 931 is provided with items
such as subject ID, specimen ID, and day and time. In the specimen
list display region 931, test result data identified by a subject
ID and a specimen ID is displayed in the form of a list.
[0173] A measurement result based on the specimen ID selected in
the specimen list display region 931 is displayed in the
measurement result display region 932. The measurement result
displayed in the measurement result display region 932 is a
counting result and ratio information based on the main cell
type.
[0174] When the user has selected the detail icon 710b on the
specimen list screen 700, or has selected (e.g., double clicked) a
record of one specimen in the specimen list display region 931, a
detail screen 800 is displayed.
[0175] FIG. 19 shows an example of the detail screen 800. The
detail screen 800 includes a tool bar 810 (710), and a data display
region 820. The tool bar 810 is the same as that displayed on the
specimen list screen 700. A measurement result region 820a for
displaying a measurement result, a flag region 820b for displaying
flag information, and a graph region 820c for displaying a graph
are displayed in the data display region 820. Measurement results
of all of the measurement channels for which measurements have been
performed on the specimen to be displayed are displayed in the
measurement result region 820a. When a flag has been set to the
specimen on the basis of a measurement result, the flag (suspicion
message) is displayed in the flag region 820b. A graph
corresponding to each measurement channel is displayed in the graph
region 820c.
[0176] In the example in FIG. 19, a WDF scattergram, a WNR
scattergram, a WPC scattergram, an RET scattergram, a PLT-F
scattergram, an RBC histogram, and a PLT histogram are displayed in
the graph region 820c. Each scattergram displayed in the graph
region 820c is created on the basis of the feature parameters
described above, and is displayed.
[0177] The user refers to the analysis result displayed on the
detail screen 800, and when the user validates test result data,
the user operates the validation icon 710d. Accordingly, the test
result data is validated. When the user does not validate the test
result data, the user may operate the measurement registration icon
710c to input a retest order, for example.
[0178] In order to confirm classification information of a cell,
the user can cause the analysis screen 900 to be displayed. In the
graph region 820c on the detail screen 800, when the user operates
(e.g., double clicks) a graph, the analysis screen 900 is displayed
in a pop-up manner, to be overlaid on the detail screen 800.
[0179] FIG. 20 is an example of the analysis screen 900. FIG. 20
shows, as an example, the analysis screen 900 that is displayed
when the user has selected a WDF scattergram.
[0180] The scattergram 901 and a detail display region 902 are
displayed on the analysis screen 900. The scattergram 901 shown in
FIG. 20 is displayed on the basis of the feature parameters of the
cell associated with the measurement channel included in the test
result data. In the scattergram 901 shown in FIG. 20, the
horizontal axis represents the peak value of side scattered light
signal (SSCP), and the vertical axis represents the peak value of
side fluorescence signal (SFLP). With respect to each individual
cell specified by a cell ID, the processor 3001 determines a
coordinate in the scattergram 901 on the basis of SSCP and SFLP,
and renders a dot on the determined coordinate, thereby plotting
the cell. The processor 3001 performs similar processes to the data
of all of the cells associated with the measurement channel that
corresponds to the scattergram 901 to be displayed, thereby
plotting the cells on the scattergram 901.
[0181] The processor 3001 sets a color of a dot on the basis of the
classification information of the cell to be plotted. Specifically,
the processor 3001 displays cells of different cell types, in the
form of dots having different colors. For example, in the
scattergram 901 corresponding to the WDF channel, neutrophils are
displayed in a first color (e.g., blue), lymphocytes are displayed
in a second color (e.g., purple), monocytes are displayed in a
third color (e.g., green), eosinophils are displayed in a fourth
color (e.g., red), and basophils are displayed in a fifth color
(e.g., yellow).
[0182] White blood cell classification performed in a conventional
blood cell counter has adopted a method of clustering analysis in
which cells are plotted on a coordinate plane such as the
scattergram 901 shown in FIG. 20, and in which cells are classified
into a plurality of populations on the basis of the coordinate of
each cell. Therefore, cells that have been plotted on the same
coordinate are classified into the same type of cell. Meanwhile, in
the present embodiment, different from the conventional white blood
cell classification, the feature parameters (SSCP and SFLP) of
cells are used only for plotting the cells onto the scattergram
901. As described above, cell classification in the present
embodiment uses the deep learning algorithm 60 to analyze waveform
data of each individual cell, thereby individually identifying the
type of each cell, without using the clustering technique. In other
words, the coordinates of the cell plots in the scattergram 901 are
fragmentary information representing the distribution of cells in a
measurement sample on the basis of parameters indicating the
features of the cells, and the coordinates are not used for
identification of cell types. Therefore, even when a plurality of
cells are plotted on the same coordinate, such cells may be
classified into different cell types by the deep learning algorithm
60. When cells of different types have been plotted on the same
coordinate, the processor 3001 renders, on the basis of the main
cell types of the plurality of cells plotted on the same
coordinate, a dot of a color that corresponds to the main cell type
that has the highest proportion.
[0183] In step S413 of the flow chart in FIG. 17, when the user has
performed an operation (e.g., double click) of selecting a dot on
the scattergram 901, the processor 3001 displays the detail display
region 902 in a pop-up manner. The detail display region 902
includes a classification information display region 902a, and the
classification information display region 902a displays the
classification information of the cell included in the dot selected
on the scattergram 901. FIG. 20 shows an example in which a
plurality of cell types and the probability corresponding to each
cell type are displayed. The classification information display
region 902a is provided with a scroll bar for allowing display of
all of the cell types and probabilities by changing the display
range in the classification information display region 902a. When a
plurality of cells have been plotted on the dot selected on the
scattergram 901, the classification information display region 902a
for each of the plurality of cells is displayed in the detail
display region 902. By operating the scroll bar, the user can cause
the classification information to be displayed in a scrolling
manner.
[0184] On the analysis screen 900 in FIG. 20, the classification
information of a cell included in the selected dot is displayed in
a pop-up manner. However, as shown in FIG. 21, the classification
information display region 902a may be provided in the analysis
screen 900. In this case, until a dot is selected, the
classification information display region 902a remains blank, and
upon selection of a dot, classification information of a cell
corresponding to the selected dot is displayed in the
classification information display region 902a.
[0185] In the above example, the detail display region 902 is
displayed in a pop-up manner in response to an operation (e.g.,
double click) performed on a dot on the scattergram 901. However,
the operation for calling the detail display region 902 can include
various operations as long as the operation can specify a
coordinate on the scattergram 901. For example, the detail display
region 902 may be displayed simply by placing a cursor on the
scattergram 901. In this case, placing the cursor on the
scattergram 901 serves as an operation for calling the detail
display region 902. Instead of a double click, the operation may be
a long press, a right click, or pressing a predetermined key (e.g.,
Enter key) of a keyboard or a software key in a state where the
cursor is placed on a coordinate.
[0186] According to the screens in FIG. 20 and FIG. 21, with
reference to the classification information display region 902a,
the user can confirm the classification information of the cell
that corresponds to the dot selected on the scattergram 901. In
addition, by changing the dot selected on the scattergram 901, the
user can dynamically switch the display content of the
classification information display region 902a, thereby being able
to continuously confirm the display content.
[0187] In the classification information display region 902a, not
limited to the probability for each cell type, another type of
information may be displayed as the classification information. For
example, when the test result data has the configuration shown in
FIG. 13, the value of the main flag for each cell type may be
displayed. When the test result data has the configuration shown in
FIG. 14, a mark may be displayed at the main cell type (the cell
type having the highest probability), or the main cell type may be
separately displayed. When the test result data has the
configuration shown in FIG. 15, the value of the research flag for
each cell type may be displayed. When the test result data has the
configuration shown in FIG. 16, the rank for each cell type may be
displayed.
[0188] In FIG. 20 and FIG. 21, the probability for each cell type
is displayed in terms of a numerical value. Instead of this, as
shown in FIG. 22 and FIG. 23, a graph according to the probability
for each cell type may be displayed. FIG. 22 shows another example
of the analysis screen 900. In the example in FIG. 22, for each
cell corresponding to the selected dot, probability information of
the cell is displayed in the form of a percentage bar chart in the
classification information display region 902a. This percentage bar
chart is displayed so as to correspond to each cell ID, and is
displayed such that the values of the probabilities for the
respective pieces of identification information are accumulated to
be a total of 100%. Accordingly, the user can visually understand
the probability of each cell type.
[0189] FIG. 23 shows another example of the analysis screen 900.
The scattergram 901, the detail display region 902, an enlargement
button 903, and an enlarged view 904 are displayed on the analysis
screen 900 in FIG. 23.
[0190] When the user has operated the enlargement button 903, a
gate 901a for a reference range is displayed on the scattergram
901. The gate 901a is for selecting a desired dot on the
scattergram 901 through range designation. The user can change the
position of the gate 901a on the scattergram 901 by performing an
operation such as dragging. In the enlarged view 904, a portion of
the scattergram 901 in the gate 901a is displayed in an enlarged
manner. When the user has performed an operation of selecting a dot
on the enlarged view 904, the classification information of the
selected cell is displayed in the classification information
display region 902a in the detail display region 902, as in the
cases of FIG. 20 and FIG. 21.
[0191] According to the screen in FIG. 23, the user can set the
gate 901a on the scattergram 901 to cause the enlarged view 904 to
be displayed, thereby being able to easily select the target cell
in the enlarged view 904. Accordingly, with respect to a cell in a
range where the dots of cells are densely present in the
scattergram 901 as well, the classification information can be
smoothly confirmed.
[0192] The size of the gate 901a may be changed by performing an
operation such as dragging on the boundary line of the gate 901a.
The size of the gate 901a may be changed stepwise every time an
operation is performed on the enlargement button 903.
[0193] In each of FIG. 20 to FIG. 23, an example of selecting one
dot on the scattergram 901 is shown. However, the number of dots
selectable at one time is not limited to one. For example, a
plurality of dots may be selected through range designation on the
scattergram 901, by dragging a cursor while clicking a point on the
scattergram 901. However, when a plurality of dots are selected at
one time, the number of cells to be displayed may become too large.
This may cause difficulty for the user to view the classification
information when the classification information is individually
displayed for each cell. Therefore, as will be later described with
reference to FIG. 29, the display method may be switched in
accordance with the number of selected dots. For example, when the
number of selected dots is not greater than a predetermined
threshold, pieces of the classification information of cells may be
individually displayed, and when the number of selected dots is
greater than the predetermined threshold, the classification
information may be statistically displayed, instead of being
individual displayed.
[0194] FIG. 24 is a flow chart showing a second example of the test
result displaying process. In FIG. 24, when compared with the flow
chart in FIG. 17, step S413 has been changed. In step S413 in FIG.
24, the classification information of each of the cells
corresponding to the dot selected on the scattergram 901 is read
out, and is displayed as statistic information on the analysis
screen 900.
[0195] In the first example of the test result displaying process
in FIG. 17, as shown in the examples in FIG. 20 to FIG. 23, the
classification information of each of cells corresponding to the
selected dot is individually displayed for each cell. Instead of
this, in the examples in FIG. 25 to FIG. 28 shown below, pieces of
the classification information of a plurality of cells
corresponding to the selected dot are aggregated, and statistic
information is displayed. Examples of the analysis screen 900 are
shown in FIG. 25 to FIG. 28.
[0196] The scattergram 901, the enlargement button 903, the
enlarged view 904, and a detail display region 905 are displayed on
the analysis screen 900 in FIG. 25.
[0197] On the basis of the classification information of all of the
cells in the gate 901a, statistic information of the probability
and the number of cells for each cell type is displayed in the form
of a histogram 905a in the detail display region 905. The histogram
905a is frequency distribution information obtained by subjecting
all of the cells in the gate 901a to counting for each probability
in the cell type. In the histogram 905a, the horizontal axis
represents probability and the vertical axis represents the number
of cells. The numbers shown in the histogram 905a include a
counting result of cells that are not of the main cell type (cells
that have a low probability of belonging to the cell type). The
detail display region 905 is provided with a scroll bar for
allowing display of the histogram 905a of all of the cell types by
changing the display range in the detail display region 905.
[0198] According to the screen shown in FIG. 25, by causing the
histogram 905a for each cell type to be displayed in the detail
display region 905, the user can understand the distribution of the
probabilities for each cell type, with respect to all of the cells
in the gate 901a. For example, in the conventional clustering
technique using a scattergram, there are cases where distribution
ranges of a plurality of cell types overlap each other on the
scattergram 901. In contrast to this, according to the present
embodiment, when the gate 901a is set to such a range where the
distributions overlap each other, and then the histogram 905a is
referred to, the frequency distribution of each cell type can be
confirmed. Accordingly, the user can perform detailed determination
for the specimen, and, for example, can determine how much the
cells of each cell type can be included in the range where the
distributions overlap each other.
[0199] FIG. 26 shows another example of the analysis screen 900.
The scattergram 901, the enlargement button 903, the enlarged view
904, a cell type selection region 906, and a three-dimensional
histogram 907 as statistic information are displayed on the
analysis screen 900.
[0200] The cell type selection region 906 is provided with check
boxes for selecting cell types. The cell type selection region 906
is provided with a scroll bar for allowing display of all of the
cell types by changing the display range in the cell type selection
region 906. The three-dimensional histogram 907 is frequency
distribution information obtained by subjecting all of the cells in
the gate 901a to counting for each probability in a selected cell
type. In the three-dimensional histogram 907, histograms for
respective selected cell types are arranged in the front-rear
direction so as to be displayed in a three-dimensional manner. In
the example shown in FIG. 26, since lymphocyte and monocyte have
been selected in the cell type selection region 906, a histogram of
lymphocyte and a histogram of monocyte are displayed together in
the three-dimensional histogram 907.
[0201] According to the analysis screen 900 in FIG. 26, by
selecting two or more cell types and referring to the
three-dimensional histogram 907, the user can simultaneously
confirm the frequency distributions of the two or more cell types.
In the example in FIG. 25, two histograms 905a are simultaneously
displayed in the screen. However, in the example in FIG. 26, when
three or more cell types are selected, three or more histograms are
displayed in the screen. Thus, in the case of the screen in FIG.
26, when compared with the screen in FIG. 25, the user can
simultaneously understand a greater number of frequency
distributions. In addition, the user can easily compare the
frequencies for each probability.
[0202] The three-dimensional histogram 907 in FIG. 26 is a
histogram in which a plurality of the histograms 905a shown in FIG.
25 are arranged in the depth direction. However, instead of this,
distribution states of probabilities of a plurality of cell types
may be displayed in terms of surface plot or the like.
[0203] FIG. 27 shows another example of the analysis screen 900.
The scattergram 901, the enlargement button 903, the enlarged view
904, and a number display region 908 are displayed on the analysis
screen 900 in FIG. 27.
[0204] In the number display region 908, with respect to all of the
cells in the gate 901a, a count value obtained by counting, for
each cell type, cells that have a probability higher than 0% is
displayed in the form of a bar graph as statistic information. In
this case, a plurality of cells in the gate 901a are redundantly
counted, and thus, the total of the count values in the number
display region 908 becomes greater than the actual number of cells
in the gate 901a. For example, it is assumed that one cell that has
probability information of monocyte 90%, neutrophil 5%, and
immature granulocyte 5% is included in the gate 901a in FIG. 27. In
this case, this cell is counted as 1 for each of monocyte,
neutrophil, and immature granulocyte. That is, when the probability
of belonging to a plurality of cell types is 0% or higher, the
number of the cell is redundantly counted for the plurality of cell
types. The aggregation method is not limited thereto, and the
number may be expressed by a decimal not greater than 1 in
accordance with the probability. For example, as in the above
example, when one cell that has probability information of monocyte
90%, neutrophil 5%, and immature granulocyte 5% is included in the
reference range of the gate 901a, aggregation may be performed such
that 0.9 monocyte, 0.05 neutrophil, and 0.05 immature granulocyte
are considered to be included.
[0205] With respect to a bar graph in the number display region
908, when the counting result of the cell type is smaller than a
predetermined threshold (e.g., less than 10), the count value of
the cell type is added to the bar graph. In the example in FIG. 27,
since the count value of immature granulocytes is five, "5" is
added to the bar graph of immature granulocyte. The number display
region 908 is provided with a scroll bar for allowing display of
all of the cell types by changing the display range in the number
display region 908.
[0206] According to the analysis screen 900 in FIG. 27, by
referring to the number display region 908, the user can confirm
the number of cells for each cell type in the gate 901a. Since a
cell that has a probability exceeding a predetermined percentage is
counted as one, this configuration is convenient when confirming a
majority cell type and a minority cell type. A cell type that has a
small count value in the number display region 908 is provided with
a count value as auxiliary display so as to allow confirmation of
the magnitude of the count value. Therefore, the count value of a
cell type that has a small count value can be clearly confirmed.
Accordingly, the user can smoothly determine a possibility that
cells of a rare cell type can be included in the specimen.
[0207] In the number display region 908, a count value obtained by
counting cells of a cell type having a probability higher than 0%
is displayed. However, the threshold for counting is not limited to
0%, and a value greater than 0% and smaller than 100% may be set. A
user interface (e.g., a probability selection region 910 in FIG.
31) that can receive a threshold for counting may be provided in
the screen.
[0208] FIG. 28 shows another example of the analysis screen 900.
The scattergram 901 and a detail display region 913 are displayed
on the analysis screen 900.
[0209] When the user has performed an operation of selecting a dot
on the scattergram 901, statistic information about a cell
corresponding to the selected dot is displayed in the detail
display region 913. The detail display region 913 has regions 913a
to 913d. In the region 913a, with respect to one or a plurality of
cells (hereinafter, "selected cell") plotted at the selected dot,
the cell types that have the highest two probabilities are
displayed. In a region 913b, with respect to the selected cell, the
cell type that has the highest probability among the five
classifications of white blood cells is displayed. In a region
913c, with respect to the selected cell, the abnormal cell type
that has the highest probability is displayed. In the region 913d,
with respect to the selected cell, cell types that may be present,
i.e., cell types that have a probability higher than 0%, are
displayed.
[0210] According to the analysis screen 900 in FIG. 28, by
referring to the detail display region 913, the user can understand
what cell type the selected cell can be classified into.
[0211] In the vicinities of the names of the cell types displayed
in the regions 913a to 913d, the probabilities of the corresponding
cell types may be additionally displayed. When the number of
selected cells is one, the probability information of the
corresponding cell can be displayed. When the number of selected
cells is a plurality, a representative value of the probabilities
of the cell types of the plurality of cells may be obtained and
displayed. The representative value can be expressed in terms of
the average value, the median, or the mode, for example. In the
region 913a, only the cell type at the highest rank may be
displayed, or cell types at the highest three or more ranks may be
displayed. In the region 913b, 913c, cell types at the highest two
or more ranks may be displayed. In the region 913d, cell types that
have probabilities higher than a predetermined threshold may be
displayed.
[0212] FIG. 29 is a flow chart showing a third example of the test
result displaying process. In step S414, the processor 3001
compares a threshold with the number of cells that correspond to
the dot selected on the scattergram 901. When the number of cells
is not greater than the threshold, the processor 3001 individually
displays the classification information for each cell in step S416.
The individual display of the classification information for each
cell has been described with reference to FIG. 20 to FIG. 23. When
the number of cells is greater than the threshold, the processor
3001 aggregates the classification information and displays
statistic information in step S415. The display of statistic
information has been described with reference to FIG. 25 to FIG.
28.
[0213] As described above, when a plurality of dots are selected at
one time, or when a wide range for the gate 901a is set, the number
of cells to be displayed may become too large. Thus, when the
classification information is individually displayed for each cell,
the visibility for the user may be impaired in some cases.
Meanwhile, when the number of cells corresponding to the selected
dot is small, individual display of the classification information
of each cell may be suitable for the purpose of the user in some
cases. For example, as in the case of the dot selected on the
scattergram 901 in FIG. 28, a position away from the center of a
cell cluster or a position where cell distribution is scarce has
been selected by the user, the user may desire to analyze cells
individually, not as a population. Therefore, in the example in
FIG. 29, when the number of selected dots is not greater than a
predetermined threshold, pieces of classification information of
cells are individually displayed, and when the number of selected
dots is greater than the predetermined threshold, the
classification information is not individually displayed but
statistically displayed. Accordingly, a screen having a high
visibility for the user can be automatically provided.
[0214] FIG. 30 shows a fourth example of the test result displaying
process. In the first example to third example of the test result
displaying process described above, when a dot of a target cell is
selected on the scattergram 901, the classification information of
the cell corresponding to the dot is displayed. In the examples in
FIG. 31 to FIG. 38 described below, the analysis screen 900 that
includes a graph is displayed (step S417), and dots of cells that
satisfy a defined extraction condition are highlighted (step S418).
In step S418, selection of a dot is not performed on the
scattergram 901. Instead, when the user has defined a condition for
a cell to be extracted, cells that satisfy the condition are
extracted and displayed in the graph.
[0215] FIG. 31 shows another example of the analysis screen 900.
The scattergram 901, a cell type selection region 909, and a
probability selection region 910 are displayed on the analysis
screen 900.
[0216] The cell type selection region 909 and the probability
selection region 910 are each implemented as a so-called pull-down
menu. The cell type selection region 909 is configured such that
one cell type can be selected from among all the cell types that
correspond to the measurement channel (in FIG. 31, the WDF channel)
selected by the user. In the probability selection region 910,
numerical values of probability, as choices, are displayed so as to
be selectable at a 5% interval in a range of 0% to 100%.
Specifically, the probability selection region 910 is configured
such that a character string composed of an equal sign or an
inequality sign and a probability, such as "=100%", ">95%",
">90%", . . . , ">5%", ">0%, or "=0%", can be
selected.
[0217] When the user has selected a target cell type in the cell
type selection region 909 and has selected a probability in the
probability selection region 910, a corresponding dot on the
scattergram 901 is highlighted. In the example shown in FIG. 31,
basophil has been selected as the cell type, and ">85%" has been
selected as the probability. In this case, dots (dots in a region
901b) of cells for which the probability of being basophil is
higher than 85% are highlighted on the scattergram 901. The method
of highlighting may be any display method that can distinguish the
extracted dot from the other dots, such as displaying the dot in a
color different from the colors for the other dots, displaying the
contour of the dot in an emphasized manner, or blinking the dot,
for example.
[0218] According to the screen in FIG. 31, by defining the target
cell type and probability as the extraction condition, the user can
understand the position of the cell that matches the condition on
the scattergram 901. Therefore, the user can understand at which
position on the conventional scattergram 901 the cell classified
according to the classification information (cell type and
probability) is present. This information can be used as a material
when the accuracy of classification according to the classification
information is determined, for example.
[0219] FIG. 32 shows another example of the analysis screen 900.
The analysis screen 900 in FIG. 32 includes a scale bar 910a for
selecting a numerical value of probability and an extraction target
selection region 910c, instead of the probability selection region
910. The scale bar 910a includes a pointer 910b slidable in the
left-right direction. The user can select a desired percentage in
the range of 0% to 100% by moving the pointer 910b in the
left-right direction. In the extraction target selection region
910c, an inequality sign combined with the percentage selected by
the scale bar 910a is displayed so as to be selectable. For
example, as shown in FIG. 32, when basophil is selected as the cell
type in the cell type selection region 909, the pointer 910b is
placed at the position of 80% in the scale bar 910a, and ">" is
selected in the extraction target selection region 910c, the
extraction condition is set as "probability of basophil>80%". In
the scattergram 901, dots of cells that satisfy the extraction
condition of "probability of basophil>80%" are highlighted.
[0220] When the user moves the pointer 910b in the scale bar 910a,
the numerical range of percentage included in the extraction
condition is continuously (stepwise) changed in association
therewith. When the extraction condition has been changed, dots
extracted on the scattergram 901 are dynamically changed,
accordingly.
[0221] For example, when the pointer 910b is caused to slide
rightward from the state shown in FIG. 32, the extraction condition
is changed from "probability of basophil>80%" to "probability of
basophil>85%". Due to this change in the extraction condition,
dots of cells highlighted on the scattergram 901 are also
dynamically changed from those corresponding to "probability of
basophil>80%" to those corresponding to "probability of
basophil>85%". That is, the number of dots highlighted
decreases. Conversely, when the pointer 910b is caused to slide
leftward, the extraction condition is changed from "probability of
basophil>80%" to "probability of basophil>75%". Due to this
change in the extraction condition, dots of cells highlighted on
the scattergram 901 are also dynamically changed from those
corresponding to "probability of basophil>80%" to those
corresponding to "probability of basophil>75%". That is, the
number of dots displayed in a highlighted manner increases.
[0222] Thus, when the user causes the pointer 910b of the scale bar
910a to slide, thereby continuously changing the extraction
condition, and, in association therewith, the user causes the
highlight display of dots of cells that match the extraction
condition displayed on the scattergram 901 to be dynamically
changed, the user can easily understand at which position on the
scattergram 901 the target cell is present.
[0223] FIG. 33 shows another example of the analysis screen 900.
FIG. 33 shows another example regarding setting of the extraction
condition. The analysis screen 900 in FIG. 33 includes condition
setting regions 910d, 910e, 910f for respectively setting an
extraction condition 1, an extraction condition 2, and an
extraction condition 3. The condition setting regions 910d, 910e,
910f each include pull-down menus that respectively correspond to
the cell type selection region 909 and the probability selection
region 910 shown in FIG. 31. Two operator setting regions 910g,
910h are provided between the condition setting regions 910d, 910e,
910f. The operator setting regions 910g, 910h are each implemented
as a pull-down menu, and can set a logical operator for combining
the extraction conditions 1 to 3. Choices of the logical operator
are "AND", "OR", "NOT", and the like. In the example in FIG. 33,
the logical operators are each "AND".
[0224] When the extraction conditions 1 to 3 are defined in the
respective condition setting regions 910d, 910e, 910f, and logical
operators are set in the operator setting regions 910g, 910h, a
complex condition in which the extraction conditions 1 to 3 are
combined by the logical operators is defined. In the scattergram
901, dots (dots in the region 901b) of cells that satisfy the
defined complex condition are highlighted. As in the example in
FIG. 33, since cells are extracted not using a single extraction
condition obtained as a one-to-one combination of cell type and
probability, but using a complex condition obtained through a
combination of a plurality of extraction conditions, more detailed
extraction condition setting can be realized.
[0225] FIG. 34 shows another example of the analysis screen 900.
The scattergram 901, a first histogram 910i, a second histogram
910j, and a display region 910k are displayed on the screen in FIG.
34.
[0226] In the example shown in FIG. 34, only the scattergram 901 is
displayed as default on the analysis screen 900. The user can
select a specific type of cell population on the displayed
scattergram 901. For example, when the user has placed a cursor on
a specific cell population on the scattergram 901, the cell
population is selected.
[0227] FIG. 34 shows a state where a population of lymphocytes has
been selected. When a specific type of cell population has been
selected, the first histogram 910i showing the probability
distribution of the selected cell population, and the second
histogram 910j showing the probability distribution regarding
abnormal cells that correspond to the selected cell population are
displayed on the analysis screen 900. The first histogram 910i and
the second histogram 910j are each similar to the histogram 905a in
FIG. 25, for example, and indicates the number of cells with
respect to the axis representing the probability. Abnormal cells
corresponding to the selected cell population mean abnormal cells
that are morphologically similar to the selected cell population.
The abnormal cells are determined on the basis of a correspondence
relationship determined in advance, in accordance with the selected
cell population. For example, when lymphocyte has been selected,
the abnormal cells can be abnormal lymphocytes or blasts. For
example, when monocyte has been selected, the abnormal cells can be
blasts. For example, when eosinophil or basophil has been selected,
the abnormal cells can be immature granulocytes.
[0228] In the example in FIG. 34, when a specific cell population
has been selected, cells that have probabilities of being abnormal
cells corresponding to the selected cell population are extracted,
and the distribution of the probabilities is displayed as the
second histogram 910j. When the user compares the probability
distribution of the selected cell population with the probability
distribution of abnormal cells, the user can understand how close
to 100% the probability distribution of the selected cell
population is. Conversely, the user can statistically understand
how many cells that have probabilities of being abnormal cells are
included in the selected cell population.
[0229] The analysis screen 900 in FIG. 34 includes the display
region 910k that displays an indicator based on statistic
information, in addition to the first histogram 910i and the second
histogram 910j. In the display region 910k, "In the lymphocyte
population, the proportion of cells that have probabilities of
being abnormal lymphocytes is 15% (previous value: 12%)." is
displayed, for example. That is, the proportion of cells that have
the selected cell population as the parent population and that have
probabilities of being corresponding abnormal cells is displayed as
an indicator in the display region 910k.
[0230] In the example in FIG. 34, the sample with respect to the
parent population is defined as "cells that have probabilities of
being abnormal cells". However, the definition of the sample is not
limited thereto. For example, the sample may be defined as "cells
that are in the second highest rank in which the cells have
probabilities of being abnormal cells. In this case, only cells
that have higher probabilities of being abnormal cells can be
extracted so as to calculate an indicator, which may lead to
reduction of noise. The indicator is not limited to a quantitative
indicator such as a proportion, and may be a qualitative indicator
that indicates the level stepwise in accordance with the absolute
number of cells that have probabilities of being abnormal
cells.
[0231] In the example in FIG. 34, the number or proportion of cells
that have probabilities of being abnormal cells can be easily
understood with respect to a specific cell population. The
configurations in FIG. 25 to FIG. 28 described above are
advantageous in that the user can designate a specific cell or cell
population on the scattergram 901 as desired. However, in those
configurations, it may be difficult for the user to designate only
a specific cell population on the scattergram 901. When a cell
present at the circumference of a cluster is also at an adjacent
cluster or is separated from a cluster, such designation may be
especially difficult. In the example in FIG. 34, simply by placing
a cursor on a specific cell population on the scattergram 901,
e.g., simply by placing a cursor on a population of lymphocytes, a
cell population that has lymphocytes as main cells can be selected.
In addition, cells that have probabilities of being abnormal
lymphocytes, which are abnormal cells that correspond to
lymphocytes, are automatically extracted, and a probability
distribution and an indicator are displayed. Therefore, the user
can easily recognize how many cells that have probabilities of
being abnormal cells are included although the main cells are
normal cells.
[0232] FIG. 35 shows another example of the analysis screen 900.
The scattergram 901, the cell type selection region 909, and a
histogram 911 are displayed on the screen in FIG. 35.
[0233] FIG. 35 shows an example in which the WPC channel has been
designated as the measurement channel to be displayed. Therefore,
the scattergram 901 is displayed on the basis of the feature
parameters of all of the cells that correspond to the measurement
channel, and the cell type selection region 909 is configured such
that, as the cell types to be displayed, all cell types, such as
mature white blood cell, abnormal lymphocyte, and blast, that are
associated with the WPC channel can be selected.
[0234] When the user has performed an operation of selecting a cell
type in the cell type selection region 909, the histogram 911 is
displayed on the basis of the selected cell type. The histogram 911
is frequency distribution information obtained by counting cells of
the selected cell type for each probability. In the histogram 911,
the horizontal axis represents probability and the vertical axis
represents the number of cells. When the user has performed an
operation of selecting a bar graph on the histogram 911,
corresponding dots (dots in a region 901c) on the scattergram 901
are highlighted. In the example shown in FIG. 35, abnormal
lymphocyte has been selected as the cell type, and the bar graph of
a probability of 70% has been selected. In this case, dots of cells
(cells distributed near the region 901c) for which probabilities of
being abnormal lymphocytes are on the order of 70% are highlighted
on the scattergram 901.
[0235] According to the analysis screen 900 in FIG. 35, similar to
the analysis screens 900 in FIG. 31 to FIG. 33, when the user
selects a cell type and a probability, the user can understand the
position of each cell that matches the condition on the scattergram
901. Therefore, the user can understand at which position on the
conventional scattergram 901 the cell classified according to the
classification information (cell type and probability) is
present.
[0236] In the examples in FIG. 31 to FIG. 35 described above, dots
of cells that satisfy an extraction condition composed of a
combination of a cell type and a probability are displayed on the
scattergram 901. Instead of the scattergram 901, the number of
cells that satisfy the extraction condition may be displayed in
terms of a histogram as shown in FIG. 36 to FIG. 38.
[0237] The probability selection region 910 and a number display
region 912 are displayed on the analysis screens 900 in FIG. 36 to
FIG. 38.
[0238] On the analysis screen 900 in FIG. 36, when the user has
performed an operation of selecting a probability in the
probability selection region 910, the number of cells corresponding
to the probability selected in the probability selection region
910, among all of the cells corresponding to the target measurement
channel, is displayed for each cell type in the number display
region 912. The number display region 912 is provided with a
descending order button 912a for performing setting such that cell
types are arranged downwardly staring from the cell type having the
greatest number; and an ascending order button 912b for performing
setting such that cell types are arranged downwardly starting from
the cell type having the smallest number. FIG. 36 shows a state
where the descending order button 912a has been operated. FIG. 37
shows a state where the ascending order button 912b has been
operated. FIG. 38 shows a state where, on the screen in FIG. 37,
">20%" has been set as a probability condition.
[0239] According to the screens in FIG. 36 to FIG. 38, after the
user has selected a probability in the probability selection region
910, the user can, by operating the descending order button 912a,
set display of the number display region 912 such that the numbers
are arranged in a descending order as shown in FIG. 36, and the
user can, by operating the ascending order button 912b, set display
of the number display region 912 such that the numbers are arranged
in an ascending order as shown in FIG. 37. Accordingly, among the
cell types corresponding to the probability of the selected
condition, the user can smoothly confirm a cell type of which the
number of cells is large or a cell type of which the numbers of
cells is small, for example.
[0240] When the user selects again another probability in the
probability selection region 910 in the state shown in FIG. 37, the
user can reset the display of the number display region 912 in
accordance with the reselected probability as shown in FIG. 38.
Thus, by operating the probability selection region 910, it is
possible to smoothly confirm the number of cells of each cell type
based on a different probability condition.
[0241] The scattergram 901 may be provided also in FIG. 36 to FIG.
38. In this case, when a bar graph in the number display region 912
has been selected, cells (dots) corresponding to the selected bar
graph may be highlighted in the scattergram 901. In addition, the
histogram 911 (see FIG. 35) may be provided also in FIG. 36 to FIG.
38. In this case, when a bar graph in the number display region 912
has been selected, the state of distribution of the probabilities
regarding the cell type of the selected bar graph may be displayed
on the histogram 911.
[0242] FIGS. 39A, 39B each schematically show another configuration
of the scattergram 901 described above.
[0243] In the scattergram 901 shown in FIG. 39A, distribution
information of cells according to the probabilities for each cell
type is superimposed on the plots of cells according to the feature
parameters (in FIG. 39A, SSCP, SFLP). The distribution information
in this case is composed of a ring pattern in which cells having
the same probability are connected in a contour line shape; and a
distribution region in which ranges each having similar
probabilities are expressed with different darkness. In FIG. 39A,
the higher the probability is, the darker color the distribution
region has. In FIG. 39A, cells (dots) based on the feature
parameters are plotted on the scattergram 901. However, as shown in
FIG. 39B, dots of cells based on the feature parameters may not
necessarily be plotted on the scattergram 901.
[0244] In the scattergrams 901 shown in FIGS. 39A, 39B,
distribution information of cells according to the probabilities
for each cell type is also displayed in the conventional
scattergram 901 based on the feature parameters. Accordingly, with
reference to the conventional scattergram 901, the user can
visually understand how the probabilities of each cell type are
distributed.
[0245] FIG. 40 is a flow chart of a fifth example of the test
result displaying process. In the first to fourth examples
described above, the analysis screen 900 is displayed when a
scattergram on the detail screen 800 has been double clicked. In
the fifth example, when a "research" tab has been selected on the
detail screen 800 in FIG. 19, a research screen 900R is displayed
instead of the analysis screen 900 (step S419). In this example, an
analysis result based on a main cell type is displayed on the
specimen list screen 700 and the detail screen 800, and an analysis
result screen or information regarding cell types other than the
main cell type is displayed only on the research screen 900R. Then,
on the basis of a research flag, classification information of the
selected cell type is read out, and a graph is displayed (step
S420).
[0246] The screen in FIG. 41 is an example of a screen that is
displayed when a research information button 921 has been operated.
A cell type selection region 933 and a number display region 934
are displayed on the screen in FIG. 41. The cell type selection
region 933 includes check boxes each for allowing selection of a
cell type regarding the designated measurement channel.
[0247] When the user has performed an operation of selecting cell
types in the cell type selection region 933, a count value obtained
by counting, with respect to the selected cell types, cells of each
cell type that has a probability higher than 0% is displayed, for
each cell type, in the form of a bar graph in the number display
region 934.
[0248] Thus, when the user has operated the research information
button 921, and then selected a cell type in the cell type
selection region 933 on the screen in FIG. 41, the user can confirm
a detailed analysis result regarding minor cells, with reference to
the number display region 934. That is, as shown in FIG. 41, in a
research information display mode, the numbers of blasts and
abnormal lymphocytes that cannot usually be main cells are also
displayed. Therefore, the user can perform a further detailed
analysis on the basis of the analysis result displayed in the
research information display mode.
[0249] The screen shown in FIG. 42 is an example of a screen that
is displayed when the research information button 921 has been
operated. The cell type selection region 933, a histogram 935, and
a scattergram 936 are displayed on the screen in FIG. 42.
[0250] When the user has performed an operation of selecting a cell
type in the cell type selection region 933, the histogram 935 is
displayed on the basis of the selected cell type. The histogram 935
is frequency distribution information obtained by counting, for
each probability of the selected cell type, all the cells that
correspond to the measurement channel. When the user has selected a
bar graph in the histogram 935, corresponding cells (dots) (dots in
a region 936a) on the scattergram 936 are highlighted. In the
example shown in FIG. 42, blast has been selected as the cell type,
and a bar graph of a probability of 50% has been selected. In this
case, cells (cells distributed near the region 936a) for which
probabilities of being blasts are on the order of 50% are
highlighted on the scattergram 936.
[0251] Thus, when the user has operated the research information
button 921, then, on the screen in FIG. 42, selected a cell type in
the cell type selection region 933 and selected a bar graph on the
histogram 935, the user can confirm a further detailed analysis
result.
[0252] When a plurality of cell types have been selected in the
cell type selection region 933 in FIG. 42, graphs of the plurality
of cell types are displayed in the histogram 935, as in the
three-dimensional histogram 907 in FIG. 26. When a plurality of bar
graphs have been selected on the histogram 935, all of the
corresponding cells (dots) on the scattergram 936 are
highlighted.
7. HOST COMPUTER TRANSMISSION PROCESS
[0253] Next, data (hereinafter, referred to as "output data") to be
sent from the processing unit 300 to the host computer 500 is
described. FIG. 43 is a flow chart showing a subroutine of host
computer transmission.
[0254] The processor 3001 of the processing unit 300 receives a
validation operation from the user and generates output data to be
transmitted to the host computer 500 in accordance with the
validation operation (step S141).
[0255] As described with reference to FIG. 18 and FIG. 19, on the
specimen list screen 700 and the detail screen 800 which are each a
test result screen, the validation icon 710d displayed in common
between these screens is displayed. The user confirms the test
result screen described above, and performs verification on the
test result. As a result of the verification, when the user has
determined that the result can be reported to the host computer
500, the user clicks the validation icon 710d. When the user does
not execute validation on the target specimen and sets a retest for
the target subject, the user operates the measurement registration
icon 710c to create a test order for the retest.
[0256] When a validation operation has been performed, the
processor 3001 generates output data to be transmitted to the host
computer 500, on the basis of analysis result data stored in the
storage 3004 (step S141). Then, the processor 3001 transmits the
generated output data to the host computer 500 (step S142).
[0257] FIG. 44A to FIG. 45B each schematically show a configuration
of output data that is generated by the processor 3001 and that is
to be transmitted to the host computer 500. The output data may be
formed by removing classification information that corresponds to a
certain cell type from the test result data, and is configured as
shown in one of those shown in FIG. 44A to FIG. 45B, for
example.
[0258] The output data shown in FIG. 44A is composed of: specimen
ID; measurement result; measurement channel; cell ID;
identification information regarding a main cell type; and
probability regarding the main cell type. That is, the output data
shown in FIG. 44A is obtained by removing data other than the
output data shown in FIG. 44A from the test result data shown in
FIGS. 12 to 16.
[0259] The output data shown in FIG. 44B is composed of: specimen
ID; measurement result; measurement channel; cell ID;
identification information regarding cell types that have the
highest two probabilities; and probabilities regarding the cell
types that have the highest two probabilities. That is, the output
data shown in FIG. 44B is obtained by adding, to the output data
shown in FIG. 44A, the identification information and probability
regarding the cell type that has the second highest probability.
The output data shown in FIG. 44B is not limited to having the
highest two cell types and probabilities, and may have cell types
and probabilities from the highest rank to a predetermined
rank.
[0260] The output data shown in FIG. 45A is composed of specimen ID
and measurement result. That is, the output data shown in FIG. 45A
is obtained by removing data other than the output data shown in
FIG. 45A from the test result data shown in FIGS. 12 to 16. That
is, the output data does not include information regarding
probability at all.
[0261] The output data shown in FIG. 45B is composed of specimen
ID, measurement result, and abnormal cell detection information.
That is, the output data shown in FIG. 45B is obtained by adding
abnormal cell detection information to the measurement result in
the output data similar to that in FIG. 45A. When not less than a
predetermined number of abnormal cells (e.g., abnormal lymphocytes
or blasts) that have a probability higher than 0% are present in
the specimen, a fact that the abnormal cells have been detected is
added as abnormal cell detection information to the measurement
result at the time of generation of the test result data. The
abnormal cell detection information is a character string such as
"abnormal lymphocytes have been detected", "Abn LYMPH?", "blasts
have been detected", or "Blast?", for example. The abnormal cell
detection information may be the cell type of the detected abnormal
cells and the number thereof. The abnormal cell detection
information is displayed in the measurement result display region
932 in FIG. 18, for example.
[0262] As described above, output data that includes a counting
result and in which at least a part of relevance information (such
as cell type and probability) has been removed is transmitted to
the host computer 500. As described above, when probabilities
associated with a plurality of cell types are obtained for each
cell, the volume of test result data becomes huge. Therefore, if
all of the test result data is transmitted to the host computer
500, the communication load due to data transmission becomes large.
In contrast to this, according to the output data shown in FIG. 44A
to FIG. 45B, not all the huge test result data is transmitted, but
only data that is necessary for management by the host computer 500
is transmitted. Accordingly, both of appropriate data management in
the host computer 500 and efficient data transmission can be
realized.
[0263] The output data shown in FIG. 44A to FIG. 45B may include
feature parameters and scattergrams.
[0264] Here, the processor 3001 generates the output data so as to
have a configuration of one of those shown in FIG. 44A to FIG. 45B.
However, the configuration of the output data may be set in
accordance with an instruction from a user.
[0265] FIG. 46A and FIG. 46B each schematically show a
configuration of a screen for selecting data to be included in the
output data.
[0266] A cell type selection region 951 and an OK button 952 are
displayed on the screen in FIG. 46A. The cell type selection region
951 is provided with check boxes each for selecting a cell type.
When the user has performed an operation of selecting a cell type
in the cell type selection region 951 and has operated the OK
button 952, the processor 3001 of the processing unit 300 stores
the set content into the storage 3004. From the next time, when
generating output data, the processor 3001 causes the set cell type
and the probability corresponding to the set cell type to be
included in the output data.
[0267] A cell range selection region 961 and an OK button 962 are
displayed on the screen in FIG. 46B. The cell range selection
region 961 is provided with check boxes each for selecting a cell
type range. A cell type range, such as "main cell type only",
"minor cell types other than main cell type", "all cell types", or
"do not transmit", can be selected in the cell range selection
region 961. When "main cell type only" has been selected, only the
identification information and probability of the main cell type
among all of the identification information and probabilities are
caused to be included in the output data. When "minor cell types
other than main cell type" has been selected, only the
identification information and probabilities of all of the minor
cell types other than the main cell type among all of the
identification information and probabilities are caused to be
included in the output data. When "all cell types" has been
selected, all of the identification information and probabilities
are caused to be included in the output data. When "do not
transmit" has been selected, none of the identification information
and probabilities are caused to be included in the output data.
[0268] When the user has performed an operation of selecting a cell
type range in the cell range selection region 961 and has operated
the OK button 962, the processor 3001 stores the set content into
the storage 3004. From the next time, when generating output data,
the processor 3001 generates output data on the basis of the set
cell type range.
8. CONFIGURATIONS OF PROCESSOR AND PARALLEL-PROCESSING
PROCESSOR
Configuration Example 1
[0269] Next, a configuration of the parallel-processing processor
4833 installed in the measurement unit 400 is described in detail
with reference FIG. 2 again. In the following, configuration
example 1 is described with respect to the processor 4831 and the
parallel-processing processor 4833.
[0270] As described with reference to FIG. 2, using the
parallel-processing processor 4833, the processor 4831 executes an
analysis process of waveform data included in a generated digital
signal, in accordance with the deep learning algorithm 60. That is,
the processor 4831 is programmed to execute an analysis process of
waveform data included in a digital signal in accordance with the
deep learning algorithm 60. The analysis software 4835a for
analyzing data of cells on the basis of the deep learning algorithm
60 may be stored in the storage 4835. In this case, the processor
4831 executes the analysis software 4835a stored in the storage
4835, thereby executing a data analysis process based on the deep
learning algorithm 60. The processor 4831 is a CPU (Central
Processing Unit), for example. For example, core i9, Core i7, or
Core i5 manufactured by Intel Corporation, Ryzen 9, Ryzen 7, Ryzen
5, or Ryzen 3 manufactured by AMD, or the like may be used as the
processor 4831.
[0271] The processor 4831 controls the parallel-processing
processor 4833. The parallel-processing processor 4833 executes
parallel processing regarding a matrix operation, for example, in
accordance with control by the processor 4831. That is, the
processor 4831 is a master processor of the parallel-processing
processor 4833, and the parallel-processing processor 4833 is a
slave processor of the processor 4831. The processor 4831 is also
referred to as a host processor or a main processor.
[0272] The parallel-processing processor 4833 executes in parallel
a plurality of arithmetic processes being at least a part of
processing regarding analysis of waveform data. The
parallel-processing processor 4833 is a GPU (Graphics Processing
Unit), an FPGA (Field Programmable Gate Array), or an ASIC
(Application Specific Integrated Circuit), for example. When the
parallel-processing processor 4833 is an FPGA, the
parallel-processing processor 4833 may have programed therein in
advance an arithmetic process regarding a trained deep learning
algorithm 60, for example. When the parallel-processing processor
4833 is an ASIC, the parallel-processing processor 4833 may have
incorporated therein in advance a circuit for executing the
arithmetic process regarding the trained deep learning algorithm
60, or may have a programmable module built therein in addition to
such an incorporated circuit, for example. As the
parallel-processing processor 4833, GeForce, Quadro, TITAN, Jetson,
or the like of NVIDIA Corporation may be used, for example. In the
case of a Jetson series, Jetson Nano, Jetson Tx2, Jetson Xavier, or
Jetson AGX Xavier is used, for example.
[0273] The processor 4831 executes a calculation process regarding
control of the measurement unit 400, for example. The processor
4831 executes a calculation process regarding control signals
transmitted/received between the apparatus mechanism part 430, the
sample preparation part 440, and the specimen suction part 450. The
processor 4831 executes a calculation process regarding
transmission/reception of information with respect to the
processing unit 300, for example. The processor 4831 executes
processes regarding reading out of program data from the storage
4835, developing a program onto the RAM 4834, and
transmission/reception of data with respect to the RAM 4834. The
processes described above and executed by the processor 4831 are
required to be executed in a predetermined order, for example. For
example, when processes needed for control of the apparatus
mechanism part 430, the sample preparation part 440, and the
specimen suction part 450 are assumed to be A, B, and C, the
processes are required to be executed in the order of B, A, and C
in some cases. Since the processor 4831 often executes such
continuous processes that depend on an order, even when the number
of arithmetic units (each may be referred to as a "processor core"
a "core", or the like) is increased, the processing speed is not
necessarily increased.
[0274] Meanwhile, the parallel-processing processor 4833 executes a
large amount of regular calculation processes such as arithmetic
operations on matrix data including a large amount of elements, for
example. In the present embodiment, the parallel-processing
processor 4833 executes parallel processing in which at least a
part of processes of analyzing waveform data in accordance with the
deep learning algorithm 60 are parallelized. The deep learning
algorithm 60 includes a large amount of matrix operations, for
example. For example, the deep learning algorithm 60 may include at
least 100 matrix operations, or may include at least 1000 matrix
operations. The parallel-processing processor 4833 has a plurality
of arithmetic units, and the respective arithmetic units can
simultaneously execute matrix operations. That is, the
parallel-processing processor 4833 can execute, in parallel, matrix
operations by a plurality of respective arithmetic units, as
parallel processing. For example, a matrix operation included in
the deep learning algorithm 60 can be divided into a plurality of
arithmetic processes that are not order-dependent with each other.
The thus divided arithmetic processes can be executed in parallel
by a plurality of arithmetic units, respectively. These arithmetic
units may be each referred to as a "processor core", a "core", or
the like.
[0275] As a result of execution of such parallel processing, speed
up of arithmetic processing in the entirety of the measurement unit
400 can be realized. A process such as a matrix operation included
in the deep learning algorithm 60 may be referred to as "Single
Instruction Multiple Data (SIMD) processing", for example. The
parallel-processing processor 4833 is suitable for such an SIMD
operation, for example. Such a parallel-processing processor 4833
may be referred to as a vector processor.
[0276] As described above, the processor 4831 is suitable for
executing diverse and complicated processes. Meanwhile, the
parallel-processing processor 4833 is suitable for executing in
parallel a large amount of regular processes. Through parallel
execution of a large amount of regular processes, the TAT (Turn
Around Time) required for the calculation processes is
shortened.
[0277] The parallel processing to be executed by the
parallel-processing processor 4833 is not limited to matrix
operations. For example, when the parallel-processing processor
4833 executes a learning process in accordance with the deep
learning algorithm 60, differential operations or the like
regarding the learning process can be the target of the parallel
processing.
[0278] As for the number of arithmetic units of the processor 4831,
a dual core (the number of cores: 2), a quad core (the number of
cores: 4), or an octa core (the number of cores: 8) is adopted, for
example. Meanwhile, the parallel-processing processor 4833 has at
least ten arithmetic units (the number of cores: 10), and can
execute ten matrix operations in parallel, for example. The
parallel-processing processor 4833 that has several-ten arithmetic
units also exists. The parallel-processing processor 4833 that has,
for example, at least 100 arithmetic units (the number of cores:
100) and that can execute 100 matrix operations in parallel also
exists. The parallel-processing processor 4833 that has, for
example, several-hundred arithmetic units also exists. The
parallel-processing processor 4833 that has, for example, at least
1000 arithmetic units (the number of cores: 1000) and that can
execute 1000 matrix operations in parallel also exists. The
parallel-processing processor 4833 that has, for example, several
thousand arithmetic units also exists.
[0279] FIG. 47 shows a configuration example of the
parallel-processing processor 4833. The parallel-processing
processor 4833 includes a plurality of arithmetic units 4836 and a
RAM 4837. The respective arithmetic units 4836 execute arithmetic
processes of matrix data in parallel. The RAM 4837 stores data
regarding the arithmetic processes executed by the arithmetic units
4836. The RAM 4837 is a memory that has a capacity of at least 1
gigabyte. The RAM 4837 may be a memory that has a capacity of 2
gigabytes, 4 gigabytes, 6 gigabytes, 8 gigabytes, 10 gigabytes, or
more. Each arithmetic unit 4836 obtains data from the RAM 4837 and
executes an arithmetic process. The arithmetic unit 4836 may be
referred to as a "processor core", "core", or the like.
[0280] FIG. 48 to FIG. 50 each show an installation example of the
parallel-processing processor 4833 to the measurement unit 400.
FIG. 48 to FIG. 50 each show an example in which the
parallel-processing processor 4833 is installed, in the cell
analyzer 100, in the form of being incorporated in the measurement
unit 400. FIG. 48 and FIG. 49 each show an installation example in
which the processor 4831 and the parallel-processing processor 4833
are provided as separate bodies. As shown in FIG. 48, the processor
4831 is installed on a substrate 4838, for example. The
parallel-processing processor 4833 is installed on a graphic board
4830, and the graphic board 4830 is connected to the substrate 4838
via a connector 4839, for example. The processor 4831 is connected
to the parallel-processing processor 4833 via a bus 485. As shown
in FIG. 49, the parallel-processing processor 4833 may be directly
installed on the substrate 4838, and connected to the processor
4831 via the bus 485, for example. FIG. 50 shows an installation
example in which the processor 4831 and the parallel-processing
processor 4833 are integrally provided. As shown in FIG. 50, the
parallel-processing processor 4833 may be built in the processor
4831 installed on the substrate 4838, for example.
[0281] FIG. 51 shows another installation example of the
parallel-processing processor 4833 to the measurement unit 400.
FIG. 51 shows an example in which the parallel-processing processor
4833 is installed to the measurement unit 400 by means of an
external apparatus 4800 connected to the measurement unit 400. For
example, the parallel-processing processor 4833 is mounted on the
external apparatus 4800 being a USB (Universal Serial Bus) device,
and this USB device is connected to the bus 485 via an interface
part 487, whereby the parallel-processing processor 4833 is
installed to the cell analyzer 100. The USB device may be a small
device such as a USB dongle, for example. The interface part 487 is
a USB interface having a transfer rate of several hundred Mbps, for
example, and more preferably, is a USB interface having a transfer
rate of several Gbps to several ten Gbps or higher. As the external
apparatus 4800 having the parallel-processing processor 4833
mounted thereon, Neural Compute Stick 2 manufactured by Intel
Corporation may be used, for example.
[0282] A plurality of USB devices each having the
parallel-processing processor 4833 mounted thereon may be connected
to the interface part 487, whereby a plurality of
parallel-processing processors 4833 may be installed to the cell
analyzer 100. The parallel-processing processor 4833 mounted on one
USB device may have a smaller number of arithmetic units 4836 than
a GPU or the like in some cases. Therefore, when a plurality of USB
devices are connected to the measurement unit 400, scale-up of the
number of cores can be realized.
[0283] FIG. 52, FIG. 53A, FIG. 53B, and FIG. 54 each show an
outline of arithmetic processes executed by the parallel-processing
processor 4833 on the basis of control of the analysis software
4835a which operates on the processor 4831.
[0284] FIG. 52 shows a configuration example of the
parallel-processing processor 4833 which executes arithmetic
processes.
[0285] The parallel-processing processor 4833 includes a plurality
of the arithmetic units 4836 and the RAM 4837. The processor 4831,
which executes the analysis software 4835a, can issue an order to
the parallel-processing processor 4833 to cause the
parallel-processing processor 4833 to execute at least a part of
arithmetic processes necessary for analysis of waveform data
performed by the deep learning algorithm 60. The processor 4831
orders the parallel-processing processor 4833 to execute arithmetic
processes regarding waveform data analysis based on the deep
learning algorithm. All or at least a part of waveform data
corresponding to the signals detected by the FCM detection part 410
is stored in the RAM 4834. The data stored in the RAM 4834 is
transferred to the RAM 4837 of the parallel-processing processor
4833. The data stored in the RAM 4834 is transferred to the RAM
4837 by a DMA (Direct Memory Access) method, for example. The
plurality of the respective arithmetic units 4836 of the
parallel-processing processor 4833 execute in parallel arithmetic
processes with respect to the data stored in the RAM 4837. Each of
the plurality of the arithmetic units 4836 obtains necessary data
from the RAM 4837, to execute an arithmetic process. Data
corresponding to the arithmetic result is stored into the RAM 4837
of the parallel-processing processor 4833. The data corresponding
to the arithmetic result is transferred from the RAM 4837 to the
RAM 4834 by a DMA method, for example.
[0286] FIG. 53A and FIG. 53B each show an outline of a matrix
operation executed by the parallel-processing processor 4833.
[0287] Prior to analyzing waveform data in accordance with the deep
learning algorithm 60, calculation of the product of a matrix
(matrix operation) is executed. The parallel-processing processor
4833 executes in parallel a plurality of arithmetic processes
regarding the matrix operation, for example. FIG. 53A shows a
calculation formula of the product of a matrix. In the calculation
formula shown in FIG. 53A, a matrix c is obtained by a product of a
matrix a of n rows.times.n columns and a matrix b of n rows.times.n
columns. As shown as an example in FIG. 53A, the calculation
formula is described in a hierarchical loop syntax. FIG. 53B shows
an example of arithmetic processes executed in parallel in the
parallel-processing processor 4833. The calculation formula shown
as an example in FIG. 53A can be divided into n.times.n arithmetic
processes, n.times.n being the number of combinations of a loop
variable i for the first hierarchical level and a loop variable j
for the second hierarchical level, for example. Such divided
arithmetic processes are arithmetic processes that are not
dependent on each other, and thus can be executed in parallel.
[0288] FIG. 54 is a conceptual diagram showing that a plurality of
arithmetic processes shown as an example in FIG. 53B are executed
in parallel in the parallel-processing processor 4833.
[0289] As shown in FIG. 54, each of the plurality of arithmetic
processes is assigned to one of the plurality of the arithmetic
units 4836 of the parallel-processing processor 4833. The
respective arithmetic units 4836 execute in parallel the assigned
arithmetic processes. That is, the respective arithmetic units 4836
simultaneously execute the divided arithmetic processes.
[0290] Through the arithmetic operations shown as an example in
FIG. 53A, FIG. 53B, and FIG. 54 and performed by the
parallel-processing processor 4833, information regarding a
probability that a cell corresponding to the waveform data belongs
to each of a plurality of cell types is obtained. On the basis of
the results of the arithmetic operations, the processor 4831, which
executes the analysis software 4835a, performs analysis regarding
the cell types of the cell that corresponds to the waveform data.
The arithmetic results are stored in the RAM 4837 of the
parallel-processing processor 4833, to be transferred from the RAM
4837 to the RAM 4834. The processor 4831 transmits a result of
analysis performed on the basis of the arithmetic results stored in
the RAM 4834, to the processing unit 300 via the bus 485 and the
interface part 489.
[0291] The arithmetic operations of the probability that a cell
belongs to each of a plurality of cell types may be performed by a
processor different from the parallel-processing processor 4833.
For example, the arithmetic results by the parallel-processing
processor 4833 may be transferred from the RAM 4837 to the RAM
4834, and, on the basis of the arithmetic results read out from the
RAM 4834, the processor 4831 may perform arithmetic operations on
the information regarding the probability that a cell corresponding
to each piece of waveform data belongs to each of a plurality of
cell types. Alternatively, the arithmetic results by the
parallel-processing processor 4833 may be transferred from the RAM
4837 to the processing unit 300, and a processor installed in the
processing unit 300 may perform arithmetic operations on the
information regarding the probability that a cell corresponding to
each piece of waveform data belongs to each of a plurality of cell
types.
[0292] In the present embodiment, the processes shown in FIG. 53A,
FIG. 53B, and FIG. 54 are applied to an arithmetic process (also
referred to as a filtering process) regarding a convolution layer
in the deep learning algorithm 60, for example.
[0293] FIG. 55A and FIG. 55B each show an outline of an arithmetic
process regarding a convolution layer.
[0294] FIG. 55A shows an example of waveform data of forward
scattered light (FSC), as waveform data to be inputted to the deep
learning algorithm 60. As described with reference to FIG. 5, the
waveform data of the present embodiment is one-dimensional matrix
data. To put it more simply, the waveform data is a sequence in
which elements are arranged in a line. Here, for convenience of
description, the number of elements of the waveform data is assumed
to ben (n is an integer 1 or greater). FIG. 55A shows a plurality
of filters. Each filter is generated through a learning process of
the deep learning algorithm 60. Each of the plurality of filters is
one-dimensional matrix data indicating features of the waveform
data. Although each filter shown in FIG. 55A is matrix data of 1
row 3 columns, the number of columns is not limited to three. A
matrix operation is performed on the waveform data inputted to the
deep learning algorithm 60 and each filter, whereby features
corresponding to the cell type of the waveform data are calculated.
FIG. 55B shows an outline of a matrix operation performed on
waveform data and a filter. As shown in FIG. 55B, a matrix
operation is executed while each filter is shifted with respect to
the elements of the waveform data, one by one. Calculation of the
matrix operation is executed according to Formula 1 below.
[ Math .times. .times. 1 ] P = 0 L - 1 .times. q = 0 M - 1 .times.
h pq .times. x i + p , j + q ( Formula .times. .times. 1 )
##EQU00001##
[0295] In Formula 1, the suffixes of x are variables that indicate
the row number and the column number of the waveform data. The
suffixes of h are variables that indicate the row number and the
column number of the filter. In the example shown in FIGS. 55A,
55B, the waveform data is one-dimensional matrix data and the
filter is matrix data of 1 row.times.3 columns, and thus, L=1, M=3,
p=0, q=0, 1, 2, i=0, j=0, 1, . . . , n-1.
[0296] The parallel-processing processor 4833 executes in parallel
the matrix operation represented by Formula 1, by means of the
plurality of the respective arithmetic units 4836. On the basis of
the arithmetic processes executed by the parallel-processing
processor 4833, classification information regarding cell types of
each cell is generated. The generated classification information is
transmitted to the processing unit 300 which is used in generation
and display of a test result of the specimen based on the
classification information.
[0297] Next, with reference to FIG. 56, a process of cell
classification in step S13 in the flow chart of FIG. 10 is
described.
[0298] The process of cell classification in step S13 is a process
performed by the processor 4831 in accordance with operation of the
analysis software 4835a. The processor 4831 causes digital signals
taken into the RAM 4834 in step S13 to be transferred to the
parallel-processing processor 4833 (step S101). As shown in FIG.
52, the processor 4831 causes each digital signal to be transferred
from the RAM 4834 to the RAM 4837 through DMA transfer. The
processor 4831 controls the bus controller 4850, for example, to
cause the digital signal from the RAM 4834 to be DMA-transferred to
the RAM 4837.
[0299] The processor 4831 instructs the parallel-processing
processor 4833 to execute parallel processing onto the waveform
data included in the digital signal (step S102). The processor 4831
instructs execution of the parallel processing by calling a kernel
function of the parallel-processing processor 4833, for example.
The process executed by the parallel-processing processor 4833 will
be described later with reference to a flow chart shown as an
example in FIG. 57. The processor 4831 instructs the
parallel-processing processor 4833 to execute a matrix operation
regarding the deep learning algorithm 60, for example. The digital
signal is decomposed into a plurality of pieces of waveform data,
to be sequentially inputted to the deep learning algorithm 60. An
index corresponding to each cell and included in the digital signal
is not inputted to the deep learning algorithm 60. The waveform
data inputted to the deep learning algorithm 60 is subjected to
arithmetic operations performed by the parallel-processing
processor 4833.
[0300] The processor 4831 receives arithmetic results obtained
through execution by the parallel-processing processor 4833 (step
S103). The arithmetic results are DMA-transferred from the RAM 4837
to the RAM 4834 as shown in FIG. 52, for example.
[0301] On the basis of the arithmetic results by the
parallel-processing processor 4833, the processor 4831 generates an
analysis result of cell types of each measured cell (step
S104).
[0302] FIG. 57 shows an operation example of the arithmetic
processes of the parallel-processing processor 4833 executed on the
basis of an instruction from the processor 4831 according to
operation of the analysis software 4835a.
[0303] The processor 4831, which executes the analysis software
4835a, causes the parallel-processing processor 4833 to execute
assignment of arithmetic processes to the arithmetic units 4836
(step S110). The processor 4831 causes the parallel-processing
processor 4833 to execute assignment of arithmetic processes to the
arithmetic units 4836 by calling a kernel function of the
parallel-processing processor 4833. As shown in FIG. 54, for
example, a matrix operation regarding the deep learning algorithm
60 is divided into a plurality of arithmetic processes, and the
respective divided arithmetic processes are assigned to the
arithmetic units 4836. A plurality of pieces of waveform data are
sequentially inputted to the deep learning algorithm 60. A matrix
operation corresponding to each piece of waveform data is divided
into a plurality of arithmetic processes, to be assigned to the
arithmetic units 4836.
[0304] The arithmetic processes are processed in parallel by a
plurality of the arithmetic units 4836 (step S111). The arithmetic
processes are executed on the plurality of pieces of waveform
data.
[0305] Arithmetic results generated through the parallel processing
by the plurality of the arithmetic units 4836 are transferred from
the RAM 4837 to the RAM 4834 (step S112). For example, the
arithmetic results are DMA-transferred from the RAM 4837 to the RAM
4834.
Configuration Example 2
[0306] With reference to FIG. 58 and FIG. 59, another configuration
example of the cell analyzer 100 composed of the measurement unit
400 and the processing unit 300 is described. In the present
configuration example 2, a parallel-processing processor is
provided in the processing unit 300.
[0307] FIG. 58 shows a block diagram of the measurement unit 400 of
configuration example 2.
[0308] Except that the measurement unit 400 shown in FIG. 58 is not
provided with the A/D converter 482, the processor 4831, the RAM
4834, the storage 4835, and the parallel-processing processor 4833,
and is provided with a connection port 4201, the measurement unit
400 shown in FIG. 58 has configurations and functions similar to
those of the measurement unit 400 of configuration example 1
described with reference to FIG. 1 to FIG. 57 and in related
description thereof. A connection cable 4202 is connected to the
connection port 4201.
[0309] FIG. 59 shows a block diagram of the processing unit 300 of
configuration example 2.
[0310] As shown in FIG. 59, the processing unit 300 includes the
processor 3001, a parallel-processing processor 3002, the storage
3004, a RAM 3005, the interface part 3006, an A/D converter 3008,
the bus controller 4850, and an interface part 3009, and these are
connected to the bus 3003. That is, in the example in FIG. 59, the
parallel-processing processor 3002 is installed, in the cell
analyzer 100, in the form of being incorporated in the processing
unit 300.
[0311] The bus 3003 is a transmission line having a data transfer
rate of not less than several hundred MB/s, for example. The bus
3003 may be a transmission line having a data transfer rate of not
less than 1 GB/s. The bus 3003 performs data transfer in accordance
with the PCI-Express or PCI-X standard, for example. Configurations
of the processor 3001, the parallel-processing processor 3002, the
storage 3004, and the RAM 3005, and processes executed by these are
similar to the configurations and processes of the processor 4831,
the parallel-processing processor 4833, the storage 4835, and the
RAM 4834 described with reference to FIG. 47 to FIG. 57 above. The
A/D converter 3008 samples each analog signal outputted from the
measurement unit 400 as described above, and generates a digital
signal including waveform data of cells. The generation method of
the digital signal has been described above.
[0312] In the example in FIG. 58 and FIG. 59, the connection cable
4202 is provided with transmission paths the number of which
corresponds to the types of analog signals transmitted from the
measurement unit 400 to the processing unit 300, for example. For
example, the connection cable 4202 is implemented as a twisted-pair
cable, and has pairs of wires, the number of pairs corresponding to
the types of analog signals transmitted to the processing unit 300.
The transmission path from a connection port 3007 to the A/D
converter 3008 may also have wires of which the number corresponds
to the types of analog signals transmitted to the processing unit
300. In the transmission path from the connection port 3007 to the
A/D converter 3008, the analog signals are transmitted as
differential signals, for example.
[0313] As shown in FIG. 60, the processor 3001 and the
parallel-processing processor 3002 have configurations and
functions similar to those of the processor 4831 and the
parallel-processing processor 4833 described above. The
parallel-processing processor 3002 includes a plurality of
arithmetic units 3200 and a RAM 3201. Analysis software 3100 for
analyzing cell types of each measured cell is executed on the
processor 3001. The parallel-processing processor 3002 may not
necessarily be directly connected to the bus 3003. For example, the
parallel-processing processor 3002 may be mounted to a USB device,
and this USB device may be connected to the bus 3003 via an
interface part (not shown), whereby the parallel-processing
processor 3002 may be installed to the cell analyzer 100 as a part
of the processing unit 300. This USB device may be a small device
such as a USB dongle, for example.
[0314] As shown in FIG. 58 and FIG. 59, the processing unit 300 is
connected to the interface part 489 of the measurement unit 400 via
the interface part 3006. The processing unit 300 transmits control
signals of the apparatus mechanism part 430 and the sample
preparation part 440, to the measurement unit 400 via the interface
part 3006. The interface part 3006 is a USB interface, for
example.
[0315] In the example shown in FIG. 58 and FIG. 59, the processing
unit 300 is connected to the connection port 4201 of the
measurement unit 400 via the connection port 3007 connected to the
A/D converter 3008 and the connection cable 4202 connected to the
connection port 3007, in addition to the interface part 3006. The
connection port 4201 is connected to the analog processing part
420. Each analog signal outputted from the connection port 4201 to
the processing unit 300 is a signal obtained as a result of the
output of the FCM detection part 410 of the measurement unit 400
being processed by the analog processing part 420, as described
above. The analog processing part 420 performs a process including
noise removal on the analog signal inputted from the FCM detection
part 410. The analog signal having been processed by the analog
processing part 420 is transmitted to the processing unit 300 via
the connection port 4201 and the connection cable 4202. The
connection cable 4202 is formed to have a length of, for example, 1
meter or less in order to reduce noise during signal transmission.
The analog signal is transmitted as a differential signal to the
processing unit 300 via the connection cable 4202, for example. In
this case, the connection cable 4202 is preferably a twisted-pair
cable.
[0316] The processing unit 300 may include a plurality of the
connection ports 3007. The processing unit 300 may obtain analog
signals from a plurality of the measurement units 400 via a
plurality of the connection ports 3007.
[0317] Each analog signal transmitted from the measurement unit 400
via the connection cable 4202 is converted into a digital signal by
the A/D converter 3008 of the processing unit 300. As described
with reference to FIG. 5, for example, the A/D converter 3008
samples the analog signal transmitted at a predetermined sampling
rate (e.g., sampling at 1024 points at a 10 nanosecond interval,
sampling at 128 points at an 80 nanosecond interval, sampling at 64
points at a 160 nanosecond interval, or the like), and generates
waveform data for each cell. The waveform data is stored into the
storage 3004 or the RAM 3005 via the bus 3003. The waveform data is
DMA-transferred to the RAM 3005, for example. The processor 3001
and the parallel-processing processor 3002 execute arithmetic
processes on the waveform data stored in the storage 3004 or the
RAM 3005.
[0318] The analysis software 3100, which operates on the processor
3001, has functions similar to those of the analysis software 4835a
shown in FIG. 52. The processor 3001 executes the analysis software
3100, thereby generating classification information regarding cell
types of each measured cell, through operations similar to those
described with reference to FIG. 52, FIG. 53A, FIG. 53B, FIG. 54,
FIG. 56, and FIG. 57 and in related description thereof.
[0319] In the case of the cell analyzer 100 of configuration
example 2 shown in FIG. 58 and FIG. 59, in the flow chart shown in
FIG. 10, generation of waveform data and feature parameters in step
S12 and cell classification by the deep learning algorithm in step
S13 are performed in the processing unit 300. Step S14
(transmission of classification information) is omitted. The
processes in FIG. 56 and FIG. 57 are performed by the processor
3001 and the parallel-processing processor 3002 of the processing
unit 300.
Configuration Example 3
[0320] With reference to FIG. 61 and FIG. 62, another configuration
example of the cell analyzer 100 composed of the measurement unit
400 and the processing unit 300 is described. In the present
configuration example 3 as well, the parallel-processing processor
3002 is installed, in the cell analyzer 100, in the form of being
incorporated in the processing unit 300.
[0321] FIG. 61 shows a block diagram of the measurement unit 400 of
configuration example 3.
[0322] Except that the measurement unit 400 shown in FIG. 61 is not
provided with the processor 4831, the RAM 4834, the storage 4835,
and the parallel-processing processor 4833, and is provided with an
interface part 4851 and a transmission line 4852 for transmitting
the digital signals generated in the A/D converter 482 to the
processing unit 300, the measurement unit 400 shown in FIG. 61 has
configurations and functions similar to those of the measurement
unit 400 of configuration example 1 described with reference to
FIG. 2 and FIG. 47 and in related description thereof.
[0323] The interface part 4851 is an interface serving as a
dedicated line having a communication band of not less than 1
gigabit/second, for example. For example, the interface part 4851
is an interface according to Gigabit Ethernet, USB 3.0, or
Thunderbolt 3. When the interface part 4851 is according to Gigabit
Ethernet, the transmission line 4852 is a LAN cable, for example.
When the interface part 4851 is according to USB 3.0, the
transmission line 4852 is a USB cable according to USB 3.0. The
transmission line 4852 is a dedicated transmission line for
transmitting the digital signals between the measurement unit 400
and the processing unit 300, for example.
[0324] FIG. 62 shows a block diagram of the processing unit 300 of
configuration example 3.
[0325] Except that the processing unit 300 shown in FIG. 62 is not
provided with the A/D converter 3008 and the connection port 3007,
and is provided with an interface part 3010, the processing unit
300 shown in FIG. 62 has configurations and functions similar to
those of the processing unit 300 of configuration example 2
described with reference to FIG. 59 and in related description
thereof. The processing unit 300 may be connected to a plurality of
the measurement units 400 via a plurality of the interface parts
3010 and a plurality of the interface parts 3006.
[0326] The processor 3001 and the parallel-processing processor
3002 have configurations and functions similar to those of the
processor 3001 and the parallel-processing processor 3002 described
with reference to FIG. 60 and in related description thereof. In
FIG. 62, the parallel-processing processor 3002 may not necessarily
be directly connected to the bus 3003. For example, the
parallel-processing processor 3002 may be mounted to a USB device,
and this USB device may be connected to the bus 3003 via an
interface part (not shown). This USB device may be a small device
such as a USB dongle, for example.
[0327] The analysis software 3100, which operates on the processor
3001, has functions similar to those of the analysis software 4835a
shown in FIG. 52. The analysis software 3100 analyzes cell types of
each measured cell, through operations similar to those described
with reference to FIG. 52, FIG. 53A, FIG. 53B, FIG. 54, FIG. 56,
and FIG. 57 and in related description thereof.
[0328] In the case of the cell analyzer 100 of configuration
example 3 shown in FIG. 61 and FIG. 62, in the flow chart shown in
FIG. 10, step S13 (cell classification) is performed in the
processing unit 300. Step S14 (transmission of classification
information) is omitted. The processes in FIG. 56 and FIG. 57 are
performed by the processor 3001 and the parallel-processing
processor 3002 of the processing unit 300.
[0329] In the configuration in FIG. 61 and FIG. 62, analog signals
(forward scattered light signal, side scattered light signal, side
fluorescence signal) of each cell generated in the FCM detection
part 410 are converted into digital signals by the A/D converter
482 in the measurement unit 400. The digital signals are sent to
the processing unit 300 via the interface part 484, the bus 485,
the interface part 4851, and the transmission line 4852. The
transmission line 4852 is a dedicated transmission line for
transmitting the digital signals between the measurement unit 400
and the processing unit 300 as described above. For example, the
measurement unit 400 and the processing unit 300 are connected in a
one-to-one relationship via the transmission line 4852. In other
words, the transmission line 4852 is a transmission line that
provides no transmission of data related to an apparatus other than
components (e.g., the measurement unit 400 and the processing unit
300) forming the cell analyzer 100, for example. The transmission
line 4852 is a transmission line different from an intranet or the
internet, for example. Accordingly, even when digital signals
generated through A/D conversion in the measurement unit 400 are
transmitted to the processing unit 300, bottleneck in the
communication speed of transmission of the digital signals can be
avoided.
Configuration Example 4
[0330] With reference to FIG. 63, FIG. 64, FIG. 65, FIG. 66, and
FIG. 67, configuration example 4 of the cell analyzer 100 is
described.
[0331] In the present configuration example 4, as shown as an
example in FIG. 63, an analysis unit 600 is provided between the
measurement unit 400 and the processing unit 300. That is, in the
configuration in FIG. 63, FIG. 64, FIG. 65, FIG. 66, and FIG. 67,
the cell analyzer 100 is composed of the measurement unit 400, the
processing unit 300, and the analysis unit 600. The analysis unit
600 analyzes cell types of each measured cell. As described later,
in the present configuration example, a parallel-processing
processor 6002 is installed, in the cell analyzer 100, in the form
of being incorporated in the analysis unit 600.
[0332] FIG. 64 shows a configuration of the measurement unit 400 of
configuration example 4.
[0333] The measurement unit 400 shown as an example in FIG. 64 has
configurations and functions similar to those of the measurement
unit 400 of configuration example 3 described with reference to
FIG. 61 and in related description thereof. The analysis unit 600
is provided between the measurement unit 400 and the processing
unit 300. The analysis unit 600 may be connected to a plurality of
the measurement units 400. The analysis unit 600 may be connected
to a plurality of the processing units 300. The interface part 4851
is an interface having a communication band of not less than 1
gigabit/second, for example. For example, the interface part 4851
is an interface according to Gigabit Ethernet, USB 3.0, or
Thunderbolt 3. When the interface part 4851 is according to Gigabit
Ethernet, the transmission line 4852 is a LAN cable, for example.
When the interface part 4851 is according to USB 3.0, the
transmission line 4852 is a USB cable according to USB 3.0. The
transmission line 4852 is a dedicated transmission line for
transmitting the digital signals between the measurement unit 400
and the processing unit 300 as described above. For example, the
measurement unit 400 and the processing unit 300 are connected in a
one-to-one relationship via the transmission line 4852.
[0334] FIG. 65 shows a configuration example of the analysis unit
600.
[0335] The analysis unit 600 includes a processor 6001, the
parallel-processing processor 6002, a bus 6003, a storage 6004, a
RAM 6005, an interface part 6006, and an interface part 6007, for
example, and these are connected to the bus 6003. The bus 6003 is a
transmission line having a data transfer rate of not less than
several hundred MB/s, for example. The bus 6003 may be a
transmission line having a data transfer rate of not less than 1
GB/s. The bus 6003 performs data transfer in accordance with the
PCI-Express or PCI-X standard, for example. The analysis unit 600
may be connected to a plurality of the measurement units 400 via a
plurality of the interface parts 6006. When a plurality of the
measurement units 400 are provided, an analysis unit 600 may be
connected to each of the measurement units 400 (e.g., a plurality
of the measurement units 400 and a plurality of the analysis units
600 are connected in a one-to-one relationship).
[0336] As shown in FIG. 66, the processor 6001 and the
parallel-processing processor 6002 have configurations and
functions similar to those of the processor 4831 and the
parallel-processing processor 4833 described above. The
parallel-processing processor 6002 includes a plurality of
arithmetic units 6200 and a RAM 6201. Analysis software 6100, which
analyzes cell types of each measured cell, operates on the
processor 6001. The analysis software 6100, which operates on the
processor 6001, has functions similar to those of the analysis
software 4835a shown in FIG. 52. The analysis software 6100
analyzes cell types of each measured cell through operations
similar to those described with reference to FIG. 52, FIG. 53A,
FIG. 53B, FIG. 54, FIG. 55A, FIG. 55B, FIG. 56, and FIG. 57, and in
related description thereof. The analysis software 6100 transmits
classification information of each measured cell to the processing
unit 300 via the interface part 6007. The interface part 6007 is of
Ethernet (registered trademark) or USB, for example. The interface
part 6007 may be an interface capable of performing wireless
communication (e.g., WiFi (registered trademark), Bluetooth
(registered trademark)).
[0337] FIG. 67 shows a configuration of the processing unit 300 of
configuration example 4.
[0338] The processing unit 300 shown in FIG. 67 may not necessarily
include the parallel-processing processor 3002, unlike the
processing unit 300 shown in FIG. 59 and FIG. 62. In addition, the
analysis software 3100 shown in FIG. 59 and FIG. 62 may not
necessarily operate on the processor 3001 shown in FIG. 67. The
processing unit 300 receives analysis results of the analysis unit
600 via the interface part 3006. The interface part 3006 is of
Ethernet or USB, for example. The interface part 3006 may be an
interface capable of performing wireless communication (e.g., WiFi,
Bluetooth).
[0339] In the configuration in FIG. 64, FIG. 65, FIG. 66, and FIG.
67, analog signals (forward scattered light signal, side scattered
light signal, side fluorescence signal) of each cell generated in
the FCM detection part 410 are converted into digital signals by
the A/D converter 482 in the measurement unit 400. Waveform data is
sent to the analysis unit 600 via the interface part 484, the bus
485, the interface part 4851, and the transmission line 4852. The
interface part 4851 is a dedicated interface that connects the
measurement unit 400 and the analysis unit 600 to each other as
described above, and the interface part 4851 connects the
measurement unit 400 and the analysis unit 600 in a one-to-one
relationship. In other words, the transmission line 4852 is a
transmission line that provides no transmission of data related to
an apparatus other than components (e.g., the measurement unit 400
and the processing unit 300) forming the cell analyzer 100, for
example. The transmission line 4852 is a transmission line
different from an intranet or the internet, for example.
Accordingly, even when digital signals generated through A/D
conversion in the measurement unit 400 are transmitted to the
processing unit 300, bottleneck in the communication speed of
transmission of the digital signals can be avoided.
[0340] In the case of the cell analyzer 100 of configuration
example 4 shown in FIG. 64, FIG. 65, FIG. 66, and FIG. 67, in the
flow chart shown in FIG. 10, step S13 (cell classification) and
step S14 (transmission of classification information) are performed
in the analysis unit 600. That is, the digital signals generated in
step S13 are transmitted from the measurement unit 400 to the
analysis unit 600, and classification information is transmitted to
the processing unit 300 in step S14. The processes in FIG. 56 and
FIG. 57 are performed by the processor 6001 and the
parallel-processing processor 6002 of the analysis unit 600.
Configuration Example 5
[0341] With reference to FIG. 63, FIG. 67, and FIG. 68,
configuration example 5 of the cell analyzer 100 is described.
[0342] The cell analyzer 100 of this configuration example 5 also
includes the measurement unit 400, the processing unit 300, and the
analysis unit 600, as in the case of configuration example 4
described above. The measurement unit 400 of configuration example
5 shown in FIG. 68 has functions and configurations similar to
those of the measurement unit 400 described with reference to FIG.
58 and in related description thereof. The measurement unit 400 of
configuration example 5 shown in FIG. 68 is connected to the
analysis unit 600 via the connection cable 4202. For example, the
connection cable 4202 is implemented as a twisted-pair cable, and
has pairs of wires, the number of pairs corresponding to the types
of analog signals transmitted to the processing unit 300. The
connection cable 4202 is formed to have a length of, for example, 1
meter or less in order to reduce noise during signal transmission.
The measurement unit 400 transmits analog signals to the analysis
unit 600 via the connection cable 4202.
[0343] The analysis unit 600 shown in FIG. 69 has functions and
configurations similar to those of the analysis unit 600 described
with reference to FIG. 65 and in related description thereof. That
is, in the example in FIG. 69, the parallel-processing processor
6002 is installed, in the cell analyzer 100, in the form of being
incorporated in the analysis unit 600. The analysis unit 600 shown
in FIG. 69 further includes a connection port 6008 and an A/D
converter 6009. Analog signals transmitted from the measurement
unit 400 via the connection cable 4202 are inputted to the A/D
converter 6009 via the connection port 6008. The A/D converter 6009
converts the analog signals into digital signals through a process
similar to that performed by the A/D converter 482.
[0344] The analysis unit 600 may be connected to a plurality of the
measurement units 400 via a plurality of the connection ports 6008.
When a plurality of the measurement units 400 are provided, an
analysis unit 600 may be connected to each of the measurement units
400 (e.g., a plurality of the measurement units 400 and a plurality
of the analysis units 600 are connected in a one-to-one
relationship).
[0345] As shown in FIG. 66, the processor 6001 and the
parallel-processing processor 6002 have configurations and
functions similar to those of the processor 4831 and the
parallel-processing processor 4833 described above. The analysis
software 6100, which analyzes cell types of each measured cell,
operates on the processor 6001. The analysis software 6100, which
operates on the processor 6001, has functions similar to those of
the analysis software 4835a shown in FIG. 52. The analysis software
6100 analyzes cell types of each measured cell through operations
similar to those described with reference to FIG. 52, FIG. 53A,
FIG. 53B, FIG. 54, FIG. 55A, FIG. 55B, FIG. 56, and FIG. 57, and in
related description thereof. The analysis software 6100 transmits
analysis results of cell types of each measured cell to the
processing unit 300 via the interface part 6007. The interface part
6007 is of Ethernet or USB, for example. The interface part 6007
may be an interface capable of performing wireless communication
(e.g., WiFi, Bluetooth).
[0346] In the case of the cell analyzer 100 of configuration
example 5 shown in FIG. 68 and FIG. 69, in the flow chart shown in
FIG. 10, generation of waveform data and feature parameters in step
S12, step S13 (cell classification), and step S14 (transmission of
classification information) are performed in the analysis unit 600.
That is, generation of waveform data (step S12) is performed in the
A/D converter 6009 of the analysis unit 600, cell classification
based on digital signals (step S13) is performed by the processor
6001 and the parallel-processing processor 6002 of the analysis
unit 600, and classification information is transmitted from the
analysis unit 600 to the processing unit 300 (step S14). The
processes in FIG. 56 and FIG. 57 are performed by the processor
6001 and the parallel-processing processor 6002 of the analysis
unit 600.
9. NEURAL NETWORK STRUCTURE AND TRAINING OF DEEP LEARNING ALGORITHM
60
(Structure of Neural Network)
[0347] FIG. 70A shows, as an example, a structure of a convolution
neural network which realizes the deep learning algorithm 60.
[0348] The neural network includes an input layer 60a, an output
layer 60b, and a middle layer 60c between the input layer 60a and
the output layer 60b. The middle layer 60c is composed of a
plurality of layers. The number of layers forming the middle layer
60c may be, for example, not less than 5, preferably not less than
50, and more preferably not less than 100.
[0349] In the neural network, a plurality of nodes 89 arranged in a
layered manner are connected between the layers. Accordingly,
information is propagated only in one direction indicated by an
arrow D in the drawing, from the input-side input layer 60a to the
output-side the output layer 60b.
(Arithmetic Operation at Each Node)
[0350] FIG. 70B is a schematic diagram showing arithmetic
operations performed at each node.
[0351] Each node 89 receives a plurality of inputs, and calculates
one output (z). In the case of the example shown in FIG. 70B, the
node 89 receives four inputs. The total input (u) received by the
node 89 is expressed by Formula 2 below, for example. In the
present embodiment, one-dimensional matrix data is used as training
input data and analysis input data. Therefore, when variables of
the arithmetic expression correspond to two-dimensional matrix
data, a process of converting the variables so as to correspond to
one-dimensional matrix data is performed.
[Math 2]
u+w.sub.1x.sub.1+w.sub.2x.sub.2+w.sub.3x.sub.3+w.sub.4x.sub.4+b
(Formula 2)
[0352] Each input is multiplied by a different weight. In Formula
2, b is a value called bias. The output (z) of the node serves as
an output of a predetermined function f with respect to the total
input (u) expressed by Formula 2, and is expressed by Formula 3
below. The function f is called an activation function.
[Math 3]
z=f(u) (Formula 3)
[0353] FIG. 70C is a schematic diagram showing arithmetic
operations between nodes.
[0354] In the neural network, with respect to the total input (u)
of each node 89 expressed by Formula 2, nodes that output results
(z) each expressed by Formula 3 are arranged in a layered manner.
Outputs of the nodes of the previous layer serve as inputs to nodes
of the next layer. In the example shown in FIG. 70C, the outputs
from nodes 89a in the left layer in FIG. 70C serve as inputs to
nodes 89b in the right layer. Each node 89b receives outputs from
the nodes 89a. The connection between each node 89a and each node
89b is multiplied by a different weight. When the respective
outputs from the plurality of nodes 89a are defined as x1 to x4,
the inputs to the respective three nodes 89b are expressed by
Formula 4-1 to Formula 4-3 below.
[Math 4]
u.sub.1+w.sub.11x.sub.1+w.sub.12x.sub.2+w.sub.13x.sub.3+w.sub.14x.sub.4+-
b.sub.1 (Formula 4-1)
u.sub.2+w.sub.21x.sub.1+w.sub.22x.sub.2+w.sub.23x.sub.3+w.sub.24x.sub.4+-
b.sub.2 (Formula 4-2)
u.sub.3+w.sub.31x.sub.1+w.sub.32x.sub.2+w.sub.33x.sub.3+w.sub.34x.sub.4+-
b.sub.3 (Formula 4-3)
[0355] When Formula 4-1 to Formula 4-3 are generalized, Formula 4-4
below is obtained. Here, i=1, . . . , I, j=1, . . . , J. I is the
total number of inputs, and J is the total number of outputs.
[Math 5]
u.sub.j=.SIGMA..sub.i=1.sup.Iw.sub.jix.sub.i+b.sub.j (Formula
4-4)
[0356] When Formula 4-4 is applied to the activation function, an
output is obtained. The output is expressed by Formula 5 below.
[Math 6]
z.sub.j=f(u.sub.j) (j=1,2,3) (Formula 5)
(Activation Function)
[0357] In the cell type analysis method according to the
embodiment, a rectified linear unit function is used as the
activation function. The rectified linear unit function is
expressed by Formula 6 below.
[Math 7]
f(u)=max(u,0) (Formula 6)
[0358] Formula 6 is a function obtained by setting u=0 to the part
u<0 in the linear function with z=u. In the example shown in
FIG. 70C, using Formula 6, the output from the node of j=1 is
expressed by the formula below.
z.sub.1=max((w.sub.11x.sub.1+w.sub.12x.sub.2+w.sub.13x.sub.3+w.sub.14x.s-
ub.4+b.sub.1),0) [Math 8]
(Neural Network Learning)
[0359] If the function expressed by use of a neural network is
defined as y(x:w), the function y(x:w) varies when a parameter w of
the neural network is varied. Adjusting the function y(x:w) such
that the neural network selects a more suitable parameter w with
respect to the input x is referred to as neural network learning.
It is assumed that a plurality of pairs of an input and an output
of the function expressed by use of the neural network have been
provided. If a desirable output for an input x is defined as d, the
pairs of the input/output are given as {(x1,d1), (x2,d2), . . . ,
(xn,dn)}. The set of pairs each expressed as (x,d) is referred to
as training data.
[0360] The neural network learning means adjusting the weight w by
use of an error function such that, with respect to any
input/output pair (xn,dn), the output y(xn:w) of the neural network
when given an input xn becomes as close to the output do as much as
possible.
y(x.sub.n:w).apprxeq.d.sub.n [Math 9]
[0361] The error function is a measure for the closeness between
the training data and the function expressed by use of the neural
network. The error function is also called a loss function. An
error function E(w) used in the cell type analysis method according
to the embodiment is expressed by Formula 7 below. Formula 7 is
called cross entropy.
[Math 10]
E(w)=-.SIGMA..sub.n=1.sup.N.SIGMA..sub.k=1.sup.Kd.sub.nk log
y.sub.k(x.sub.n:w) (Formula 7)
[0362] A method for calculating the cross entropy of Formula 7 is
described. In the output layer of the neural network used in the
cell type analysis method according to the embodiment, i.e., in the
last layer of the neural network, an activation function for
classifying inputs x into a finite number of classes according to
the contents, is used. The activation function is called a softmax
function, and expressed by Formula 8 below. It is assumed that, in
the output layer 60b, the nodes are arranged by the same number as
the number of classes k. It is assumed that the total input u of
each node k (k=1, . . . , K) of an output layer L is given as
uk.sup.(L) from the outputs of the previous layer L-1. Accordingly,
the output of the k-th node in the output layer is expressed by
Formula 8 below.
[ Math .times. .times. 11 ] y k .ident. z k ( L ) = exp ( u k ( L )
) j = 1 K .times. exp ( u j ( L ) ) ( Formula .times. .times. 8 )
##EQU00002##
[0363] Formula 8 is the softmax function. The sum of output y 1, .
. . , yk determined by Formula 8 is always 1.
[0364] When each class is expressed as C1, . . . , Ck, output yk of
node k in the output layer L (i.e., uk.sup.(L)) represents the
probability that the given input x belongs to class Ck. The input x
is classified into a class in which the probability expressed by
Formula 9 below becomes largest.
[Math 12]
p(C.sub.k|x)=y.sub.k=z.sub.k.sup.(L) (Formula 9)
[0365] In the neural network learning, a function expressed by the
neural network is considered as a model of the posterior
probability of each class, the likelihood of the weight w with
respect to the training data is evaluated under such a probability
model, and a weight w that maximizes the likelihood is
selected.
[0366] It is assumed that target output dn by the softmax function
of Formula 8 is 1 only if the output is a correct class, and
otherwise, target output dn is 0. When the target output is
expressed in a vector format dn[dn1, . . . , dnk], if, for example,
the correct class of input xn is C3, only target output dn3 becomes
1, and the other target outputs become 0. When coding is performed
in this manner, the posterior distribution is expressed by Formula
10 below.
[Math 13]
p(d|x)=.PI..sub.k=1.sup.Kp(C.sub.k|x).sup.d.sup.k (Formula 10)
[0367] Likelihood L(w) of weight w with respect to the training
data {(xn,dn)}(n=1, . . . , N) is expressed by Formula 11 below.
When the logarithm of likelihood L(w) is taken and the sign is
inverted, the error function of Formula 7 is derived.
[ Math .times. .times. 14 ] L .function. ( w ) = n = 1 N .times. p
.function. ( d n | x n .times. : .times. w ) = n = 1 N .times. k =
1 K .times. p .function. ( C k | x n ) d nk = n = 1 N .times. k = 1
K .times. ( y k .function. ( x .times. : .times. w ) ) d nk (
Formula .times. .times. 11 ) ##EQU00003##
[0368] Learning means minimizing the error function E(w) calculated
on the basis of the training data, with respect to parameter w of
the neural network. In the cell type analysis method according to
the embodiment, the error function E(w) is expressed by Formula
7.
[0369] Minimizing the error function E(w) with respect to parameter
w has the same meaning as finding a local minimum point of the
function E(w). Parameter w is a weight of connection between nodes.
The local minimum point of weight w is obtained by iterative
calculation of repeatedly updating parameter w from an arbitrary
initial value used as a starting point. An example of such
calculation is the gradient descent method.
[0370] In the gradient descent method, a vector expressed by
Formula 12 below is used.
[ Math .times. .times. 15 ] .gradient. E = .differential. E
.differential. w = [ .differential. E .differential. w 1 , .times.
, .differential. E .differential. w M ] T ( Formula .times. .times.
12 ) ##EQU00004##
[0371] In the gradient descent method, a process of moving the
value of the current parameter w in the negative gradient direction
(i.e., -.gradient.E) is repeated many times. When the current
weight is AO and the weight after the moving is w.sup.(t+1), the
arithmetic operation according to the gradient descent method is
expressed by Formula 13 below. Value t means the number of times
the parameter w is moved.
[Math 16]
w.sup.(t+1)=w(t)-.di-elect cons..gradient.E (Formula 13)
[0372] The symbol shown in Formula 14 below used in Formula 13 is a
constant that determines the magnitude of the update amount of
parameter w, and is called a learning coefficient.
[Math 17]
.di-elect cons. (Formula 14)
[0373] As a result of repetition of the arithmetic operation
expressed by Formula 13, error function E(w.sup.(t)) decreases in
association with increase of value t, and parameter w reaches a
local minimum point.
[0374] It should be noted that the arithmetic operation according
to Formula 13 may be performed on all of the training data (n=1, .
. . , N) or may be performed on only a part of the training data.
The gradient descent method performed on only a part of the
training data is called a stochastic gradient descent method. In
the cell type analysis method according to the embodiment, the
stochastic gradient descent method is used.
[0375] An outline of training of the deep learning algorithm 60 is
described using FIG. 71 as an example. As described above, the deep
learning algorithm 60 is implemented as a neural network that
includes a multi-layered middle layer. As the training data,
waveform data of a cell of which the cell type has been identified
in advance is inputted to the input layer 60a of the neural
network. As the training data, probability data 78 is inputted to
the output layer 60b of the neural network. When the number of cell
types to be classified by the deep learning algorithm 60 is nine as
shown in FIG. 8, the number of nodes of the output layer 60b is
nine, and one of the cell types is assigned to each node. The
probability data 78 is a data group in which the probability of the
label value of the cell type that has been identified in advance
and that corresponds to the waveform data inputted to the input
layer 60a is set to 100% and the probabilities of the other cell
types are set to 0%. In this manner, the training data is inputted
to each of the input layer 60a and the output layer 60b, whereby
the neural network is trained.
10. EFFECTS OF EMBODIMENT
[0376] According to the embodiment described above, as
representatively shown in FIG. 20, for example, a result in which
each individual cell is classified into a plurality of cell types
is displayed. In a conventional cell classification method, a cell
that has features of a neutrophil and an immature granulocyte, for
example, is alternatively classified into either one of these
types. However, in the embodiment described above, with respect to
such a cell, the probabilities of belonging to neutrophil and
immature granulocyte are obtained by the deep learning algorithm,
and the values of the probabilities corresponding to the respective
cell types are displayed. Thus, the user can take an action of
conducting more detailed analysis such as, for example, performing
a retest or a visual inspection for detection of abnormal cells, on
a specimen containing a cell that has a probability of being an
abnormal cell.
11. REMARKS
[0377] The present disclosure includes following items 1-20.
[0378] Item 1: An analysis method for analyzing a specimen
containing cells, the analysis method comprising: applying light to
a measurement sample prepared from the specimen and detecting light
generated from cells; obtaining, with respect to each of a
plurality of cells contained in the specimen, feature data of the
cell on the basis of the detected light; analyzing the feature data
with use of an artificial intelligence algorithm, thereby
classifying each of the cells into a plurality of cell types; and
displaying information based on a result of the classifying.
[0379] Item 2: The analysis method of item 1, wherein the
classifying includes specifying, on the basis of the feature data,
a plurality of cell types to each of which each of the cells has a
possibility of morphologically belonging, among a plurality of cell
types determined in advance.
[0380] Item 3: The analysis method of item 2, wherein the plurality
of cell types include at least one normal cell and at least one
abnormal cell, and the classifying includes obtaining a probability
that each of the cells belongs to the abnormal cell.
[0381] Item 4: The analysis method of any one of items 1 to 3,
wherein the plurality of cell types include lymphocyte, monocyte,
eosinophil, neutrophil, basophil, and abnormal blood cell.
[0382] Item 5: The analysis method of item 3 or 4, wherein the
abnormal blood cell includes at least one of immature granulocyte,
blast, and abnormal lymphocyte.
[0383] Item 6: The analysis method of any one of items 1 to 5,
further comprising counting cells on the basis of the result of the
classifying and displaying a result of the counting of the
cells.
[0384] Item 7: The analysis method of any one of items 1 to 6,
wherein the analyzing of the data includes inputting the feature
data of one cell into a deep learning algorithm, and obtaining, as
an output from the deep learning algorithm, a result of classifying
the cell into a plurality of cell types.
[0385] Item 8: The analysis method of item 7, wherein the deep
learning algorithm has been trained using, as teaching data,
feature data of cells and information regarding types of the
cells.
[0386] Item 9: The analysis method of any one of items 1 to 8,
wherein the detecting includes detecting light generated as a
result of a cell passing through a flow cell to which light is
applied, and the obtaining of the feature data includes obtaining a
waveform signal that changes over time in accordance with the
detected light.
[0387] Item 10: The analysis method of item 9, wherein the feature
data is waveform data generated by sampling the waveform signal at
a plurality of time points.
[0388] Item 11: The analysis method of any one of items 1 to 10,
wherein the analyzing of the data is performed by using a processor
and a parallel-processing processor that operates under an order of
the processor.
[0389] Item 12: The analysis method of any one of items 1 to 11,
further comprising: displaying a graph on which a plurality of
cells contained in the specimen are plotted on the basis of the
feature data; and displaying, in accordance with one or a plurality
of cells having been selected on the graph, a result of classifying
the selected cell.
[0390] Item 13: The analysis method of any one of items 1 to 12,
wherein the analyzing of the data includes obtaining a probability
that one cell corresponds to each of a plurality of the cell
types.
[0391] Item 14: The analysis method of item 13, wherein the
displaying of the result of the classifying includes displaying the
probability.
[0392] Item 15: The analysis method of item 14, wherein the
displaying of the result of the classifying includes displaying a
cell type that has a highest probability.
[0393] Item 16: The analysis method of item 15, wherein displaying
of information based on the result of the classifying includes
displaying statistic information generated on the basis of the
probability.
[0394] Item 17: The analysis method of any one of items 1 to 16,
further comprising: displaying a graph on which a plurality of
cells contained in the specimen are plotted on the basis of the
feature data; and displaying, in accordance with one or a plurality
of cells having been selected on the graph, a result of classifying
the selected one or plurality of cells.
[0395] Item 18: The analysis method of any one of items 1 to 17,
further comprising: displaying a graph on which a plurality of
cells contained in the specimen are plotted on the basis of the
feature data; and displaying, in accordance with a plurality of
cells having been selected on the graph, statistic information
based on a result of classifying the selected plurality of
cells.
[0396] Item 19: The analysis method of any one of items 1 to 18,
further comprising: displaying a graph on which a plurality of
cells contained in the specimen are plotted on the basis of the
feature data; receiving an input of an extraction condition of a
cell; extracting a cell that satisfies the extraction condition on
the basis of classification information; and displaying, among the
cells displayed on the graph, the extracted cell so as to be
distinguished from other cells.
[0397] Item 20: An analyzer configured to analyze a specimen
containing cells, the analyzer comprising: a sample preparation
part configured to prepare a measurement sample from the specimen;
a detection part configured to apply light to the measurement
sample prepared by the sample preparation part, and to detect light
generated from cells; a signal processing part configured to
obtain, with respect to each of a plurality of cells contained in
the specimen, feature data of the cell on the basis of the detected
light; a display part; and at least one processor configured to
analyze the feature data with use of an artificial intelligence
algorithm, thereby classifying each of the cells into a plurality
of cell types, the at least one processor being configured to cause
the display part to display information based on a result of the
classifying.
* * * * *