Sample analysis equipment, sample analysis method, method

Abstract

This method enables screening of CML patients in the early stages of the disease, which was difficult with blood cell count tests. [Solution] The sample analyzer comprises a measurement unit that obtains optical information of cells by irradiating cells contained in the sample with multiple diffracted lights generated by incident light on a diffracting optical element, and an analysis unit that obtains information on leukemia cells contained in the sample by analyzing the optical information obtained by the measurement unit using an artificial intelligence algorithm.

Description

Technical Field

[0001] The present invention relates to a specimen analyzer and the like.

Background Art

[0002] Chronic myeloid leukemia (CML) is a leukemia caused by abnormalities in pluripotent hematopoietic stem cells and is characterized by the Philadelphia chromosome formed by t(9;22)(q34;q11). In CML, the BCR::ABL1 tyrosine kinase (TK) encoded by the BCR::ABL1 fusion gene on the Philadelphia chromosome is constantly activated, is involved in the proliferation of leukemia cells, and progresses through three disease stages. The three disease stages are the chronic phase with few subjective symptoms (3 to 5 years), the transition phase in which abnormal differentiation of granulocytes progresses (3 to 9 months), and the acute transformation phase in which undifferentiated blast cells increase and resemble acute leukemia (3 to 6 months), and ultimately it is fatal.

[0003] Non-Patent Document 1 discloses a CML diagnosis method based on an increase in white blood cells, particularly characteristic increases in neutrophils, eosinophils, and basophils, by a blood cell count test.

Prior Art Documents

Non-Patent Documents

[0004]

Non-Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] CML is a disease whose prognosis has dramatically improved with the advent of tyrosine kinase inhibitors (TKIs), and it is believed that early detection further improves treatment effectiveness. Although CML is characterized by an increase in white blood cells and basophils in the peripheral blood, in clinical practice, CML is suspected and diagnosed based on abnormalities in white blood cell count and basophil count. Therefore, it is difficult to screen patients in the early stages of the disease, before abnormalities in blood cell counts appear, using conventional blood cell count tests such as those disclosed in Non-Patent Document 1.

[0006] In view of these challenges, one objective of the present invention is to enable screening of CML patients in the early stages of the disease, which has been difficult with blood cell count tests. [Means for solving the problem]

[0007] According to a first aspect of the present invention, the sample analyzer comprises a measurement unit that acquires optical information of cells by irradiating cells contained in a sample with multiple diffracted lights generated by incident light on a diffracting optical element, and an analysis unit that acquires information on leukemia cells contained in the sample by analyzing the optical information obtained by the measurement unit using an artificial intelligence algorithm. According to a second aspect of the present invention, a sample analysis method involves irradiating cells contained in a sample with multiple diffracted lights generated by incident light on a diffracting optical element to acquire optical information of the cells, and then analyzing the obtained optical information using an artificial intelligence algorithm to acquire information on leukemia cells contained in the sample. According to a third aspect of the present invention, the method involves irradiating cells contained in a sample with multiple diffracted lights generated by incidenting light on a diffracting optical element to acquire optical information of the cells, the sample includes a first sample taken from a patient with chronic myeloid leukemia before the start of treatment and a second sample taken from a healthy person, a classification model for classifying leukemia cells is generated based on the optical information obtained from the first and second samples, an index regarding the classification performance of leukemia cells and normal cells by the generated classification model is obtained, and information regarding the efficacy of a chronic myeloid leukemia treatment drug for the patient is output based on the index. [Effects of the Invention]

[0008] According to the present invention, for example, information on leukemia cells contained in a sample can be obtained by performing analysis using an artificial intelligence algorithm based on the optical information of the cells contained in the sample. Leukemia cells are cancerous cells found in the blood of CML patients and are genetically Philadelphia chromosome (BCR::ABL gene) positive white blood cells. Morphologically, leukemia cells are indistinguishable from normal white blood cells and could not be detected by blood cell count tests or blood smear tests performed as screening tests. Therefore, conventionally, CML was diagnosed by performing genetic testing on patients who showed characteristic findings of CML, such as leukocytosis or basophilia, in blood cell count tests. Since leukocytosis and basophilia occur as a result of abnormal proliferation of leukemia cells in the bone marrow, CML is often already advanced by the time these symptoms appear. According to the present invention, it is possible to obtain information on leukemia cells present in the blood of CML patients, thus enabling early detection of CML. [Brief explanation of the drawing]

[0009] [Figure 1] This is a schematic front view showing the configuration of the sample analyzer according to the first embodiment. [Figure 2] A block diagram showing an example of the functional configuration of the GCM measurement unit according to the first embodiment. [Figure 3] A block diagram showing an example of the functional configuration of the sample preparation section of the GCM measurement unit according to the first embodiment. [Figure 4] A schematic diagram showing the configuration of the optical measurement section of the GCM measurement unit according to the first embodiment. [Figure 5] A schematic diagram showing the flow cell and diffraction illumination light of the GCM measurement unit according to the first embodiment. [Figure 6] A diagram schematically showing the distribution pattern of diffracted light contained in the diffracted illumination light according to the first embodiment. [Figure 7]Block diagram showing an example of the functional configuration of the control unit according to the first embodiment. [Figure 8] Schematic diagram showing the AI algorithm before and after training according to the first embodiment. [Figure 9] Diagram schematically showing the configuration of the cell analysis result screen according to the first embodiment. [Figure 10] Flowchart showing an example of the control process for measurement by the control unit according to the first embodiment. [Figure 11] Diagram showing the test information regarding the specimen of the verification experiment according to the first embodiment. [Figure 12] Diagram showing the test information regarding the specimen of the verification experiment according to the first embodiment. [Figure 13] Diagram showing the result of the verification experiment according to the first embodiment. [Figure 14] Diagram showing the result of the verification experiment according to the first embodiment. [Figure 15] Diagram showing the result of the verification experiment according to the first embodiment. [Figure 16] Front view schematically showing the configuration of the specimen analyzer according to the second embodiment. [Figure 17] Block diagram showing an example of the functional configuration of the FCM measurement unit according to the second embodiment. [Figure 18] Block diagram showing an example of the functional configuration of the sample preparation unit of the FCM measurement unit according to the second embodiment. [Figure 19] Diagram schematically showing the configuration of the optical measurement unit of the FCM measurement unit according to the second embodiment. [Figure 20] Diagram schematically showing the flow cell and the direct illumination light of the FCM measurement unit according to the second embodiment. [Figure 21] Flowchart showing an example of the control process for measurement by the control unit according to the second embodiment. [Figure 22] Flowchart showing an example of the FCM measurement process according to the second embodiment. [Figure 23] Diagram schematically showing the configuration of the cell analysis result screen according to the second embodiment. [Figure 24]A diagram schematically showing the configuration of a cell analysis result screen when the GCM measurement process according to the second embodiment is not performed. [Figure 25] A diagram schematically showing the configuration of a cell analysis result screen when the GCM measurement process according to the second embodiment is performed. [Figure 26] A front view schematically showing the configuration of a specimen analyzer according to the third embodiment. [Figure 27] A block diagram showing an example of the functional configuration of an integrated measurement unit according to the third embodiment. [Figure 28] A block diagram showing an example of the functional configuration of a sample preparation unit of an integrated measurement unit according to the third embodiment. [Figure 29] A diagram schematically showing the configuration of an optical measurement unit of an integrated measurement unit according to the third embodiment. [Figure 30] A flowchart showing an example of control processing related to measurement by a control unit according to the third embodiment. [Figure 31] A flowchart showing an example of control processing related to measurement by a control unit according to a modification of the third embodiment. [Figure 32] A block diagram showing an example of the functional configuration of a sample preparation unit of an integrated measurement unit according to a modification of the third embodiment. [Figure 33] A diagram showing a statistical comparison of PCR method values in a CML patient specimen during treatment according to the fourth embodiment. [Figure 34] A diagram showing an example of the results of a verification experiment according to the fourth embodiment. [Figure 35] A flowchart showing an example of control processing related to measurement by a control unit according to the fourth embodiment.

Embodiments for Carrying Out the Invention

[0010] <First Embodiment> The first embodiment relates to a basic example of obtaining information on leukemia cells contained in a specimen by irradiating a plurality of diffracted lights generated by incident light on a diffractive optical element onto cells contained in the specimen to obtain optical information of the cells, and analyzing the obtained optical information by an artificial intelligence algorithm. In this embodiment, as an example, information on leukemia cells contained in a sample is obtained based on the technique of ghost cytometry.

[0011] Figure 1 is a schematic front view showing the configuration of sample analyzer 1A, which is an example of a sample analyzer according to this embodiment. The sample analyzer 1A includes, for example, a ghost cytometry measurement unit (hereinafter referred to as the "GCM measurement unit") 20, a control unit 30, and a transport unit 40.

[0012] Sample analyzer 1A is, for example, a device that automatically analyzes samples. The sample may be blood collected from a subject, and the sample container 51 containing the sample may be transported while being held in a sample rack 50.

[0013] The laboratory technician, who is the operator of the sample analyzer 1A, for example, sets the sample container 51 containing the sample into the sample rack 50, and places the sample rack 50 in the rightmost area of ​​the transport unit 40. The transport unit 40 transports the sample rack 50 and positions it in front of the GCM measurement unit 20.

[0014] The GCM measurement unit 20 takes a sample container 51 from the sample rack 50, transfers it into the GCM measurement unit 20, and measures the sample in the sample container 51. Once the measurement of the sample in the sample container 51 is complete, the GCM measurement unit 20 returns the sample container 51 to its original position in the sample rack 50. Once all the necessary measurements have been completed for all the sample containers 51 on one sample rack 50, the transport unit 40 transports the sample rack 50 to the leftmost area of ​​the transport unit 40. The laboratory technician then retrieves the sample rack 50 from the leftmost area.

[0015] The GCM measurement unit 20 is configured to measure samples transported on the transport unit 40. The transport unit 40 is configured to automatically supply sample racks 50 containing samples to the GCM measurement unit 20. By having the transport unit 40 automatically supply samples to the GCM measurement unit 20, the workload of laboratory technicians required to transport samples to the GCM measurement unit 20 can be reduced.

[0016] The control unit 30 controls, for example, the GCM measurement unit 20 and the transport unit 40. Furthermore, the control unit 30 analyzes the measurement information obtained by the GCM measurement unit 20.

[0017] Figure 2 is a block diagram showing an example of the functional configuration of the GCM measurement unit 20.

[0018] The GCM measurement unit 20 includes, for example, a measurement control unit 21, a storage unit 22, a communication unit 23, a reading unit 24, a sample preparation unit 25, and a measurement unit 26.

[0019] The measurement control unit 21 is composed of, for example, an FPGA or a CPU. The memory unit 22 is composed of, for example, an HDD, an SSD, RAM, and ROM. The measurement control unit 21 performs various processes based on a program stored in the memory unit 22, for example, and controls each part of the GCM measurement unit 20.

[0020] The communication unit 23 is configured, for example, with a connection terminal based on the USB standard, and communicates with the control unit 30.

[0021] The reading unit 24 is configured, for example, as a barcode reader, and reads the barcode from the barcode label attached to the sample container 51 to obtain the sample ID.

[0022] The sample preparation unit 25, for example, aspirates a sample from the sample container 51 and mixes a reagent with the aspirated sample to prepare a measurement sample.

[0023] The measuring unit 26 includes, for example, an optical measuring unit 200, and the optical measuring unit 200 includes, for example, a fluid adjustment unit 220a.

[0024] The fluid adjustment unit 200a comprises, for example, a container for holding the sheath fluid, a syringe for transferring the measurement sample, and a pneumatic source (pump) for transferring the sheath fluid. The fluid adjustment unit 200a supplies the sheath fluid together with the measurement sample prepared in the sample preparation unit 25 to the flow cell 201 (see Figure 4) of the optical measurement unit 200, and adjusts the flow rate of the measurement sample flowing through the flow cell 201 per unit time. The optical measurement unit 200 measures the measurement sample supplied to the flow cell 201.

[0025] The optical measurement unit 200 includes an amplifier and an A / D conversion unit, performs signal processing on the detection signal acquired by the measurement, and outputs the processed measurement information to the measurement control unit 21. The measured information is, for example, a ghost motion imaging (GMI) waveform signal based on ghost cytometry (GCM). This can also be called ghost motion imaging waveform information (GMI waveform information), and it is an example of cellular optical information obtained by irradiating cells contained in a sample with multiple diffracted light rays generated when light is incident on a diffractive optical element. In the following, it will be simply referred to as a "waveform signal."

[0026] The measurement control unit 21 stores the waveform signal output from the measurement unit 26 in the storage unit 22. When the measurement of one sample is completed, the measurement control unit 21 transmits the waveform signal stored in the storage unit 22 to the control unit 30, associating it with the sample ID read by the reading unit 24.

[0027] Figure 3 is a block diagram showing an example of the functional configuration of the sample preparation unit 25 for preparing a measurement sample.

[0028] The sample preparation unit 25 includes, for example, a stirring unit 25a, a suction tube 25b, and a reaction chamber C30.

[0029] The stirring unit 25a is configured to, for example, grip the sample container 51 and shake the gripped sample container 51 to stir the sample inside the sample container 51. The suction tube 25b is, for example, a nozzle with a pointed lower end, and is configured to penetrate the lid of the sample container 51, which is made of an elastic material. The suction tube 25b aspirates the sample from inside the sample container 51 after stirring and dispenses the aspirated sample into the reaction chamber C30.

[0030] In reaction chamber C30, the sample is mixed with a hemolytic agent (an example of a reagent) for lysing red blood cells and a staining solution containing a fluorescent dye (a staining agent: an example of a reagent) for staining specific parts of the cells, thereby preparing the measurement sample. The hemolytic agent mixed in reaction chamber C30 is, for example, a WDF hemolytic agent. The staining solution mixed in reaction chamber C30 is, for example, WDF staining solution. The sample prepared in reaction chamber C30 is measured by the optical measurement unit 200.

[0031] The optical measurement unit 200 acquires a detection signal corresponding to the blood cells in the measurement sample, and performs signal processing on the acquired detection signal to obtain a waveform signal. The control unit 31 and calculation unit 32 of the control unit 30 analyze the waveform signal obtained from the measurement of the sample, classify it as CML cells or not, and obtain the number of each blood cell.

[0032] In this case, the waveform signal includes, for example, time-series data of forward scattered light corresponding to each cell, showing the change in intensity of forward scattered light received by the light-receiving unit 225 as each cell in the sample being measured flowing through the flow cell 201 (see Figures 4 and 5) passes through the illumination range R of the illumination light; time-series data of side scattered light corresponding to each cell, showing the change in intensity of side scattered light received by the light-receiving unit 233 as each cell in the sample being measured flowing through the flow cell 201 passes through the illumination range R of the illumination light; and time-series data of fluorescence corresponding to each cell, showing the change in fluorescence intensity received by the light-receiving unit 243 as each cell in the sample being measured flowing through the flow cell 201 passes through the illumination range R of the illumination light.

[0033] As will be described later, the control unit 31 and the calculation unit 32 of the control unit 30 input, for example, time-series data of forward-scattered light and side-scattered light into the trained AI algorithm 62 and analyze it.

[0034] The waveform signal is not limited to the time-series data described above, and may be any information that reflects the size, shape, internal structure, or nucleic acid content of each cell, obtained by irradiating each cell in the measurement sample with light in which multiple diffracted rays generated by the incident diffracting optical element 215 are distributed.

[0035] Furthermore, the hemolytic agent and staining solution mixed in reaction chamber C30 are not limited to the reagents mentioned above.

[0036] Furthermore, the mixing of the staining solution in reaction chamber C30 may be omitted. In other words, the reagent does not need to contain a staining agent, and the sample and reagent (hemolytic agent) may be mixed in reaction chamber C30 to prepare the measurement sample. Alternatively, a diluent may be mixed instead of the staining solution. In other words, the sample and reagents (hemolytic agent, diluent) may be mixed in reaction chamber C30 to prepare the measurement sample. In these cases, information on leukemia cells contained in the sample can be obtained without labeling (no fluorescent labeling required), and the fluorescence focusing optical system 205 and light receiving unit 243, which will be described later in Figure 4, can be omitted.

[0037] Figure 4 is a schematic diagram showing the configuration of the optical measurement unit 200. For convenience, the X, Y, and Z axes, which are orthogonal to each other, are indicated in Figure 4. The Z axis direction is the flow direction of the sample to be measured in the flow cell 201.

[0038] The optical measuring unit 200 includes, for example, a flow cell 201, a light source 211, an illumination optical system 202, a forward focusing optical system 203, a side focusing optical system 204, a fluorescence focusing optical system 205, and light receiving units 225, 233, and 243.

[0039] The illumination optical system 202 includes, for example, a collimator lens 212, cylindrical lenses 213 and 214, a diffractive optical element (DOE) 215, and a focusing lens 216. The illumination optical system 202 irradiates the flow channel 201a of the flow cell 201 with light from the light source 211. Hereinafter, the light emitted from the light source 211 and irradiated onto the channel 201a will be referred to as "diffractive illumination light." Diffractive illumination light is light in which multiple diffracted light beams generated by the diffractive optical element 215 are distributed. More specifically, diffractive illumination light is light having a structured illumination pattern (structured light illumination).

[0040] The forward-focusing optical system 203 includes, for example, a focusing lens 221, a beam stopper 222, a focusing lens 223, and an optical filter 224. The forward-focusing optical system 203 focuses the forward-scattered light generated from the blood cells onto the light-receiving unit 225 and blocks the diffracted illumination light that has passed through the flow cell 201 without irradiating the blood cells.

[0041] The lateral focusing optical system 204 includes, for example, a focusing lens 231 and an optical filter 232. The lateral focusing optical system 204 focuses the lateral scattered light generated from blood cells onto the light receiving unit 233.

[0042] The fluorescence focusing optical system 205 includes, for example, a focusing lens 241 and an optical filter 242. The fluorescence focusing optical system 205 focuses the fluorescence emitted from blood cells onto the light receiving unit 243.

[0043] The light source 211 is, for example, a semiconductor laser light source. The light source 211 emits light of a predetermined wavelength λ20 in the X-axis direction. The wavelength λ20 is, for example, 405 nm. The fast axis direction and slow axis direction of the light source 211 are parallel to the Y axis direction and the Z axis direction, respectively. The collimator lens 212 converts the light emitted from the light source 211 into parallel light.

[0044] Cylindrical lens 213 is a concave cylindrical lens, and cylindrical lens 214 is a convex cylindrical lens. Cylindrical lens 213 directs light emitted from light source 211 into cylindrical lens 214 by increasing the width in the Z-axis direction while keeping the width in the Y-axis direction unchanged, thereby creating a nearly circular shape. Cylindrical lens 214 converts the light emitted from light source 211 into parallel light.

[0045] The collimator lens 212 and cylindrical lenses 213 and 214 are positioned such that the light emitted from the light source 211 and transmitted through them forms a nearly perfect circle when viewed in the X-axis direction. As a result, the light incident on the diffractive optical element 215 forms a nearly perfect circle.

[0046] The configuration of the light source 211, collimator lens 212, and cylindrical lenses 213 and 214 is limited to any configuration that ensures the light incident on the diffractive optical element 215 forms a nearly circular shape, and other configurations are also acceptable. For example, the light source 211, collimator lens 212, and cylindrical lenses 213 and 214 may each be rotated 90 degrees with respect to the X-axis. In this case, the fast axis direction and slow axis direction of the light source 211 will be parallel to the Z axis direction and Y axis direction, respectively. Alternatively, a light source that emits nearly circular light may be used as the light source 211, and the collimator lens 212 and cylindrical lenses 213 and 214 may be omitted.

[0047] The diffractive optical element 215 has a diffraction pattern formed on it with a complex uneven shape, such as grooves and inclinations, to impart a diffraction effect to the incident light. The diffractive optical element 215 can be manufactured, for example, according to the description in U.S. Patent No. 9,477018, which is incorporated herein by reference. The diffractive optical element 215 diffracts the light incident in the X-axis direction from the cylindrical lens 214 side with respect to the X-axis direction, generating multiple diffracted lights with different propagation directions. These multiple diffracted lights are the spectrally separated portions of the incident light. The diffraction orders of the multiple diffracted lights are different from each other. The focusing lens 216 focuses the multiple diffracted lights generated from the diffractive optical element 215 onto the flow cell 201. The multiple diffracted lights with different propagation directions generated in the diffractive optical element 215 are focused onto the flow cell 201 to form diffracted illumination light.

[0048] The sample prepared in reaction chamber C30 (Figure 3) flows through flow cell 201. Diffractive illumination light irradiates cells in the sample flowing through flow cell 201, generating forward scattered light, side scattered light, and fluorescence from the irradiated cell sites. Forward scattered light occurs in the X-axis direction, while side scattered light and fluorescence occur in directions intersecting the X-axis direction (e.g., the Y-axis direction).

[0049] The focusing lens 221 focuses the forward scattered light generated from the cells and the diffracted illumination light that passes through the flow cell 201 without irradiating the cells. The beam stopper 222 allows forward scattered light generated from the cells to pass through, while blocking diffracted illumination light that has passed through the flow cell 201. The focusing lens 223 focuses the forward scattered light that has passed through the beam stopper 222 onto the light receiving unit 225. The optical filter 224 is configured to transmit only light with a wavelength of λ20. The light-receiving unit 225 receives forward-scattered light that has passed through the optical filter 224 and outputs a detection signal corresponding to the received light intensity. The light-receiving unit 225 is, for example, a photomultiplier tube (PMT).

[0050] The focusing lens 231 concentrates the lateral scattered light generated from the cells onto the light-receiving unit 233. The optical filter 232 is configured to transmit only light with a wavelength of λ20. The light-receiving unit 233 receives the laterally scattered light that has passed through the optical filter 232 and outputs a detection signal corresponding to the light-receiving intensity. The light-receiving unit 233 is, for example, a photomultiplier tube (PMT).

[0051] The focusing lens 241 concentrates the fluorescence emitted from the cells onto the light-receiving unit 243. The optical filter 242 is configured to transmit only light with a wavelength of λ21. The light-receiving unit 243 receives fluorescence transmitted through the optical filter 242 and outputs a detection signal corresponding to the light-receiving intensity. The light-receiving unit 243 is, for example, a photomultiplier tube (PMT).

[0052] Note that the light-receiving units 225, 233, and 243 are not limited to photomultiplier tubes (PMTs), but may also be photodiodes (PDs), for example.

[0053] Figure 5 schematically shows the flow cell 201 and the diffracted illumination light. Figure 5 includes the same X, Y, and Z axes as in Figure 4.

[0054] Inside the flow cell 201, a channel 201a through which the sample to be measured flows is formed parallel to the Z-axis. By flowing the sheath fluid along with the sample to be measured through channel 201a, the cells contained in the sample are enveloped in the sheath fluid and pass through the central region CE of channel 201a. The diffracted illumination light focused by the focusing lens 216 is irradiated onto a predetermined irradiation area R located in the central region CE of channel 201a. The flow rate of the sample to be measured per unit time is adjusted so that only one cell is positioned in the irradiation area R at a time, in other words, so that two or more cells do not pass through the irradiation area R simultaneously.

[0055] The lower part of Figure 5 shows an example of an image obtained by shining diffractive illumination light generated by the diffractive optical element 215 into a darkroom and capturing it with a camera. In the diffractive illumination light image in Figure 5, the black areas indicate areas without light, and the white dots indicate areas with light. The white dots in the diffractive illumination light image represent the diffractive light generated by the diffractive optical element 215. In the example shown in Figure 5, the diffractive light includes 0th order diffractive light, +1 to +300th order diffractive light, and -1 to -300th order diffractive light, and is shown in the diffractive illumination light image as a total of 601 white dots. The diffractive optical element 215 has a diffraction pattern (steps or grooves) formed on it so that each type of diffractive light is distributed as shown in such a diffractive illumination light image.

[0056] Figure 6 schematically shows the distribution pattern of diffracted light contained in diffracted illumination light.

[0057] Figure 6 shows an image of the illumination area R (see Figure 5) divided into a grid by multiple squares with sides the same length as the diameter of the diffracted light spot. The black squares indicate the region containing the diffracted light spot. The white squares indicate the region not containing the diffracted light spot. In Figure 6, cells passing through the illumination area R are shown as dashed circles. Since the diameter of the diffracted light spot in the image of diffracted illumination shown in Figure 6 is approximately 1 μm, in this case, the size of each square is 1 μm × 1 μm. The size of the cells is approximately 10 μm.

[0058] The size and length of the diffracted illumination light in the irradiation area R can be expressed in pixels, assuming each region of the grid is 1 pixel. In the example shown in Figure 6, the length of the diffracted illumination light in the flow direction of the sample (Z-axis direction) is px1 (pixel), the length of the diffracted illumination light in the short direction (Y-axis direction) is px2 (pixel), and the size of the diffracted illumination light is px1 × px2 (pixels).

[0059] The diffractive optical element 215 is designed so that the multiple diffracted light beams constituting the diffractive illumination light are distributed in a predetermined pattern. In this case, the predetermined pattern is a random pattern. The pattern may not have any repetition of a particular pattern, or it may have periodicity in which a particular pattern is repeated. However, it is preferable that at least one diffracted light beam is placed in a region that is 1 pixel long in the Y-axis direction and extends in the Z-axis direction, so that the entire cell region is exposed to the diffractive illumination light at least once.

[0060] When the sample to be measured flows into the flow path 201a of the flow cell 201 during measurement, the cells in the sample move in the Z-axis direction within the irradiation range R of the diffractive illumination light. At this time, the fluid adjustment unit 200a (see Figure 2) adjusts the flow rate per unit time to be approximately constant. When diffractive light contained in the diffractive illumination light is irradiated onto cells moving in the Z-axis direction, forward scattered light and side scattered light are generated from the part of the cell irradiated with the diffractive light. Also, when diffractive light is irradiated onto cells stained with a fluorescent dye, fluorescence is generated from the fluorescent dye irradiated with the diffractive light. The light receiving unit 225 (see Figure 4) receives forward scattered light generated by multiple diffractive lights irradiated onto the cells, the light receiving unit 233 (see Figure 4) receives side scattered light generated by multiple diffractive lights irradiated onto the cell's position, and the light receiving unit 243 (see Figure 4) receives fluorescence generated by multiple diffractive lights irradiated onto a predetermined location of the stained cell.

[0061] As cells move along the Z-axis, the number of diffracted light beams irradiated onto the cells changes, and the location of each diffracted light beam that hits the cell changes. As a result, the intensity of forward scattered light, side scattered light, and fluorescence emitted from the cells changes over time. Consequently, the detection signals from each of the light-receiving units 225, 233, and 243 also change over time. As will be described later, the calculation unit 32 (see Figure 7) classifies the cells using the AI ​​algorithm 62 based on the waveform signals obtained from these detection signals.

[0062] Figure 7 is a block diagram showing an example of the functional configuration of the control unit 30.

[0063] The control unit 30 includes, for example, a control unit 31, a calculation unit 32, a storage unit 33, a display unit 34, an input unit 35, and a communication unit 36.

[0064] The control unit 31 is configured, for example, by a CPU. The processing unit 32 is composed of, for example, a GPU (Graphics Processing Unit) or an NPU (Neural Network Processing Unit). The memory unit 33 is composed of, for example, an HDD, an SSD, RAM, and ROM. The control unit 31 executes a program stored in the memory unit 33 to control each part of the control unit 30, and also performs cell analysis based on the measurement information acquired by the GCM measurement unit 20.

[0065] The control unit 31 instructs the calculation unit 32 to perform analysis using the AI ​​algorithm 62, analyzes the waveform signal acquired by the optical measurement unit 200 of the GCM measurement unit 20, and obtains the GCM analysis result. In this case, the AI ​​algorithm 62 may consist of, for example, a statistical machine learning model such as SVM (Support Vector Machine) or a neural network model such as MLP (Multi-Layer Perceptron).

[0066] Alternatively, the control unit 31 and the calculation unit 32 may be configured as an integrated control calculation unit, and this control calculation unit may have the functions of both the control unit 31 and the calculation unit 32.

[0067] The display unit 34 is composed of, for example, a liquid crystal display. The input unit 35 is comprised of, for example, a keyboard, mouse, and pointing devices including a touch panel. The liquid crystal display of the display unit 34 and the touch panel of the input unit 35 may be integrated into a single unit. The communication unit 36 ​​is configured, for example, with a connection terminal based on the USB standard, and communicates with the GCM measurement unit 20 and the transport unit 40.

[0068] Figure 8 is a schematic diagram showing the AI ​​algorithm 61 before training and the AI ​​algorithm 62 after training.

[0069] In training the AI ​​algorithm 61, samples in which the majority of leukocytes in the peripheral blood are thought to have been replaced by CML cells based on genetic testing can be used to train the AI ​​algorithm. For example, samples in which the percentage of leukocytes positive for the BCR::ABL fusion gene is above a predetermined value can be used. For example, samples in which the Major BCR::ABL1 mRNA (%) is 80% or higher by PCR testing can be used, and more preferably 90% or higher. Peripheral blood collected from CML-positive patients who have not received TKI treatment (hereinafter referred to as "CML-positive untreated samples") are preferably used as such samples.

[0070] The control unit 31, for example, causes the sample rack 50 containing the sample container 51 containing the CML-positive untreated sample to be transported by the transport unit 40 and supplied to the GCM measurement unit 20. The control unit 31 then acquires, for example, a training waveform signal indicating CML positivity (hereinafter referred to as the "CML(+) waveform signal" for convenience) from the GCM measurement unit 20.

[0071] Furthermore, the control unit 31, for example, causes the sample rack 50 containing sample containers 51 from healthy individuals to be transported to the transport unit 40 and supplied to the GCM measurement unit 20. The control unit 31, for example, acquires a training waveform signal indicating CML negativity (hereinafter, for convenience, referred to as the "CML(-) waveform signal") from the GCM measurement unit 20. Samples from healthy individuals can be assumed to contain normal white blood cells (but not CML cells). For example, the "CML(+) waveform signal" and the "CML(-) waveform signal" may be a set of waveform signals (training dataset) consisting of waveform signals from multiple cells.

[0072] As shown in the upper part of Figure 8, the training waveform signals ("CML(+) waveform signal" and "CML(-) waveform signal") used to train the AI ​​algorithm 61 before training are, for example, information obtained by measuring specific cells (normal leukocytes and CML cells) with the GCM measurement unit 20.

[0073] For example, among the "CML(+) waveform signals" and "CML(-) waveform signals" of whole blood, only the waveform signals corresponding to white blood cells (e.g., all white blood cells, granulocytes, lymphocytes, monocytes) may be used as the "CML(+) waveform signal" or "CML(-) waveform signal."

[0074] In this case, for example, by using a waveform signal (GMI waveform signal), it is possible to determine whether the cells being analyzed are white blood cells, and what type of white blood cell they are. The types of white blood cells can be, for example, based on the white blood cell differential, such as granulocytes (neutrophils, eosinophils, basophils), lymphocytes, and monocytes. In this case, for example, the method disclosed in "Pooled CRISPR screening of high-content cellular phenotypes using ghost cytometry" Tsubouchi A, An Y, Kawamura Y [..] Ota S. Cell Rep Methods 2024-03-25 (https: / / doi.org / 10.1016 / j.crmeth.2024.100737) may be applied as a method for determination.

[0075] The AI ​​algorithm 61 is composed of, for example, a neural network including multiple hidden layers. In this case, the neural network may be, for example, a convolutional neural network (CNN) having convolutional layers. The AI ​​algorithm 61 has an input layer, an output layer, and hidden layers. The AI ​​algorithm 61 is trained when a data set of waveform signals obtained by sampling an analog detection signal obtained from a single cell at a predetermined sampling period ("CML(+) waveform signal" or "CML(-) waveform signal") is input to the input layer, and a label value corresponding to the cell type (normal leukocyte or CML cell) is input to the output layer. By repeatedly performing this training in advance, the trained AI algorithm 62 is generated.

[0076] As shown in the lower part of Figure 8, for example, the trained AI algorithm 62 also has an input layer, an output layer, and an intermediate layer. The waveform signal acquired based on the subject's sample is input to the input layer. This allows the output layer to output classification information regarding the cell type corresponding to the waveform signal (whether or not it is a CML cell).

[0077] For example, if the output layer has one node, the label value corresponding to CML cells may be set to "1" and the label value corresponding to normal white blood cells may be set to "0". Furthermore, for example, if the output layer has two nodes, the learning process may be configured so that when a "CML(+) waveform signal" is input to the input layer, the node corresponding to the CML cell outputs "1" and the node corresponding to the normal white blood cell outputs "0". Similarly, the learning process may be configured so that when a "CML(-) waveform signal" is input to the input layer, the node corresponding to the CML cell outputs "0" and the node corresponding to the normal white blood cell outputs "1".

[0078] The classification information may include the probability that the target cell is a CML cell. For example, if the output layer has one node, the classification information is the output value of the output layer and takes a value in the range of "0 to 1". Also, for example, if the output layer has two nodes, the classification information is the output value of the node corresponding to the CML cell and takes a value in the range of "0 to 1". The classification information indicates that, for example, the closer the value is to "1", the higher the probability that it is a CML cell. Furthermore, based on the classification information, the control unit 31 performs, for example, a threshold determination to determine whether a target cell is a CML cell in the waveform signal corresponding to a particular cell in the subject's sample.

[0079] Training of AI algorithm 61 and classification using AI algorithm 62 are performed, for example, by inputting a data set of waveform signals obtained for each cell by one or more of the three light-receiving units 225, 233, and 243 (see Figure 4) as input data into the input layer. Specifically, if a waveform signal obtained from any one of the light-receiving units 225, 233, and 243 is used to obtain n data sets from the detection signals obtained for each individual cell, the number of detection signal data points input to AI algorithm 61 and AI algorithm 62 corresponding to one cell will be n, and the number of nodes in the input layer will also be n. Alternatively, if a data set of three detection signals obtained from each of the three light-receiving units 225, 233, and 243 is input as input data into the input layer, 3n data sets will be obtained from the three detection signals, and the number of nodes in the input layer will also be 3n.

[0080] In this embodiment, the arithmetic unit 32 performs cell classification using the AI ​​algorithm 62, but the control unit 31 may also perform cell classification using the AI ​​algorithm 62. However, the arithmetic unit 32, which consists of a GPU or the like, can perform cell classification using the AI ​​algorithm 62 more quickly. Alternatively, the aforementioned control arithmetic unit may also perform cell classification using the AI ​​algorithm 62.

[0081] Furthermore, AI algorithm 62 is not limited to neural network models. For example, it may be a machine learning model (classification model) such as SVM. In this case, during the learning phase, the model may be trained by using "CML(+) waveform signal" and "CML(-) waveform signal" as explanatory variables and defining the target variable so that both can be classified.

[0082] The control unit 30 repeatedly determines whether a target cell is a CML cell based on the waveform signal corresponding to a single cell in the subject's sample, and accumulates this determination to calculate the percentage (e.g., "%) of multiple blood cells contained in the sample that are CML cells.

[0083] This waveform signal (GMI waveform signal) may correspond to each cell in the whole blood. That is, it may be possible to determine whether all cells are CML cells or not, regardless of whether they are white blood cells or not, and the number of CML cells in the total number of cells or the ratio of CML cells to the total number of cells may be calculated. Furthermore, the waveform signal may correspond to the waveform signal of cells identified as white blood cells. That is, it is possible to determine whether or not only the cells identified as white blood cells are CML cells, and to calculate the number of CML cells in the total number of white blood cells or the ratio of the number of CML cells to the total number of white blood cells. Furthermore, the waveform signal may correspond to a specific type of white blood cell (granulocyte, lymphocyte, or monocyte). In other words, it is possible to determine whether or not a cell is a CML cell only for cells that have been identified as a specific type, and to calculate the number of CML cells among the cells of a specific type, or the ratio of the number of CML cells to the total number of cells of a specific type.

[0084] Figure 9 schematically shows the configuration of the cell analysis result screen 600 that is displayed on the display unit 34 during the analysis result output processing (step S150 in Figure 10), which will be described later. The cell analysis result screen 600 can also be called the output screen for CML cell analysis result information based on classification information.

[0085] The cell analysis results screen 600 consists, for example, of a count value display area 610 and a CML cell content display area 620. In this example, the count display area 610 is configured to display, for example, the total white blood cell count analyzed in the sample, the number of cells identified as CML cells among the total white blood cells (CML cell count), and the number of cells identified as not being CML cells (normal cells) among the total white blood cells (normal cell count).

[0086] The CML cell content display area 620 is configured to display, for example, the percentage of CML cells contained in the sample (CML cell content) in units of "%", based on the number of CML cells and the total number of white blood cells.

[0087] Furthermore, in the CML cell content display area 620, if the CML cell content exceeds a predetermined percentage (e.g., "20%"), or is above a predetermined percentage, a flag information 621 may be displayed to indicate that a more sensitive genetic test, such as Major BCR::ABL1 mRNA quantitative PCR or FISH, is recommended for the subject from whom the sample was collected.

[0088] Next, with reference to Figure 10, the measurement process of the sample analyzer 1A will be described. Figure 10 is a flowchart showing an example of the control process flow related to measurement by the control unit 30.

[0089] First, in step S110, the control unit 31 of the control unit 30 controls the GCM measurement unit 20 to perform sample preparation by the sample preparation unit 25 (see Figures 2 and 3). As a result, the measurement sample is prepared in the reaction chamber C30.

[0090] In step S120, the control unit 31 controls the GCM measurement unit 20 so that the optical measurement unit 200 performs the measurement. As a result, the optical measurement unit 200 irradiates the cells in the sample contained in the sample flowing through the flow cell 201 with diffracted illumination light, as shown in Figure 5, and acquires a waveform signal. The control unit 31 then acquires the waveform signal acquired by the measurement unit 26 from the GCM measurement unit 20.

[0091] In step S130, the control unit 31 and the calculation unit 32 input the waveform signal acquired in step S120 into the trained AI algorithm 62 for analysis.

[0092] In step S140, the control unit 31 and the calculation unit 32 generate CML cell analysis information based on the analysis in step S130. The CML cell analysis information may include counts of blood cells, CML cells, and non-CML cells, and / or CML cell content, obtained by analyzing waveform signals based on measurements from the optical measurement unit 200. The CML cell analysis information may be, for example, information based on classification information, and may also be called GML analysis information.

[0093] For example, when a medical technologist inputs a display instruction via the input unit 35, in step S150, the control unit 31 displays the cell analysis results screen 300, which includes the CML cell analysis information generated in step S140, on the display unit 34 (an example of output).

[0094] Furthermore, the "output" of information, including analysis results, may include not only the display of information on the device itself (display output), but also, for example, the output of information to other functional units of the device (internal output), or the output (external output) or transmission (external transmission) of information to devices other than the device itself (external devices). Sound output (including voice output) may also be included.

[0095] In step S160, for example, based on the input operation of the laboratory technician via the input unit 35, the control unit 31 determines whether or not to terminate the process. If it determines to continue the process (S160: NO), the control unit 31 controls, for example, to take another sample container 51 from the sample rack 50 and transfer it into the GCM measurement unit 20, and prepares to measure the sample in the new sample container 51. Then, for example, the process returns to step S110. If it is determined that the process should be terminated (S160: YES), the control unit 31, for example, controls the return of the sample container 51 from the GCM measurement unit 20 to its original position in the sample rack 50, thereby terminating the process.

[0096] <Verification experiment> To confirm the effectiveness of this method, we assumed that CML patients undergoing TKI treatment, which is used as standard treatment for CML, were CML patients in the early stages of the disease, and conducted verification experiments to see if it was possible to distinguish between samples from CML patients undergoing TKI treatment and samples from healthy individuals. In the verification experiments, as will be described later with reference to Figure 29, we used an optical measurement unit 400 that can acquire flow cytometry information in addition to GMI waveform information from a single cell.

[0097] Figure 11 compares blood test values ​​from samples of CML patients before TKI treatment (n=6) and samples of CML patients during TKI treatment (n=11). The bottom row shows the values ​​of gene testing by PCR. In CML patients before TKI treatment, the percentage of CML cells was over 90% by BCR::ABL1 mRNA (IS%), indicating that almost all white blood cells in the blood of CML patients are CML cells. On the other hand, in CML patients during TKI treatment, the median percentage of CML cells in white blood cells was about 50%. However, among the complete blood count values ​​of CML patients during TKI treatment, the median white blood cell count (WBC), which is used as an indicator for diagnosing CML, was 49.9 (normal range: 36.0~88.0 × 10^2 / μL), and the median basophil percentage (BASO%) was 1.0 (normal range: 0~1.0%), both of which are within the normal range.

[0098] Figure 12 is a graph showing whether an abnormal flag was raised (Flag positive) or not (Flag negative) for each sample when a complete blood count (CBC) was performed using an automated hematology analyzer for the samples shown in Figure 11. The "Untreated" bar graph corresponds to the group of samples (n=6) from CML patients before TKI treatment in Figure 11, and the "Treated" bar graph corresponds to the group of samples (n=11) from CML patients undergoing TKI treatment in Figure 11. This graph shows that abnormal flags indicating characteristic findings suggestive of CML, such as increased WBC (white blood cell count) and increased basophils (Baso), are almost never raised in CML patients undergoing TKI treatment.

[0099] The results in Figures 11 and 12 show that even in samples where CML cells account for more than 50% of the white blood cells in the blood, it is difficult to suspect CML using a complete blood count with an automated hematologist. Therefore, in this validation experiment, we investigated whether this method could be used to differentiate CML patients undergoing TKI treatment, assuming they were early-stage CML patients for whom screening by abnormal flags in complete blood counts with automated hematologists is difficult.

[0100] Figure 13 shows an example of training (learning) AI algorithm 61 using waveform signals from all white blood cells of TKI-naive CML-positive patients, referred to as "CML(+) waveform signals."

[0101] <Learning Phase> The left side of the figure shows a scattergram of a sample from a CML-positive patient who had not received TKI treatment. As mentioned above, this validation experiment used an optical system capable of acquiring flow cytometry signals in addition to cellular waveform signals. In the scattergram, the horizontal axis represents the peak value (peak intensity) of forward scattered light obtained as a flow cytometry signal, and the vertical axis represents the peak value (peak intensity) of side scattered light obtained as a flow cytometry signal. In the scattergram, cells within the dashed line were identified as leukocytes. The AI ​​algorithm 61 was trained using the waveform signal of the identified leukocytes as the "CML(+) waveform signal". The AI ​​algorithm 61 was also trained using the waveform signal of leukocytes contained in a sample from a healthy individual as the "CML(-) waveform signal". In this way, the trained AI algorithm 62 was obtained.

[0102] <Inference Phase> Cellular waveform signals and flow cytometry signals were obtained from samples from healthy individuals (n=5) and from samples assumed to be from CML patients in the early stages of the disease, i.e., CML patients undergoing TKI treatment (n=11). Leukocytes were identified based on the peak values ​​of forward and side scattering light. The waveform signals of the identified leukocytes were input into a trained AI algorithm 62 to calculate the CML cell content.

[0103] The right side of the scattergram shows the CML cell content of samples from healthy individuals (labeled "HC" in the figure) and samples from hypothetical early-stage CML patients (labeled "CML" in the figure). This graph shows a significant difference in the CML cell content among the identified white blood cells between HC and CML, suggesting the possibility of early diagnosis of CML patients in the early stages of the disease. To the right of that is the ROC curve for AI algorithm 61 during training. From this ROC curve, the AUC value is "0.98", indicating that the trained AI algorithm 62 is an effective classification model. The right side of the figure shows the correlation analysis results between the CML cell content in leukocytes identified by this method and the Philadelphia chromosome BCR::ABL1 gene content detected by PCR. From this figure, it can be seen that there is a positive correlation between this method and the PCR results, demonstrating the effectiveness of this method.

[0104] Figure 14 shows a verification example of training (learning) AI algorithm 61 using the waveform signal of granulocytes among the white blood cells of TKI-naive CML-positive patients as the "CML(+) waveform signal". Similar to the verification example in Figure 13, in order to match the population of cells to be analyzed in the learning phase and the inference phase, the waveform signal of granulocytes was input to AI algorithm 62 in the inference phase. Similarly, the waveform signal of granulocytes contained in a sample from a healthy individual was used as the "CML(-) waveform signal". The interpretation of each figure is the same as in Figure 13. These results show that when AI algorithm 61 is trained using waveform signals from granulocytes, there is no superposition in the CML cell content of the differentiated leukocytes between HC and CML, and a significant difference is observed. Furthermore, the AUC value is "1," indicating that the trained AI algorithm 62 is a very effective classification model. In other words, these results strongly suggest that training AI algorithm 61 specifically on granulocytes could lead to the early diagnosis of CML patients in the early stages of the disease.

[0105] Figure 15 shows a verification example of training (learning) AI algorithm 61 using the waveform signal of lymphocytes from leukocytes of CML-positive patients who have not received TKI treatment, as the "CML(+) waveform signal". Similar to the verification example in Figure 13, in order to match the population of cells to be analyzed in the learning phase and the inference phase, the waveform signal of lymphocytes was input to AI algorithm 62 in the inference phase. Similarly, the waveform signal of lymphocytes contained in a sample from a healthy individual was used as the "CML(-) waveform signal". The interpretation of each figure is the same as in Figure 13. When training AI algorithm 61 specifically on lymphocytes, the AUC value decreases to "0.92". However, considering the characteristic of lymphocytes to withstand long-term storage, the results suggest that it may be helpful in the early diagnosis of CML even in samples collected some time ago.

[0106] The results of the verification experiments shown in Figures 13-15 suggest that this method can identify CML cells in the blood and accurately distinguish early-stage CML patients, which are difficult to screen using blood cell counts. The reason why GCM can identify CML cells is thought to be that GCM can examine the morphological characteristics unique to CML cells in more detail than flow cytometry signals used in blood cell counts. The inventors have revealed that the BCR::ABL fusion gene, which is characteristic of CML cells, alters the morphology of mitochondria in cells and causes excessive mitochondrial fragmentation. It is thought that GCM can distinguish between normal cells and CML cells by comprehensively capturing subtle morphological changes caused by cell tumorigenesis, such as morphological changes of mitochondria within cells, which are difficult to capture with conventional flow cytometry, thus leading to the results shown above.

[0107] In the verification experiment described above, an optical system capable of acquiring flow cytometry signals in addition to GMI waveform signals was used to identify specific types of cells (e.g., leukocytes, granulocytes, lymphocytes). However, the configuration for acquiring flow cytometry signals is not necessarily required, and this method can be implemented using only GMI waveform signals. For example, in addition to the AI ​​algorithm trained using CML waveform signals, an AI algorithm (type identification AI) that determines whether a cell is a specific type of cell (e.g., leukocyte) based on the waveform signal may be used. In this case, for example, in the learning phase, the AI ​​algorithm 61 can be trained by inputting the waveform signals of cells identified as leukocytes by the type identification AI from among the cells of a TKI-untreated CML-positive patient as CML(+) waveform signals. In the inference phase, the discrimination result can be obtained by inputting the waveform signals of cells identified as leukocytes by the type identification AI from among the cells of the patient sample into the trained AI algorithm 62. Alternatively, learning and inference may be performed using the waveform signals of all particles from which waveform signals are obtained among the particles contained in the sample, without performing gating as shown in the scattergrams in Figures 13-15. As can be seen from the scattergram in Figure 13, since most of the particles in the blood lysed with a hemolytic agent are white blood cells, it is possible to analyze only white blood cells without performing gating.

[0108] <Effects of the First Embodiment> According to this embodiment, in a sample analyzer (e.g., sample analyzer 1A), the measurement unit (e.g., GCM measurement unit 20) irradiates cells in a sample (e.g., blood) with multiple diffracted lights generated by the incident light on a diffracting optical element (e.g., diffracting optical element 215) to acquire optical information (e.g., GMI waveform signal (waveform information)). Then, the analysis unit (e.g., control unit 30) analyzes the optical information obtained by the measurement unit using an artificial intelligence algorithm (e.g., trained AI algorithm 62) to acquire information about leukemia cells contained in the sample (e.g., whether or not they are leukemia cells, the percentage of leukemia cells, the number of leukemia cells). This allows for easy and accurate acquisition of information about leukemia cells contained in a sample by analyzing the optical information obtained by the measurement unit using an artificial intelligence algorithm. In other words, it may be possible to detect CML with high accuracy through a simple blood test in the early stages of CML development, potentially leading to early diagnosis of CML.

[0109] Furthermore, according to this embodiment, the analysis unit acquires information on leukemia cells, specifically the percentage of leukemia cells among white blood cells (for example, the CML cell content). This allows for the easy and accurate acquisition of the percentage of leukemia cells among the white blood cells contained in a sample by analyzing the optical information obtained by the measurement unit using an artificial intelligence algorithm.

[0110] Furthermore, according to this embodiment, the analysis unit acquires the number of leukemia cells (for example, the number of CML cells) as information about leukemia cells. This allows for the easy and accurate acquisition of the number of leukemia cells in a sample by analyzing the optical information obtained by the measurement unit using an artificial intelligence algorithm.

[0111] Furthermore, according to this embodiment, the sample analyzer includes a sample preparation unit (for example, a sample preparation unit 25) that mixes the sample and reagents to prepare a measurement sample. The sample preparation unit then uses a hemolytic agent (for example, a WDF hemolytic agent) as a reagent to prepare a measurement sample in which the red blood cells contained in the sample have been lysed. By using a hemolytic agent as a reagent, the sample containing lysed red blood cells can be appropriately prepared for measurement.

[0112] In this case, the reagents may also be made to not contain any staining agents. This allows for the preparation of a measurement sample by mixing the sample with a reagent that does not contain a staining agent. Furthermore, because the reagent does not contain a staining agent, information on leukemia cells contained in the sample can be obtained without labeling (no fluorescent labeling required), and the optical system for collecting fluorescence and the fluorescence receiving unit can be omitted. The above reagent may also include a diluent instead of a staining agent.

[0113] Furthermore, in this embodiment, the artificial intelligence algorithm (for example, AI algorithm 62) is trained using, for example, optical information of leukocytes contained in a sample taken from a patient with chronic myeloid leukemia (for example, waveform signals in all leukocytes of a TKI-naive CML-positive patient) as training data. The artificial intelligence algorithm is trained using optical information of leukocytes contained in samples taken from patients with chronic myeloid leukemia as training data, enabling it to appropriately acquire information about leukemia cells contained in the samples.

[0114] Furthermore, according to this embodiment, the artificial intelligence algorithm (for example, AI algorithm 62) is trained using, for example, optical information of granulocytes contained in a sample taken from a patient with chronic myeloid leukemia (for example, waveform signals in granulocytes of CML-positive patients who have not received TKI treatment) as training data. The artificial intelligence algorithm is trained using optical information of granulocytes contained in samples taken from patients with chronic myeloid leukemia as training data, enabling it to appropriately acquire information about leukemia cells contained in the samples.

[0115] <Second Embodiment> In the first embodiment, the sample was measured and CML cells were identified based on ghost cytometry performed by the GCM measurement unit 20. The second embodiment relates to a method for identifying CML cells by pre-measuring a sample using flow cytometry, which has a higher throughput than ghost cytometry, and then applying ghost cytometry to the screened sample (a sample suspected of having CML).

[0116] The contents described in the second embodiment are equally applicable to any of the other embodiments and other variations.

[0117] Figure 16 is a schematic front view showing the configuration of sample analyzer 1B, which is an example of a sample analyzer in the second embodiment. The sample analyzer 1B includes, for example, a flow cytometry measurement unit (hereinafter referred to as the "FCM measurement unit") 10, a GCM measurement unit 20, a control unit 30, and a transport unit 40.

[0118] The laboratory technician, who is the operator of the sample analyzer 1B, sets the sample container 51 containing the sample into the sample rack 50 and places the sample rack 50 in the rightmost area of ​​the transport unit 40. The transport unit 40 transports the sample rack 50 and positions it in front of the FCM measurement unit 10 and the GCM measurement unit 20 as appropriate.

[0119] The FCM measurement unit 10 removes the sample container 51 from the sample rack 50, transfers it into the FCM measurement unit 10, and measures the sample in the sample container 51. Once the measurement of the sample in the sample container 51 is complete, the FCM measurement unit 10 returns the sample container 51 to its original position in the sample rack 50. Similarly, the GCM measurement unit 20 removes the sample container 51 from the sample rack 50, transfers it into the GCM measurement unit 20, and measures the sample in the sample container 51. Once the measurement of the sample in the sample container 51 is complete, the GCM measurement unit 20 returns the sample container 51 to its original position in the sample rack 50. Once all the necessary measurements have been completed for all the sample containers 51 on one sample rack 50, the transport unit 40 transports the sample rack 50 to the leftmost area of ​​the transport unit 40. The laboratory technician then retrieves the sample rack 50 from the leftmost area.

[0120] The FCM measurement unit 10 and the GCM measurement unit 20 are configured to measure samples transported on the transport unit 40. The transport unit 40 is configured to automatically supply the sample rack 50 containing the samples to the FCM measurement unit 10 and the GCM measurement unit 20. Since samples can be automatically supplied to the FCM measurement unit 10 and the GCM measurement unit 20 via the transport unit 40, the workload of laboratory technicians required to transfer samples between the FCM measurement unit 10 and the GCM measurement unit 20 can be reduced.

[0121] The control unit 30 controls, for example, the FCM measurement unit 10, the GCM measurement unit 20, and the transport unit 40. The control unit 30 also analyzes the measurement information obtained from, for example, the FCM measurement unit 10 and the GCM measurement unit 20.

[0122] Figure 17 is a block diagram showing an example of the functional configuration of the FCM measurement unit 10.

[0123] The FCM measurement unit 10 includes, for example, a measurement control unit 11, a storage unit 12, a communication unit 13, a reading unit 14, a sample preparation unit 15, and a measurement unit 16.

[0124] The measurement control unit 11 is composed of, for example, an FPGA or a CPU. The memory unit 12 is composed of, for example, an HDD, an SSD, RAM, and ROM. The measurement control unit 11 performs various processes based on the program stored in the memory unit 12 and controls each part of the FCM measurement unit 10. The communication unit 13 is configured, for example, with a connection terminal based on the USB standard, and communicates with the control unit 30.

[0125] The reading unit 14 is composed of, for example, a barcode reader. The reading unit 14 reads the barcode from the barcode label attached to the sample container 51 and obtains the sample ID. The sample preparation unit 15 aspirates a sample from the sample container 51 and mixes the aspirated sample with a reagent to prepare a sample for measurement.

[0126] The measuring unit 16 includes, for example, an electrical measuring unit 16a, an HGB measuring unit 16b (HGB: hemoglobin), and an optical measuring unit 100. The electrical measurement unit 16a measures cells (blood cells) in the sample, for example, using the sheath flow DC detection method. The HGB measurement unit 16b measures hemoglobin in cells (blood cells) in the sample, for example, using the SLS-hemoglobin method. The optical measurement unit 100 measures cells (blood cells) in the sample, for example, by flow cytometry.

[0127] The electrical measurement unit 16a and the HGB measurement unit 16b include, for example, an amplifier and an A / D converter, and perform signal processing on the detection signal acquired by the measurement, and output the processed measurement information to the measurement control unit 11.

[0128] The optical measurement unit 100 includes, for example, an amplifier and an A / D converter, and performs signal processing on the detection signal acquired by the measurement, and outputs the processed measurement data to the measurement control unit 11.

[0129] The measurement control unit 11 stores the measurement data output from the measurement unit 16 in the storage unit 12. When the measurement of one sample is completed, the measurement control unit 11 transmits the measurement data stored in the storage unit 12 to the control unit 30, associating it with the sample ID read by the reading unit 14.

[0130] Figure 18 is a block diagram showing an example of the functional configuration of the sample preparation unit 15 for preparing a measurement sample.

[0131] The sample preparation unit 15 includes, for example, a stirring unit 15a, a suction tube 15b, and reaction chambers C11, C12, C21-C24.

[0132] The stirring unit 15a is configured to, for example, grip the sample container 51 and shake the gripped sample container 51 to stir the sample inside the sample container 51. The suction tube 15b is, for example, a nozzle with a pointed lower end, and is configured to penetrate the lid of the sample container 51, which is made of an elastic material. The suction tube 15b aspirates the sample from inside the sample container 51 after stirring and dispenses the aspirated sample into reaction chambers C11, C12, C21-C24 as needed.

[0133] In reaction chamber C11, the sample and RBC / PLT diluent are mixed to prepare an RBC / PLT measurement sample. The RBC / PLT diluent is, for example, CellPak® DCL. The RBC / PLT measurement sample prepared in reaction chamber C11 is measured by the electrical measurement unit 16a. The electrical measurement unit 16a acquires a detection signal corresponding to the blood cells in the RBC / PLT measurement sample and performs signal processing on the acquired detection signal to obtain measurement information. The control unit 31 of the control unit 30 (see Figure 7) analyzes the measurement information obtained from the measurement of the RBC / PLT measurement sample and acquires the number of red blood cells, platelets, etc.

[0134] In reaction chamber C12, the sample, HGB hemolytic agent, and HGB diluent are mixed to prepare an HGB measurement sample. The HGB hemolytic agent is, for example, Sulfolyzer®, and the HGB diluent is, for example, Cellpak® DCL. The HGB measurement sample prepared in reaction chamber C12 is measured in the HGB measurement unit 16b. The HGB measurement unit 16b acquires a detection signal corresponding to the hemoglobin concentration and performs signal processing on the acquired detection signal to obtain measurement information. The control unit 31 of the control unit 30 analyzes the measurement information obtained from the measurement of the HGB measurement sample and obtains the hemoglobin concentration, etc.

[0135] In reaction chamber C21, the sample, WDF hemolytic agent, and WDF staining solution are mixed to prepare a WDF measurement sample. The WDF hemolytic agent is, for example, LyzaCel® WDFII, and the WDF staining solution is, for example, FluoroCel® WDF. The WDF measurement sample prepared in reaction chamber C21 is measured by the optical measurement unit 100. The optical measurement unit 100 acquires detection signals corresponding to blood cells in the WDF measurement sample and performs signal processing on the acquired detection signals to obtain measurement data. The control unit 31 of the control unit 30 analyzes the measurement data obtained from the measurement of the WDF measurement sample and the measurement data obtained from the measurement of the WNR measurement sample (described later), classifies the blood cells into neutrophils, normal lymphocytes, monocytes, eosinophils, basophils, blast cells, abnormal lymphocytes, atypical lymphocytes, immature granulocytes, nucleated red blood cells, etc., and obtains the number of each blood cell.

[0136] In this case, the measurement data includes time-series data of lateral scattered light corresponding to each cell, showing the change in intensity of lateral scattered light received by the light-receiving unit 133 as each cell in the WDF measurement sample flowing through the flow cell 101 (see Figure 19) passes through the beam spot BS, and time-series data of fluorescence corresponding to each cell, showing the change in intensity of fluorescence received by the light-receiving unit 143 as each cell in the WDF measurement sample flowing through the flow cell 101 passes through the beam spot BS. The control unit 31 of the control unit 30 obtains the peak values ​​of lateral scattered light and fluorescence corresponding to each cell from the time-series data of lateral scattered light and the time-series data of fluorescence, and generates a scattergram.

[0137] In reaction chamber C22, the sample, WNR hemolytic agent, and WNR staining solution are mixed to prepare a WNR measurement sample. The WNR hemolytic agent is, for example, LyzaCel® WNR, and the WNR staining solution is, for example, FluoroCel® WNR. The WNR measurement sample prepared in reaction chamber C22 is measured by the optical measurement unit 100. The optical measurement unit 100 acquires detection signals corresponding to blood cells in the WNR measurement sample and performs signal processing on the acquired detection signals to obtain measurement data. The control unit 31 of the control unit 30 analyzes the measurement data obtained from the measurement of the WNR measurement sample, classifies leukocytes and nucleated red blood cells, etc., and obtains the number of each blood cell.

[0138] In this case, the measurement data includes time-series data of forward scattered light corresponding to each cell, showing the change in intensity of forward scattered light received by the light-receiving unit 124 as each cell in the WNR measurement sample flowing through the flow cell 101 (see Figure 19) passes through the beam spot BS, and time-series data of fluorescence corresponding to each cell, showing the change in intensity of fluorescence received by the light-receiving unit 143 as each cell in the WNR measurement sample flowing through the flow cell 101 passes through the beam spot BS. The control unit 31 of the control unit 30 obtains the peak values ​​of forward scattered light and fluorescence corresponding to each cell from the time-series data of forward scattered light and the time-series data of fluorescence, and generates a scattergram.

[0139] In reaction chamber C23, the sample, RET diluent, and RET staining solution are mixed to prepare the RET measurement sample. The RET diluent is, for example, Cellpack® DFL, and the RET staining solution is, for example, Fluorocell® RET. The RET measurement sample prepared in reaction chamber C23 is measured by the optical measurement unit 100. The optical measurement unit 100 acquires detection signals corresponding to blood cells in the RET measurement sample and performs signal processing on the acquired detection signals to obtain measurement data. The control unit 31 of the control unit 30 analyzes the measurement data obtained from the measurement of the RET measurement sample, classifies reticulocytes, etc., and obtains the number of each blood cell.

[0140] In this case, the measurement data includes time-series data of forward scattered light corresponding to each cell, showing the change in intensity of forward scattered light received by the light-receiving unit 124 as each cell in the RET measurement sample flowing through the flow cell 101 (see Figure 19) passes through the beam spot BS, and time-series data of fluorescence corresponding to each cell, showing the change in intensity of fluorescence received by the light-receiving unit 143 as each cell in the RET measurement sample flowing through the flow cell 101 passes through the beam spot BS. The control unit 31 of the control unit 30 obtains the peak values ​​of forward scattered light and fluorescence corresponding to each cell from the time-series data of forward scattered light and the time-series data of fluorescence, and generates a scattergram.

[0141] In reaction chamber C24, the sample, PLT-F diluent, and PLT-F staining solution are mixed to prepare a PLT-F measurement sample. The PLT-F diluent is, for example, Cellpack® DFL, and the PLT-F staining solution is, for example, Fluorocell® PLT. The PLT-F measurement sample prepared in reaction chamber C24 is measured by the optical measurement unit 100. The optical measurement unit 100 acquires detection signals corresponding to blood cells in the PLT-F measurement sample and performs signal processing on the acquired detection signals to obtain measurement data. The control unit 31 of the control unit 30 analyzes the measurement data obtained from the measurement of the PLT-F measurement sample, classifies platelets, etc., and obtains the number of each blood cell.

[0142] In this case, the measurement data includes time-series data of forward scattered light corresponding to each cell, showing the change in intensity of forward scattered light received by the light-receiving unit 124 as each cell in the PLT-F measurement sample flowing through the flow cell 101 (see Figure 19) passes through the beam spot BS, and time-series data of fluorescence corresponding to each cell, showing the change in intensity of fluorescence received by the light-receiving unit 143 as each cell in the PLT-F measurement sample flowing through the flow cell 101 passes through the beam spot BS. The control unit 31 of the control unit 30 obtains the peak values ​​of forward scattered light and fluorescence corresponding to each cell from the time-series data of forward scattered light and the time-series data of fluorescence, and generates a scattergram.

[0143] The measurement data is not limited to the peak values ​​mentioned above and may be any information that reflects the size, shape, internal structure, or nucleic acid content of each cell, obtained by irradiating each cell in the sample with at least one beam of light having a single beam spot. Furthermore, the diluents, hemolytic agents, and staining solutions mixed in reaction chambers C11, C12, and C21-C24 are not limited to the reagents mentioned above.

[0144] Figure 19 is a schematic diagram showing the configuration of the optical measurement unit 100. For convenience, the X, Y, and Z axes, which are orthogonal to each other, are indicated in Figure 19. The Z axis direction is the flow direction of the sample to be measured in the flow cell 101.

[0145] The optical measuring unit 100 includes a flow cell 101, a light source 111, a collimator lens 112, a cylindrical lens 113, a focusing lens 114, focusing lenses 121 and 131, a beam stopper 122, optical filters 123, 132 and 142, light receiving units 124, 133 and 143, and a dichroic mirror 141.

[0146] The light source 111 is, for example, a semiconductor laser light source. The light source 111 emits light of a predetermined wavelength λ10 in the X-axis direction. The wavelength λ10 is, for example, 488 nm or 642 nm. The collimator lens 112 converts the light emitted from the light source 111 into parallel light. The cylindrical lens 113 focuses the light from the light source 111 in the Y-axis direction. The focusing lens 114 converges the light from the light source 111 in the Y-axis and Z-axis directions, flattens it at the position of the flow cell 101, and focuses it into the flow channel 101a of the flow cell 101.

[0147] Figure 20 is a schematic side view showing the configuration of the flow cell 101.

[0148] Light from the light source 111 is directed onto the irradiation position of the flow channel 101a of the flow cell 101 as a single, flattened beam spot BS with a small width in the Z-axis direction, due to the action of the cylindrical lens 113 and the focusing lens 114. Hereinafter, the light emitted from the light source 111 and directed onto the irradiation position of the flow channel 101a will be referred to as "straight illumination light". When straight illumination light irradiates cells flowing through the flow channel 101a, forward scattered light, side scattered light, and fluorescence are generated from the irradiated area of ​​the cell. Here, it is assumed that when straight illumination light with wavelength λ10 irradiates a fluorescent dye used to stain cells, light with wavelength λ11 is generated from the fluorescent dye.

[0149] As mentioned above, the flow rate is adjusted in the fluid adjustment section 200a of the flow cell 201 so that the flow rate per unit time is almost constant. Therefore, the throughput of measurement by the GCM measurement unit 20 may be lower than the throughput of measurement by the FCM measurement unit 10. In addition, the control unit 31 can control the operation of the GCM measurement unit 20 in accordance with the flow rate adjustment so that the number of cells measured by the GCM measurement unit 20 is less than the number of cells measured by the FCM measurement unit 10.

[0150] Returning to Figure 19, the focusing lens 121 focuses the forward scattered light of wavelength λ10 generated from the cells onto the light receiving unit 124. The beam stopper 122 blocks the light of wavelength λ10 that has passed through the flow cell 101 without irradiating the cells, and allows the forward scattered light of wavelength λ10 generated from the cells to pass through. The optical filter 123 is configured to transmit only light of wavelength λ10. The light receiving unit 124 receives the forward scattered light of wavelength λ10 that has passed through the optical filter 123 and outputs a detection signal according to the light detection intensity. The light receiving unit 124 is, for example, a photodiode (PD).

[0151] The focusing lens 131 focuses the side-scattered light of wavelength λ10 emitted from the cells onto the light-receiving unit 133 and focuses the fluorescence of wavelength λ11 emitted from the cells onto the light-receiving unit 143. The dichroic mirror 141 transmits light of wavelength λ10 and reflects light of wavelength λ11. The optical filter 132 is configured to transmit only the light of wavelength λ10 from the dichroic mirror 141. The light-receiving unit 133 receives the side-scattered light of wavelength λ10 that has passed through the optical filter 132 and outputs a detection signal according to the light-receiving intensity. The light-receiving unit 133 is, for example, a photodiode (PD).

[0152] The optical filter 142 is configured to transmit only light of wavelength λ11 from the dichroic mirror 141. The light-receiving unit 143 receives the fluorescence of wavelength λ11 that has passed through the optical filter 142 and outputs a detection signal according to the light-receiving intensity. The light-receiving unit 143 is, for example, a photomultiplier tube (PMT), an avalanche photodiode (APD), or a photodiode (PD).

[0153] The control unit 31 performs cell analysis based on measurement information acquired by, for example, the FCM measurement unit 10 and the GCM measurement unit 20.

[0154] The control unit 31 analyzes the measurement information acquired by the electrical measurement unit 16a and HGB measurement unit 16b of the FCM measurement unit 10 and the measurement data acquired by the optical measurement unit 100 of the FCM measurement unit 10 to obtain FCM analysis results. For example, the control unit 31 generates a scattergram for each sample based on the measurement data and classifies the cells based on the generated scattergram.

[0155] For example, the control unit 31 generates a scattergram based on measurement data obtained by measuring a WDF measurement sample, and groups multiple cell groups corresponding to plots on the generated scattergram. In grouping cell groups, for example, the control unit 31 calculates the centroid of a predetermined region of plots on the scattergram, performs cluster analysis on the cell groups based on the distance from each plot to the centroid, and classifies each cell. An example of a scattergram will be described later.

[0156] By grouping the cell population, regions are set up, for example, for normal lymphocytes, monocytes, neutrophils and basophils, and eosinophils. If the sample contains blast cells, abnormal lymphocytes, atypical lymphocytes, immature granulocytes, or nucleated red blood cells, regions corresponding to these blood cells are also set up. The control unit 31 counts the plots within the regions set up in the scattergram to obtain the number of blood cells in each category.

[0157] The control unit 31 does not necessarily have to actually generate a scattergram to group the cell populations; for example, it may perform cell population-based grouping by processing data corresponding to the scattergram.

[0158] Next, the measurement process of the sample analyzer 1B will be described with reference to Figures 21 and 22.

[0159] Figure 21 is a flowchart showing an example of the control process flow related to measurement by the control unit 30.

[0160] In step S11, the control unit 31 of the control unit 30 controls the FCM measurement unit 10 so that the FCM measurement process is performed. As a result, the control unit 31 analyzes the measurement information and measurement data acquired by the FCM measurement unit 10 and obtains the FCM analysis results.

[0161] The FCM analysis results may include, for example, the cell count in the sample (number of cells per unit volume, etc.), an abnormal cell flag indicating the presence of abnormal cells, and a scattergram or histogram.

[0162] The count values ​​from the FCM analysis results may be, for example, count values ​​for red blood cells, white blood cells, neutrophils, normal lymphocytes, monocytes, eosinophils, basophils, platelets, abnormal cells, etc. The abnormal cell flag in the FCM analysis results may include, for example, a flag indicating that the number of abnormal cells per unit volume in the sample, such as blast cells, abnormal lymphocytes, atypical lymphocytes, immature granulocytes, and nucleated red blood cells, is above a predetermined threshold, and a flag indicating that the classification status of white blood cells is abnormal. Abnormal cells, in the case of blood samples, may be defined as cells that are not present in the peripheral blood of healthy individuals, or are present in only small numbers. Furthermore, a number of abnormal cells per unit volume exceeding a predetermined threshold can be considered as the presence of abnormal cells.

[0163] If the number of abnormal cells per unit volume in the sample, such as blast cells, abnormal lymphocytes, atypical lymphocytes, immature granulocytes, and nucleated red blood cells, does not meet the specified threshold, the abnormal cell flag will not be included in the FCM analysis results.

[0164] The FCM measurement process will be explained later with reference to Figure 22.

[0165] In step S12, the control unit 31 determines whether the FCM analysis results obtained in step S11 include an abnormal cell flag. If the FCM analysis results are determined to include an abnormal cell flag (step S12: YES), in step S13, the control unit 31 controls the GCM measurement unit 20 so that the GCM measurement process is performed. As a result, the control unit 31 analyzes the waveform signal acquired by the GCM measurement unit 20 and obtains the GCM analysis results.

[0166] The GCM measurement process can be performed, for example, according to steps S110 to S140 in Figure 10.

[0167] Next, in step S14, the control unit 31 generates cell analysis results (cell analysis results) for the target sample based on the FCM analysis results obtained in step S11 and the GCM analysis results obtained in step S13.

[0168] On the other hand, if the FCM analysis results are determined not to include abnormal cell flags (step S12: NO), in step S15, the control unit 31 generates cell analysis results (cell analysis results) for the target sample based on the FCM analysis results obtained in step S11.

[0169] In other words, in step S12, the control unit 31 selectively determines whether to generate cell analysis results based on FCM analysis results or based on both FCM analysis results and GCM analysis results.

[0170] For example, when a medical technologist inputs a display instruction via the input unit 35, in step S16, the control unit 31 displays a cell analysis results screen 300, including the cell analysis results generated in step S14 or step S15, on the display unit 34. The cell analysis results screen 300 will be described later with reference to Figures 23-30.

[0171] As described above, in the example in Figure 21, the GCM measurement process is executed when the FCM analysis results include an abnormal cell flag, so the measurement frequency by the GCM measurement unit 20 is lower than that of the FCM measurement unit 10. In other words, the control unit 31 can control each measurement unit so that the measurement frequency by the GCM measurement unit 20 is lower than that of the FCM measurement unit 10. As explained with reference to Figure 5, in the GCM measurement unit 20, the flow rate of the sample to be measured is adjusted so that only one cell is positioned in the irradiation range R of the diffraction illumination light at one time. Therefore, the throughput of measurement by the GCM measurement unit 20 may be lower than that of measurement by the FCM measurement unit 10. Consequently, if the measurement frequency by the GCM measurement unit 20 becomes the same as that of the FCM measurement unit 10, the overall throughput of the laboratory may decrease. By making the measurement frequency of the GCM measurement unit 20 lower than that of the FCM measurement unit 10, it is possible to take advantage of the benefits of the GCM measurement unit 20 while suppressing a decrease in the overall throughput of the laboratory.

[0172] Figure 22 is a flowchart showing an example of the FCM measurement process flow.

[0173] In step S101, the control unit 31 of the control unit 30 controls the FCM measurement unit 10 so that sample preparation is performed by the sample preparation unit 15 (see Figures 17 and 18) of the FCM measurement unit 10. As a result, the measurement samples are prepared in reaction chambers C11, C12, and C21-C24.

[0174] For convenience, this explanation assumes that the measurement sample is prepared in all reaction chambers C11, C12, and C21-C24. However, in reality, the necessary measurement samples are prepared according to the specified measurement items for the target specimen, and in the subsequent steps S102 and S103, the measurement samples are measured in the electrical measurement unit 16a, HGB measurement unit 16b, and optical measurement unit 100, respectively, according to the prepared measurement samples.

[0175] In step S102, the control unit 31 controls the FCM measurement unit 10 so that measurements are performed by the electrical measurement unit 16a and the HGB measurement unit 16b, and acquires measurement information based on these measurements from the FCM measurement unit 10.

[0176] In step S103, the control unit 31 controls the FCM measurement unit 10 so that the optical measurement unit 100 performs the measurement. As a result, the optical measurement unit 100 irradiates the cells in the sample contained in the sample flowing through the flow cell 101 with a straight illumination light having a single beam spot BS, as shown in Figure 20, and acquires measurement data. The control unit 31 then acquires the measurement data acquired by the measurement unit 16 from the FCM measurement unit 10.

[0177] In step S104, the control unit 31 analyzes the measurement information acquired in step S102 and the measurement data acquired in step S103.

[0178] In step S105, the control unit 31 generates FCM analysis information based on the analysis in step S104. The FCM analysis information includes the blood cell count obtained by analyzing the measurement information based on the measurements of the electrical measurement unit 16a and the HGB measurement unit 16b, and the blood cell count and abnormal cell flag obtained by analyzing the measurement data based on the measurements of the optical measurement unit 100.

[0179] Next, with reference to Figures 23 to 25, an example of the cell analysis results screen 300 displayed in step S16 of Figure 21 will be described.

[0180] Figure 23 is a schematic diagram showing the configuration of the cell analysis results screen 300.

[0181] The cell analysis results screen 300 includes, for example, a count value display area 310 and an abnormal cell flag display area 320.

[0182] The count value display area 310 includes, for example, display areas 311 to 314 corresponding to the analysis modes CBC, DIFF, RET, and PLT-F, respectively.

[0183] The display area 311 shows count values ​​corresponding to the CBC mode, such as white blood cell count, red blood cell count, hemoglobin level, hematocrit value, mean corpuscular volume, etc.

[0184] The display area 312 shows count values ​​corresponding to DIFF mode, such as neutrophil count, lymphocyte count, monocyte count, eosinophil count, and basophil count.

[0185] The display area 313 shows count values ​​corresponding to RET mode, such as reticulocyte ratio, reticulocyte count, reticulocyte immaturity index, and reticulocyte hemoglobin equivalent.

[0186] Display area 314 displays count values ​​corresponding to PLT-F mode, such as the immature platelet ratio.

[0187] Additionally, the cell analysis results screen 300 may be configured to display the scattergram and histogram included in the FCM analysis results.

[0188] The abnormal cell flag display area 320 includes display areas 321 to 323 that display abnormal cell flags for white blood cells, red blood cells, and platelets, respectively.

[0189] The display area 321 is marked with a label 321a, which displays the number 1 or 2. A label 321a displaying "1" indicates that the abnormal cell flag for leukocytes displayed in the display area 321 is based on the FCM analysis results, while a label 321a displaying "2" indicates that the abnormal cell flag for leukocytes displayed in the display area 321 is based on the GCM analysis results.

[0190] When step S15 in Figure 21 is performed, the GCM measurement process is not performed, so no GCM analysis results are obtained, and the cell analysis results are generated based on the FCM analysis results based on the FCM measurement process. Therefore, in this case, as shown in Figure 24, the cell analysis results screen 300 displays cell analysis results based only on the FCM analysis results, and "1" is displayed in label 321a in the display area 321 that displays abnormal cell flags related to white blood cells, to indicate that abnormal cell flags based on the FCM analysis results are displayed. Note that in the example shown in Figure 23, there were no abnormal cell flags based on the FCM analysis results, so no abnormal cell flags are displayed in the display areas 321 to 323.

[0191] When step S14 in Figure 21 is performed, the GCM measurement process is executed, and the cell analysis results are generated based on the FCM analysis results based on the FCM measurement process and the GCM analysis results. In this case, as shown in Figure 25, on the cell analysis results screen 300, the count value display area 310 and display areas 322 and 323 display the count value and abnormal cell flag based on the FCM analysis results, and the display area 321 that displays the abnormal cell flag for leukocytes displays the display based on the GCM analysis results instead of the abnormal cell flag based on the FCM analysis results. To indicate that the GCM analysis results are displayed in display area 321, "2" is displayed on label 321a. In the example shown in Figure 25, the display area 321 displays "CML?", indicating a suspected case of CML, and the CML cell content "(20%)", based on the GCM analysis results. In other words, in this example, at least a portion of the results based on the FCM analysis results is supplemented with the GCM analysis results. For example, the "CML?" label may be displayed if the CML cell content exceeds a predetermined percentage (e.g., "20%") or is equal to or greater than a predetermined percentage. Alternatively, the CML cell content may not be displayed at all, or only the CML cell content may be displayed.

[0192] Furthermore, if GCM measurement is performed and the GCM analysis results indicate a low suspicion of CML (for example, if the CML cell content is less than or equal to a predetermined percentage), the comparative example "CML?" may not be displayed in the display area 321, and "2" may be displayed on the label 321a.

[0193] Thus, regardless of the presence or type of anomaly flag in the FCM analysis results, if GCM measurement processing is performed, the display will be based on the GCM analysis results. This allows the laboratory technician to accurately determine whether abnormal cells are present in the sample, enabling them to accurately decide whether or not to prepare a smear at the end of the sample analysis device 1B. Even if a smear is prepared, it can be smoothly verified based on an accurate abnormal cell flag derived from the GCM analysis results.

[0194] <Effects of the second embodiment> In this embodiment, the sample analyzer (e.g., sample analyzer 1B) further comprises a first measurement unit (e.g., FCM measurement unit 10) that acquires measurement results regarding information on blood cells contained in the sample. The aforementioned measurement unit is a second measurement unit (e.g., GCM measurement unit 20) different from the first measurement unit, and the second measurement unit performs a measurement on the sample and acquires optical information of the cells contained in the sample when the measurement results from the first measurement unit on the sample satisfy predetermined conditions (e.g., the number of abnormal cells per unit volume in the sample, such as blast cells, abnormal lymphocytes, atypical lymphocytes, immature granulocytes, and nucleated red blood cells, is above a predetermined threshold). According to this, when the measurement result of the first measurement unit on the sample meets predetermined conditions, a second measurement unit, different from the first measurement unit, performs measurement on the sample and acquires optical information of the cells contained in the sample. As a result, the measurement frequency of the second measurement unit becomes lower than that of the first measurement unit, which helps to suppress a decrease in throughput, as mentioned above.

[0195] In this case, the specified conditions described above may include an increase in a particular type of white blood cell or the appearance of blast cells. According to this, if the measurement results from the first measurement unit on the sample meet the conditions for an increase in a specific type of white blood cell or the appearance of blast cells, the second measurement unit performs a measurement on the sample to obtain optical information of the cells contained in the sample. If an increase in a specific type of white blood cell or the appearance of blast cells is observed, it may indicate CML. Therefore, the second measurement unit can perform a measurement of the sample and acquire optical information of the cells contained in the sample, based on predetermined conditions such as an increase in basophils or the appearance of blast cells.

[0196] In this case, the specific type of white blood cell mentioned above may be a basophil. According to this, if the measurement results from the first measurement unit on the sample meet the conditions for an increase in basophils or the appearance of blast cells, the second measurement unit performs a measurement on the sample to obtain optical information of the cells contained in the sample. An increase in basophils may indicate CML. Therefore, the second measurement unit can be configured to perform a measurement on the sample and acquire optical information of the cells contained in the sample, based on a predetermined condition of an increase in basophils.

[0197] Furthermore, the above-mentioned conditions may include the appearance of abnormal cells other than blast cells, such as abnormal lymphocytes, atypical lymphocytes, immature granulocytes, and nucleated red blood cells.

[0198] <Modified form of the second embodiment> In the above embodiment, when GCM measurement processing was performed, not only the GCM analysis results but also the FCM analysis results were displayed as cell analysis results. However, this is not limited to this, and the FCM analysis results do not necessarily have to be displayed.

[0199] For example, in step S14 of Figure 21, the control unit 31 of the control unit 30 may generate cell analysis results based solely on the GCM analysis results. The configuration of the cell analysis results screen 300 when the GCM measurement process is performed may be the same as, for example, the cell analysis results screen 600 in Figure 9.

[0200] In this modified example, according to the GCM analysis results, the number of false positives for laboratory technicians can be reduced, for example, by decreasing the number of samples that result in false positives. This allows for more accurate verification of smears based on the cell analysis results, and the preparation and verification of smears can be omitted. Therefore, the burden on laboratory technicians can be reduced.

[0201] <Third Embodiment> In the second embodiment, the FCM measurement process and the GCM measurement process were performed by the FCM measurement unit 10 and the GCM measurement unit 20, respectively. In contrast, the third embodiment is an embodiment relating to performing both FCM measurement processing and GCM measurement processing using an integrated flow cytometry / ghost cytometry measurement unit (hereinafter referred to as the "integrated measurement unit").

[0202] The contents described in the third embodiment are equally applicable to any of the other embodiments and any of the other modifications.

[0203] Figure 26 is a schematic front view showing the configuration of sample analyzer 1C, which is an example of a sample analyzer according to the third embodiment.

[0204] Compared to the second embodiment shown in Figure 16, the sample analyzer 1C includes an integrated measurement unit 70 instead of the FCM measurement unit 10 and the GCM measurement unit 20.

[0205] Figure 27 is a block diagram showing an example of the functional configuration of the integrated measurement unit 70.

[0206] Compared to the GCM measurement unit 20 of the first embodiment shown in Figure 2, the integrated measurement unit 70 includes the electrical measurement unit 16a and HGB measurement unit 16b shown in Figure 17, and replaces the sample preparation unit 25 and optical measurement unit 200 with a sample preparation unit 27 and an optical measurement unit 400. The fluid adjustment unit 400a within the optical measurement unit 400 adjusts the flow rate of the sample to be measured per unit time in the flow cell 201 of the optical measurement unit 400, and is configured similarly to the fluid adjustment unit 200a in Figure 2. The sample preparation unit 27 will be described later with reference to Figure 28. The optical system of the optical measurement unit 400 will be described later with reference to Figure 29.

[0207] The electrical measurement unit 16a and the HGB measurement unit 16b perform signal processing on the detection signal acquired by the measurement and output the processed measurement information to the measurement control unit 21.

[0208] The optical measurement unit 400 performs signal processing on the detection signal acquired by the measurement and outputs the measurement data to the measurement control unit 21.

[0209] The measurement control unit 21 stores the measurement data output from the measurement unit 26 in the storage unit 22. When the measurement of one sample is completed, the measurement control unit 21 associates the measurement data stored in the storage unit 22 with the sample ID read by the reading unit 24 and transmits it to the control unit 30.

[0210] Figure 28 is a block diagram showing an example of the functional configuration of the sample preparation unit 27.

[0211] The sample preparation unit 27, compared to the sample preparation unit 25 of the first embodiment shown in Figure 3, includes reaction chambers C11, C12, and C21-C24 as shown in Figure 18. Reaction chambers C11 and C12 are connected to the electrical measurement unit 16a and the HGB measurement unit 16b, respectively, while reaction chambers C21-C24 and C30 are connected to the optical measurement unit 400.

[0212] The suction tube 25b aspirates the sample from the sample container 51, which has been stirred by the stirring unit 25a, and dispenses the aspirated sample into reaction chambers C11, C12, C21-C24, and C30 as needed.

[0213] The samples prepared in reaction chambers C21-C24 and C30 are each individually flowed into flow cell 201 and measured by the optical measurement unit 400.

[0214] The optical measurement unit 400 measures the sample prepared in reaction chambers C21 to C24 to acquire a detection signal, and performs signal processing on the acquired detection signal to obtain measurement data. Furthermore, the optical measurement unit 400 measures the sample prepared in the reaction chamber C30 to acquire a detection signal, and performs signal processing on the acquired detection signal to obtain a waveform signal.

[0215] Figure 29 is a schematic diagram showing the configuration of the optical measuring unit 400.

[0216] Compared to the optical measuring unit 200 in Figure 4, the optical measuring unit 400 includes the light source 111, collimator lens 112, cylindrical lens 113, beam stopper 122, optical filters 123, 132, 142, and light receiving units 124, 133, 143 of the optical measuring unit 100 shown in Figure 19, and further includes dichroic mirrors 115, 125, 134, 144 and a focusing lens 126.

[0217] The dichroic mirror 115 reflects light of wavelength λ10 from the light source 111 and transmits light of wavelength λ20 from the light source 211. The dichroic mirror 115 aligns the optical axis of the light from the light source 111 with the central axis of the light from the diffractive optical element 215. The focusing lens 216 focuses the light from the light sources 111 and 211 into the flow channel 201a of the flow cell 201. The focusing lens 216 is configured to suppress chromatic aberration for light of wavelengths λ10 and λ20. The beam spot BS (see Figure 20) of the straight illumination light from the light source 111 is positioned at the center of the illumination range R shown in Figure 5. The diffractive illumination light from the light source 211 is irradiated into the illumination range R, as in Figure 5.

[0218] The collimator lens 112, cylindrical lens 113, dichroic mirror 115, and focusing lens 216 constitute an illumination optical system 206 that irradiates cells passing through the flow cell 201 with light from the light source 111 as straight illumination light.

[0219] Similar to the second embodiment, when cells flowing through the flow cell 201 are irradiated with rectifying light of wavelength λ10, forward scattered light of wavelength λ10, side scattered light of wavelength λ10, and fluorescence of wavelength λ11 are generated from the irradiated area of ​​the cell. When cells flowing through the flow cell 201 are irradiated with diffracting light of wavelength λ20, forward scattered light of wavelength λ20, side scattered light of wavelength λ20, and fluorescence of wavelength λ21 are generated from the irradiated area of ​​the cell.

[0220] The dichroic mirror 125 reflects the rectified illumination light and the forward scattered light based on the rectified illumination light, and transmits the diffracted illumination light and the forward scattered light based on the diffracted illumination light. The rectified illumination light and the diffracted illumination light that have passed through the flow cell 201 are blocked by the beam stoppers 122 and 222, respectively. The focusing lens 126 focuses the forward scattered light based on the diffracted illumination light that has passed through the beam stopper 122 onto the light receiving unit 124. The dichroic mirror 134 reflects the side scattered light based on the rectified illumination light and transmits the side scattered light based on the diffracted illumination light. The dichroic mirror 144 reflects the fluorescence based on the rectified illumination light and transmits the fluorescence based on the diffracted illumination light. The light receiving units 124, 133, 143, 225, 233, and 243 receive the corresponding light and output a detection signal.

[0221] Figure 30 is a flowchart showing an example of the control process flow related to measurement by the control unit 30.

[0222] In the control process shown in Figure 30, compared to the second embodiment shown in Figure 21, the GCM measurement process in step S13 is executed between steps S11 and S12. That is, in the second embodiment, the GCM measurement process is executed regardless of the FCM analysis results. In the FCM measurement process of the second embodiment, the integrated measurement unit 70 performs the same processing as in Figure 22, and in the GCM measurement process, the integrated measurement unit 70 performs the same processing as, for example, steps S110 to S140 in Figure 10.

[0223] <Effects of the Third Embodiment> According to this embodiment, in a sample analyzer (e.g., sample analyzer 1C), the measurement unit (e.g., measurement unit 26: optical measurement unit 400) acquires information that identifies at least leukocytes among the cells contained in the sample. Then, the analysis unit (e.g., control unit 30) acquires the number (e.g., CML cell count) or percentage (e.g., CML cell content) of leukemia cells among the leukocytes identified based on the information acquired by the measurement unit, based on the optical information obtained by the measurement unit. This makes it possible to obtain, based on optical information, the number or percentage of leukemia cells among the leukocytes identified based on information that identifies at least leukocytes among the cells contained in the sample.

[0224] Furthermore, according to this embodiment, the measurement unit acquires information to identify at least granulocytes among the cells contained in the sample, and the analysis unit acquires the number or percentage of leukemia cells among the granulocytes identified based on the information acquired by the measurement unit, based on the optical information obtained by the measurement unit. This makes it possible to obtain, based on optical information, the number or percentage of leukemia cells among the granulocytes identified based on information that identifies at least granulocytes among the cells contained in the sample.

[0225] In this case, the information used to identify at least granulocytes among the cells contained in the above sample may include the intensity of scattered light obtained by irradiating the cells with multiple diffracted lights or a single irradiated light. This makes it possible to obtain, based on optical information, the number or percentage of leukemia cells among the granulocytes identified, which are identified based on information that identifies at least granulocytes among the cells contained in the sample, including the intensity of scattered light obtained by irradiating cells with multiple diffracted lights or a single irradiated light.

[0226] <First modified example of the third embodiment> In the third embodiment, as shown in Figure 30, the first analysis result was determined in step S12, and either step S14 or S15 was executed according to the first analysis result. However, the embodiment is not limited to this, and the determination in step S12 may be omitted.

[0227] Figure 31 is a flowchart showing an example of the control process flow related to measurement by the control unit 30 in this modified example.

[0228] In this modified control process, steps S41 and S42 are added in place of steps S12 and S14-S16, compared to the flowchart shown in Figure 30.

[0229] In step S41, the control unit 31 of the control unit 30 generates cell analysis results based on the GCM analysis results. In this case, the cell analysis results include count values ​​(e.g., total white blood cell count and CML cell count) and CML cell content included in the GCM analysis results. In step S42, the control unit 31 outputs (e.g., displays) the cell analysis results generated in step S41 to the cell analysis results screen 300, and also outputs (e.g., displays) the FCM analysis results obtained in the FCM measurement process in step S11 as reference information. In this case, the FCM analysis results may be added to the cell analysis results screen 300 when a button provided on the cell analysis results screen 300 is operated, for example, or the FCM analysis results may be displayed on the cell analysis results screen 300 along with a label indicating that it is reference information.

[0230] In steps S41 and S42 of Figure 31, cell analysis results are generated based on GCM analysis results and FCM analysis results are displayed as reference information. However, cell analysis results may be generated based on FCM analysis results and GCM analysis results may be displayed as reference information. For example, the control unit 30 may selectively determine whether to generate cell analysis results based on FCM analysis results or GCM analysis results. For example, depending on the settings and operations of the medical technologist, the control unit 30 may selectively determine whether to generate cell analysis results based on FCM analysis results or GCM analysis results.

[0231] <Second modified example of the third embodiment> In the third embodiment, the sample used in the GCM measurement process was prepared in reaction chamber C30, but it is not limited to this; an RBC / PLT sample prepared in reaction chamber C11 may also be used in the GCM measurement process.

[0232] Figure 32 is a block diagram showing the functional configuration of the sample preparation unit 27 in this modified example.

[0233] In this modified example, the sample preparation unit 27 omits the reaction chamber C30 compared to the third embodiment shown in Figure 28, and the reaction chamber C11 is connected to the optical measurement unit 400. In this modified example, for the second measurement sample, the RBC / PLT measurement sample prepared in the reaction chamber C11 is flowed into the flow cell 201, and the measurement sample is irradiated with diffractive illumination light to acquire a waveform signal. In this case, since fluorescence based on diffractive illumination light is not acquired, the GCM analysis results are obtained by analyzing the waveform signals based on forward scattered light and side scattered light based on diffractive illumination light.

[0234] According to this modified example, the reaction chamber C30 can be omitted, thus simplifying the configuration of the sample analyzer 1C.

[0235] In this modified example, in the GCM measurement process, an RBC / PLT measurement sample was flowed through the flow cell 201 to acquire a waveform signal. However, this is not the only case, and a WDF measurement sample may be flowed through the flow cell 201 to acquire a waveform signal.

[0236] In this case, the GCM measurement process may be performed simultaneously with the FCM measurement process. That is, when the WDF measurement sample prepared in the reaction chamber C21 is flowed through the flow cell 201, information for identifying white blood cells and a waveform signal may be acquired simultaneously. However, in this case, in order to appropriately acquire the waveform signal, it is necessary to reduce the flow rate per unit time of the WDF measurement sample flowing through the flow cell 201 under the control of the fluid adjustment unit 400a as compared with the case of acquiring only the information for identifying white blood cells from the WDF measurement sample. However, since information for identifying white blood cells and a waveform signal can be acquired simultaneously, it becomes possible to shorten the throughput of specimen analysis.

[0237] <Fourth Embodiment> As a standard treatment for CML, for example, with the emergence of TKI, which is one of the aforementioned molecular target drugs, the prognosis of CML has been dramatically improved. However, many patients need to take TKI orally for a long period, and various adverse events have been reported with long-term oral administration. As a treatment goal for CML, it is important to aim for treatment-free remission (TFR) that maintains remission even after discontinuation of TKI. However, only about 50% of patients can actually achieve TFR.

[0238] BCR::ABL1, the causative gene for chronic myeloid leukemia (CML), is measured by quantitative PCR, and it has been reported that a rapid initial decrease in BCR::ABL1 after the initiation of a treatment-free agent (TKI) correlates with an increased likelihood of achieving total remission (TFR) (Shanmuganathan N, et al, Early BCR-ABL1 kinetics are predictive of subsequent achievement of treatment-free remission in chronic myeloid leukemia, Blood. 2021, DOI: 10.1182 / blood.2020005514). In other words, predicting the response to TKI treatment in CML treatment is considered to be useful information for the attending physician in deciding whether to select or discontinue TKIs.

[0239] This embodiment relates to predicting (predicting) the treatment response to CML based on the results of GCM classification of leukocytes in CML patients at the time of initial examination.

[0240] The contents described in the fourth embodiment are equally applicable to any of the other embodiments and any of the other modifications.

[0241] Figure 33 shows a statistical comparison of the IS% distribution in 10 CML patients who received second-generation TKI treatment. The graph compares patients with a BCR::ABL1 mRNA (IS%) of 1% or higher at 3 months of treatment (i.e., patients with a low response to the drug, hereinafter referred to as "late responders") and patients with a BCR::ABL1 mRNA (IS%) of less than 1% at 3 months of treatment (i.e., patients with a high response to the drug, hereinafter referred to as "early responders"). The graph on the left corresponds to "late responders" (n=4 (4 cases)), and the graph on the right corresponds to "early responders" (n=6 (6 cases)). The vertical axis represents (IS%).

[0242] This graph shows that at 3 months of treatment, there is a significant difference in BCR::ABL1 mRNA (IS%) levels between late responders and early responders. As mentioned above, the rapid initial decrease in BCR::ABL1 after TKI initiation has been reported to correlate with an increased likelihood of achieving TFR. Therefore, it is important to assume that early responders are patients with a high probability of achieving TFR and late responders are patients with a low probability of achieving TFR, and to develop treatment plans accordingly.

[0243] While late responders and early responders can be distinguished by tracking time-series data using the PCR method described above, the BCR::ABL1 mRNA (IS%) value at the initial consultation is around 100% in 10 CML patients, making distinction impossible at this point. Therefore, the inventors investigated a method that can distinguish between late responders and early responders even at the initial consultation. In this method, a classification model is generated (constructed) that classifies cells in samples taken from patients with chronic myeloid leukemia before the start of treatment (hereinafter referred to as "target patients") as CML cells and cells in samples taken from healthy individuals as normal cells. By using an index related to the classification performance of this classification model between CML cells and normal cells, it is possible to determine whether the target patient is a late responder (a group with low treatment response) or an early responder (a group with high treatment response). Any index that represents the classification performance of the classification model can be used, such as the F1 score or AUC. A higher classification performance indicates a significant morphological difference between CML cells and normal cells in the target patient's sample.

[0244] The upper part of Figure 34 shows a graph comparing the F1 scores used by an AI algorithm trained on samples from CML patients to classify CML cells and normal cells, for late responders and early responders as shown in Figure 33. The graph on the left corresponds to "late responders" (n=4), and the graph on the right corresponds to "early responders" (n=6). The vertical axis represents the F1 score, which is the discrimination result when comparing CML patient samples and healthy control samples using ghost cytometry. In other words, the vertical axis can be said to represent the degree of morphological difference between cells from CML patient samples and cells from healthy control samples. This graph shows that late responders have significantly higher F1 scores compared to early responders.

[0245] The lower part of Figure 34 shows the ROC curves when classifying late responders and early responders based on their F1 scores. The results of this ROC analysis show that, with a cutoff value of "85.5%", the detection could be performed with a sensitivity of "75%" and a specificity of "100%".

[0246] In other words, by focusing on the F1 score obtained by differentiating between peripheral blood leukocytes of CML patients and healthy individuals at the time of initial examination, it may be possible to distinguish between late responders and early responders based on the CML cell content at the time of initial examination.

[0247] Figure 35 is a flowchart showing an example of the control process flow related to measurement by the control unit 30. In this embodiment, for example, the control unit 30 in the sample analyzer 1 described in the various embodiments above may perform the following processing.

[0248] In step S13, the control unit 31 controls the GCM measurement unit 20 so that GCM measurement processing is performed on the target patient's sample. As a result, the control unit 31 acquires waveform signals from the target patient's CML cells in the GCM measurement unit 20.

[0249] In step S51, the control unit 31 trains the AI ​​algorithm 61 using the waveform signals of some cells (e.g., 75% of all cells) from among the waveform signals of multiple cells contained in the sample of the target patient as training data, i.e., as "CML(+) signals". Furthermore, the control unit 31 trains the AI ​​algorithm 61 using the waveform signals of some cells (e.g., 75% of all cells) from among the waveform signals of multiple cells contained in the sample of one or more healthy individuals as training data, i.e., as "CML(-) signals". As a result, the control unit 31 obtains the trained AI algorithm 62. The control unit 31 inputs waveform signals from multiple waveform signals obtained from a sample of a target patient that were not used as training data (for example, waveform signals from the remaining 25% of cells) into the trained AI algorithm 62, causing it to classify CML cells as positive or negative. In this case, if the AI ​​algorithm 62 classifies a signal as positive, it becomes TP (True Positive), and if it classifies a signal as negative, it becomes FN (False Negative). Similarly, the control unit 31 inputs waveform signals from multiple waveform signals obtained from a sample of a healthy person that were not used as training data into the trained AI algorithm 62, causing it to classify CML cells as positive or negative. In this case, if the AI ​​algorithm 62 classifies a signal as positive, it becomes FP (False Positive), and if it classifies a signal as negative, it becomes TN (True Negative).

[0250] Next, in step S52, the control unit 31 calculates an index (e.g., F1 score) that represents the classification performance of the classification model based on the classification results from the AI ​​algorithm 62, namely TP, FN, FP, and TN.

[0251] In step S53, the control unit 31 predicts the TKI treatment response of the target patient (whether the target patient is an early responder or a late responder) based on the calculated indicator (for example, by comparing the F1 score with a threshold (cutoff value)).

[0252] Then, in step S54, the control unit 31 outputs (for example, displays) a cell analysis results screen including the predicted TKI treatment response to the display unit 34. The TKI treatment response may include, for example, the F1 score.

[0253] <Effects of the 4th Embodiment> According to the method of this embodiment, optical information of cells is obtained by irradiating cells contained in a sample with multiple diffracted lights generated by incident light on a diffracting optical element. The sample includes a first sample taken from a patient with chronic myeloid leukemia before the start of treatment (for example, a sample taken from a patient with chronic myeloid leukemia before the start of treatment (target patient)) and a second sample taken from a healthy person. A classification model for classifying leukemia cells is generated based on the optical information obtained from the first and second samples, an index regarding the classification performance of leukemia cells and normal cells by the generated classification model is obtained, and information regarding the efficacy of a chronic myeloid leukemia treatment drug for the patient is output based on the index. By using optical information obtained from a first sample taken from a patient with chronic myeloid leukemia before the start of treatment and a second sample taken from a healthy individual, a classification model for classifying leukemia cells can be easily generated. Then, an index regarding the classification performance of the generated classification model between leukemia cells and normal cells can be obtained, and based on the obtained index, information on the efficacy of chronic myeloid leukemia treatment drugs in the aforementioned patients with chronic myeloid leukemia before the start of treatment can be output, making it possible for attending physicians and others to review the information.

[0254] In this case, the treatment drug for chronic myeloid leukemia may also be a tyrosine kinase inhibitor (TKI).

[0255] <Modified form of the fourth embodiment> While tyrosine kinase inhibitors, a type of molecularly targeted drug, were given as an example of a treatment for chronic myeloid leukemia, other molecularly targeted drugs may also be used. Furthermore, although molecularly targeted therapy is currently often used as standard treatment, other treatment methods may also be applied, and these drugs may be selected as part of those treatment options.

[0256] <Other Embodiments> In the above embodiment, a single rectilinear illumination light having a single beam spot BS irradiates the cells flowing through the flow cell, but a plurality of rectilinear illumination lights having a single beam spot may irradiate the cells flowing through the flow cell. That is, in the optical measurement unit 100 shown in FIG. 19, other rectilinear illumination light having a single beam spot based on light from another light source may irradiate the cells flowing through the flow cell 101. Also, in the optical measurement unit 400 shown in FIG. 29, other rectilinear illumination light having a single beam spot based on light from another light source may irradiate the cells flowing through the flow cell 201. The wavelength of the light irradiated from the other light source is preferably different from the wavelength of the light irradiated from the light source 111 or the light source 211.

[0257] In the above embodiment, the diffractive optical element 215 may have a condensing effect. In this case, for example, the diffraction pattern itself formed on the diffractive optical element 215 may have a condensing effect, a diffraction pattern for generating diffracted light may be formed on the incident surface of the diffractive optical element 215, and a pattern having a lens effect or a Fresnel lens may be formed on the exit surface of the diffractive optical element 215. Also, when the diffractive optical element 215 has a condensing effect, the condensing lens 216 may be omitted.

[0258] In the above embodiment, the diffractive optical element 215 is a transmissive diffractive optical element, but it may be a reflective diffractive optical element.

[0259] In the above embodiment, the arithmetic unit 32 of the control unit 30 classifies the cells by the AI algorithm 62 based on the detection signals of the light receiving units 225, 233, and 243. However, the present invention is not limited to this, and the cells may be classified by comparing the pattern of the detection signals of the light receiving units 225, 233, and 243 with the pattern stored in the storage unit 33 in advance.

[0260] In the above embodiment, count values ​​and abnormal cell flags related to CML cells were obtained during the GCM measurement process. However, count values ​​and abnormal cell flags related to cells other than CML cells, such as neutrophils, normal lymphocytes, monocytes, eosinophils, basophils, blast cells, abnormal lymphocytes, atypical lymphocytes, immature granulocytes, and nucleated erythrocytes, may also be obtained.

[0261] In the above embodiment, the sample was blood, but it is not limited to blood, and other bodily fluids may be used.

[0262] In the above embodiment, the GCM measurement process was performed when the FCM analysis result met the predetermined conditions shown in step S12. However, the embodiment is not limited to this, and the GCM measurement process may be performed regardless of the FCM analysis result.

[0263] In the above embodiment, the control unit 31 of the control unit 30 may, if the FCM analysis result meets the predetermined conditions shown in step S12, place the sample container 51 measured by the FCM measurement unit 10 at the sample intake position (sample supply position) by the GCM measurement unit 20, and if the FCM analysis result does not meet the predetermined conditions, control the transport unit 40 so that the sample container 51 passes through the GCM measurement unit 20.

[0264] Embodiments of the present invention can be modified in various ways as appropriate within the scope of the technical idea set forth in the claims. [Explanation of Symbols]

[0265] 1(1A,1B,1C) Sample analyzer 10 FCM measurement units 20 GCM measurement units 30 Control Units 40 Conveyor Units 50 sample racks

Claims

1. A measurement unit that obtains optical information of cells by irradiating cells contained in a sample with multiple diffracted lights generated when light is incident on a diffracting optical element, A sample analyzer comprising: an analysis unit that obtains information on leukemia cells contained in the sample by analyzing the optical information obtained by the measurement unit using an artificial intelligence algorithm; and a sample analyzer.

2. The sample analyzer according to claim 1, wherein the analysis unit obtains the percentage of leukemia cells among white blood cells as information on leukemia cells.

3. The sample analyzer according to claim 1, wherein the analysis unit obtains the number of leukemia cells as information on leukemia cells.

4. The system further includes a sample preparation unit for mixing the sample and reagent to prepare a measurement sample. The sample analysis apparatus according to claim 1, wherein the sample preparation unit prepares a measurement sample in which red blood cells contained in the sample are hemolyzed by using a hemolytic agent as the reagent.

5. The specimen analyzer according to claim 4, wherein the reagent does not contain a staining agent.

6. The measurement unit is configured to acquire information that identifies at least white blood cells among the cells contained in the sample. The specimen analyzer according to claim 1, wherein the analysis unit obtains the number or percentage of leukemia cells among the white blood cells identified based on the information based on the optical information.

7. The measurement unit is configured to acquire information that identifies at least granulocytes among the cells contained in the sample. The specimen analyzer according to claim 1, wherein the analysis unit obtains the number or percentage of leukemia cells among the granulocytes identified based on the information based on the optical information.

8. The specimen analyzer according to claim 6 or 7, wherein the information includes the intensity of scattered light obtained by irradiating cells with a single irradiation light.

9. The specimen analyzer according to claim 1, wherein the artificial intelligence algorithm is trained using optical information of leukocytes contained in a specimen taken from a patient with chronic myeloid leukemia as training data.

10. The specimen analyzer according to claim 1, wherein the artificial intelligence algorithm is trained using optical information of granulocytes contained in a specimen taken from a patient with chronic myeloid leukemia as training data.

11. It further comprises a first measurement unit that acquires measurement results regarding information about blood cells contained in the sample, The aforementioned measuring unit is a second measuring unit that is different from the first measuring unit. The specimen analyzer according to claim 1, wherein the second measuring unit performs a measurement on the specimen to obtain optical information of cells contained in the specimen when the measurement result of the first measuring unit on the specimen satisfies predetermined conditions.

12. The specimen analyzer according to claim 11, wherein the predetermined conditions include an increase in a specific type of white blood cell or the appearance of blast cells.

13. The specimen analyzer according to claim 12, wherein the aforementioned specific type of white blood cell is a basophil.

14. By injecting light into a diffractive optical element and irradiating the cells contained in the sample with multiple diffracted lights, the optical information of the cells is obtained. A sample analysis method that obtains information on leukemia cells contained in the sample by analyzing the obtained optical information using an artificial intelligence algorithm.

15. Optical information of the cells is obtained by irradiating cells contained in a sample with multiple diffracted lights generated by incident light on a diffracting optical element, and the sample includes a first sample taken from a patient with chronic myeloid leukemia before the start of treatment and a second sample taken from a healthy person. Based on the optical information obtained from the first and second samples, a classification model for classifying leukemia cells is generated. We obtained an index regarding the classification performance of the generated classification model between leukemia cells and normal cells. A method for outputting information regarding the efficacy of a drug for treating chronic myeloid leukemia in the patient based on the aforementioned indicators.

16. The method according to claim 15, wherein the therapeutic agent is a tyrosine kinase inhibitor.

Images