Sample analyzer and body fluid sample detection method

Abstract

The invention relates to a sample analyzer and a body fluid sample detection method. The method comprises the following steps: acquiring optical signal data of a to-be-detected liquid prepared based on a body fluid sample; the light signal data comprises at least two of a forward scattering light signal, a lateral scattering light signal and a fluorescence signal; generating a scatter diagram of the to-be-detected liquid according to the optical signal data; determining a scatter diagram type and / or an abnormal region in the scatter diagram based on the scatter diagram; the scatter diagram type comprises a normal type, a detection anomaly type and a quality anomaly type; under the condition that the scatter diagram type is abnormal, determining at least one of an auxiliary diagnosis result and a sample quality inspection result of the body fluid sample according to the scatter features of the abnormal region and the scatter diagram type; the scatter features are used for representing the particle distribution condition of the abnormal region. According to the method, by determining the abnormal area of the scatter diagram, targeted recognition and judgment can be performed according to the scatter features in the abnormal area of the scatter diagram, and the sensitivity and specificity of an auxiliary diagnosis result and / or a sample quality inspection result are / is improved.

Description

Technical Field

[0001] This application relates to the field of in vitro diagnostic technology, and in particular to a sample analyzer and a method for detecting body fluid samples. Background Technology

[0002] Body fluid analysis is crucial for screening and diagnosing patients. Currently, clinically used body fluid analysis methods mainly include routine body fluid analysis, biochemical body fluid analysis, and tumor marker body fluid analysis. However, routine body fluid analysis and biochemical body fluid analysis typically use gating to identify specific cell particles for analysis, such as identifying tumor cells by designating high-fluorescence regions. However, cells with high fluorescence signals include not only tumor cells but also normal cells such as mesothelial cells and macrophages. In particular, mesothelial cells have the greatest impact on the detection of tumor cells, resulting in relatively low sensitivity and specificity of body fluid analysis. On the other hand, tumor marker body fluid analysis is costly and relies heavily on the diagnostic experience of healthcare professionals.

[0003] Therefore, how to achieve high sensitivity, specificity, and accessibility in body fluid testing is an urgent problem to be solved. Summary of the Invention

[0004] Therefore, it is necessary to provide a sample analyzer and a method for detecting body fluids that can ensure the sensitivity and specificity of body fluid detection, in response to the above-mentioned technical problems.

[0005] In a first aspect, this application provides a sample analyzer, the sample analyzer comprising:

[0006] A sampling component for obtaining bodily fluid samples from a sample container;

[0007] A sample preparation assembly has at least one reaction chamber and a reagent supply unit; wherein the at least one reaction chamber is used to receive the body fluid sample acquired by the sampling assembly; the reagent supply unit provides a hemolytic reagent and a fluorescent reagent to the at least one reaction chamber, thereby mixing the body fluid sample with the hemolytic reagent and the fluorescent reagent in the reaction chamber to prepare a test solution;

[0008] An optical detection component is used to irradiate the test liquid flowing through the detection area with light and collect various light signal data generated by the cells due to light irradiation.

[0009] A memory, wherein the memory stores a computer program;

[0010] A processor, connected to the memory, is configured to perform the following steps when executing the computer program:

[0011] The optical detection component acquires optical signal data of the test liquid; the optical signal data includes at least two of forward-scattered light signals, side-scattered light signals, and fluorescence signals; a scatter plot of the test liquid is generated based on the optical signal data; the scatter plot type and / or abnormal regions in the scatter plot are determined based on the scatter plot; the scatter plot type includes normal type, detection abnormal type, and quality abnormal type; if the scatter plot type is not the normal type, at least one of the auxiliary diagnostic result and sample quality inspection result of the body fluid sample is determined based on the scatter characteristics of the abnormal region and the scatter plot type; the scatter characteristics are used to characterize the particle distribution of the abnormal region.

[0012] Secondly, this application also provides a method for detecting bodily fluid samples, the method comprising:

[0013] Acquire optical signal data of the test liquid prepared based on a body fluid sample; the optical signal data includes at least two of forward-scattered light signals, side-scattered light signals, and fluorescence signals;

[0014] A scatter plot of the test liquid is generated based on the optical signal data;

[0015] The scatter plot type and / or abnormal regions in the scatter plot are determined based on the scatter plot; the scatter plot type includes normal type, detection abnormal type and quality abnormal type;

[0016] If the scatter plot type is not the normal type, at least one of the auxiliary diagnostic result and sample quality inspection result of the body fluid sample is determined based on the scatter plot characteristics of the abnormal region and the scatter plot type; the scatter plot characteristics are used to characterize the particle distribution of the abnormal region.

[0017] In one embodiment, determining the scatter plot type and / or outlier regions in the scatter plot based on the scatter plot includes:

[0018] The scatter plot is input into an AI model for recognition to obtain the scatter plot type.

[0019] In the case where the scatter plot type is not the normal type, a distribution feature map is obtained based on the identification information of the scatter plot type in the AI ​​model;

[0020] The region of interest is selected from the distribution feature map as the abnormal region.

[0021] In one embodiment, when the scatter plot type is not the normal type, obtaining the distribution feature map based on the scatter plot type includes:

[0022] Based on the scatter plot type, the feature layer and predicted values ​​are obtained through forward propagation;

[0023] Backpropagation is performed on the predicted values ​​of the scatter plot type to obtain the gradient information returned to the feature layer; the gradient information characterizes the sensitivity of the channels in the feature layer to the scatter plot type;

[0024] Based on the gradient information of each channel in the feature layer, determine the weight vector corresponding to each channel;

[0025] The distribution feature map is obtained by weighting and summing the weight vector with the corresponding channel.

[0026] In one embodiment, determining at least one of the auxiliary diagnostic results and sample quality inspection results of the body fluid sample based on the scatter plot characteristics of the abnormal region and the scatter plot type includes:

[0027] If the scatter plot type is the detected anomaly type, the body fluid type of the body fluid sample is obtained;

[0028] Based on the abnormal region, obtain the scatter features that match the body fluid type;

[0029] Based on the scatter plot characteristics that match the body fluid type, it is determined whether the auxiliary diagnostic result indicates an abnormality in the test result of the body fluid sample.

[0030] In one embodiment, determining whether the auxiliary diagnostic result indicates an abnormality in the test result of the body fluid sample based on the scatter plot characteristics matching the body fluid type further includes:

[0031] Obtain particle classification information based on the scatter plot;

[0032] Based on at least one of the following: target particles matching the body fluid type in the particle classification information, the difference information between the scatter feature and the particle classification information, and the scatter feature matching the body fluid type, a preset diagnostic function is used for matching. If the matching is successful, the auxiliary diagnostic result is determined to be that the detection result is abnormal, and the disease diagnosis information corresponding to the diagnostic function is output.

[0033] In one embodiment, determining at least one of the auxiliary diagnostic results and sample quality inspection results of the body fluid sample based on the scatter plot characteristics of the abnormal region and the scatter plot type includes:

[0034] Obtain particle classification information based on the scatter plot;

[0035] When the scatter plot type is the quality anomaly type, at least one of the scatter features and particle classification information, and the scatter features that match the body fluid type, is matched with a preset quality discrimination function. If the match is successful, the sample quality inspection result is determined to be that the body fluid sample has an abnormal sample quality.

[0036] In one embodiment, the method further includes:

[0037] If the auxiliary diagnostic result indicates that the test result of the body fluid sample is abnormal, disease diagnostic information matching the body fluid type of the body fluid sample is output based on the auxiliary diagnostic result.

[0038] If the sample quality inspection result indicates that the body fluid sample has abnormal quality, a prompt message will be output based on the sample quality inspection result; the prompt message is used to indicate that the body fluid sample is not properly stored and / or prepared.

[0039] In one embodiment, the training process of the AI ​​model includes:

[0040] Obtain a training sample set; the training sample set includes scatter plots of normal samples, scatter plots of samples with detected anomalies, and scatter plots of samples with abnormal quality.

[0041] Based on the anomaly type, the scatter plots of the non-normal samples in the training sample set are divided into patterns, and combined with the scatter plots of the normal samples, to obtain the divided training sample set.

[0042] The initial model is trained using the partitioned training sample set. When the initial model meets the iteration termination condition, the trained AI model is obtained.

[0043] Thirdly, this application also provides a method for detecting bodily fluid samples, the method comprising:

[0044] Acquire optical signal data of the test solution prepared based on a body fluid sample; the optical signal data includes at least two of forward-scattered light signals, side-scattered light signals, and fluorescence signals;

[0045] A scatter plot of the test liquid is generated based on the optical signal data;

[0046] Based on the scatter plot, identify the abnormal regions in the scatter plot;

[0047] Based on the scatter plot characteristics of the abnormal region, the auxiliary diagnostic results of the body fluid sample are determined; wherein, the scatter plot characteristics are used to characterize the particle distribution in the abnormal region.

[0048] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the body fluid sample detection method described in any embodiment of this application.

[0049] In the aforementioned sample analyzer and body fluid sample detection method, by determining the scatter plot type and abnormal regions within the scatter plot, and based on the scatter plot type and the scatter characteristics of the abnormal regions, at least one of the auxiliary diagnostic results and sample quality inspection results for the body fluid sample is determined. On one hand, by classifying scatter plot types, targeted identification and judgment can be performed based on the scatter plot type, improving the sensitivity and specificity of auxiliary diagnostic results and / or sample quality inspection results. On the other hand, identifying abnormal regions within the scatter plot and judging them based on their scatter characteristics, compared to the existing technology of delineating specific cell particles through gating, can eliminate interference from other cell particles, further improving the sensitivity and specificity of body fluid detection. Furthermore, by identifying and judging only the scatter characteristics of abnormal regions, there is no need to process the scatter characteristics of all regions of the scatter plot, improving processing efficiency, and resulting in more refined and accurate auxiliary diagnostic results and / or sample quality inspection results. This method is relatively low-cost and widely applicable, thus achieving rapid and low-cost body fluid detection. Attached Figure Description

[0050] Figure 1 This is an application environment diagram illustrating a body fluid sample detection method according to an exemplary embodiment;

[0051] Figure 2 This is a schematic flowchart illustrating a method for detecting bodily fluid samples according to an exemplary embodiment;

[0052] Figure 3A This is a schematic diagram of a scatter plot according to an exemplary embodiment;

[0053] Figure 3B This is a schematic diagram of a scatter plot according to an exemplary embodiment;

[0054] Figure 3C This is a schematic diagram of a scatter plot according to an exemplary embodiment;

[0055] Figure 4 This is a schematic diagram of a three-dimensional scatter plot according to an exemplary embodiment;

[0056] Figure 5 This is a schematic diagram of a scatter plot according to an exemplary embodiment;

[0057] Figure 6 This is a schematic diagram of a scatter plot according to an exemplary embodiment;

[0058] Figure 7 This is a schematic diagram of a heat map according to an exemplary embodiment;

[0059] Figure 8 This is a schematic diagram of a region of interest in a heat map according to an exemplary embodiment;

[0060] Figure 9 This is a schematic diagram of anomaly regions in a scatter plot according to an exemplary embodiment;

[0061] Figure 10 This is a scatter plot and schematic diagram of abnormal regions of the test liquid according to an exemplary embodiment;

[0062] Figure 11 This is a scatter plot and schematic diagram of abnormal regions of the test liquid according to an exemplary embodiment;

[0063] Figure 12 This is a structural block diagram of a body fluid sample detection device according to an exemplary embodiment;

[0064] Figure 13 This is an internal structural diagram of a computer device according to an exemplary embodiment. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0066] The terms "first," "second," and "third" used in the embodiments of this application are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first," "second," or "third" may explicitly or implicitly include at least one of that feature. In the description of this application, "at least one" is used to indicate one or more; "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0067] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0068] This application provides a sample analyzer, the sample analyzer comprising:

[0069] A sampling component for obtaining bodily fluid samples from a sample container;

[0070] A sample preparation assembly has at least one reaction chamber and a reagent supply unit; wherein the at least one reaction chamber is used to receive the body fluid sample acquired by the sampling assembly; the reagent supply unit provides a hemolytic reagent and a fluorescent reagent to the at least one reaction chamber, thereby mixing the body fluid sample with the hemolytic reagent and the fluorescent reagent in the reaction chamber to prepare a test solution;

[0071] An optical detection component is used to irradiate the test liquid flowing through the detection area with light and collect various light signal data generated by the cells due to light irradiation.

[0072] A memory, wherein the memory stores a computer program;

[0073] A processor, connected to the memory, is configured to perform the following steps when executing the computer program:

[0074] The optical detection component acquires optical signal data of the test liquid; the optical signal data includes at least two of forward-scattered light signals, side-scattered light signals, and fluorescence signals; a scatter plot of the test liquid is generated based on the optical signal data; the scatter plot type and / or abnormal regions in the scatter plot are determined based on the scatter plot; the scatter plot type includes normal type, detection abnormal type, and quality abnormal type; if the scatter plot type is not the normal type, at least one of the auxiliary diagnostic result and sample quality inspection result of the body fluid sample is determined based on the scatter characteristics of the abnormal region and the scatter plot type; the scatter characteristics are used to characterize the particle distribution of the abnormal region.

[0075] In this embodiment, the sample analyzer is applied in the field of in vitro assisted diagnostics. For example, the sample analyzer can be a human cell analyzer, an animal cell analyzer, a point-of-care testing device, etc. The function of the sample analyzer is to analyze the test solution. In related technologies, the sample analyzer is equipped with interactive components, such as a touch screen, or a non-touch screen, keyboard, and mouse.

[0076] In this embodiment, the optical detection component may include a scattered light detector and a fluorescence detector. The scattered light detector may be a forward-scattering light detector for detecting forward-scattered light or a side-scattering light detector; the fluorescence detector is used to detect the fluorescence signal generated by particles in the test liquid passing through the detection area after being irradiated by light.

[0077] Optionally, the body fluid sample detection method in this application embodiment can be applied to a sample analyzer or to the software program of a blood cell analyzer.

[0078] Optionally, the body fluid sample detection method provided in this application embodiment can be applied to, for example, Figure 1 In the application environment shown, the sample analyzer 10 is communicatively connected to a cloud server / local server 20. The sample analyzer acquires the optical signal data of the test liquid through an optical detection component and transmits it to the cloud server / local server 20. The cloud server / local server 20 generates a scatter plot of the test liquid based on the optical signal data; determines the scatter plot type and abnormal regions in the scatter plot based on the scatter plot; and, in the case where the scatter plot type is abnormal, determines at least one of the auxiliary diagnostic results and sample quality inspection results of the body fluid sample based on the scatter characteristics of the abnormal regions and the scatter plot type.

[0079] In some embodiments, such as Figure 2 As shown, a method for detecting bodily fluid samples is provided, which can be applied to... Figure 1 Taking the sample analyzer 10 as an example, the following steps are included:

[0080] S201, acquire optical signal data of the test liquid prepared based on the body fluid sample; the optical signal data includes at least two of forward scattered light signal, side scattered light signal and fluorescence signal.

[0081] In this embodiment, the test solution can be obtained by reacting a body fluid sample with a reagent. The body fluid sample may include, but is not limited to, at least one of body cavity fluid, cerebrospinal fluid, pleural effusion, ascites, cardiac sac fluid, or synovial fluid.

[0082] In this embodiment, the optical signal data can be obtained by illuminating the test liquid with light as it flows through the detection area using an optical detection component. The optical signal data can be pulse signal data directly output by the optical detection component, or it can be data that has undergone filtering, amplification, and / or noise reduction processing.

[0083] In this embodiment, the forward scattering (FSC) signal indicates the scattered light signal generated along the laser propagation direction when a laser beam irradiates a cell or particle; the forward scattering signal can reflect the number and volume of particles. The side scattering (SSC) signal reflects the complexity of the cell's internal structure (such as intracellular particles or the cell nucleus). The fluorescence (FL) signal reflects the content of nucleic acid material in the cell. These light signals can be used to classify and count particles in the test solution.

[0084] In some embodiments, when a laser beam irradiates the test liquid in the optical detection assembly, the light is scattered forward at a relatively small angle (e.g., 0.5° to 10°), generating a forward-scattered light signal. The intensity of the forward-scattered light signal is proportional to the size of the cell and is therefore often used to detect the surface properties of cells or other particles.

[0085] In some embodiments, when a laser beam irradiates the test liquid in the optical detection component, a portion of the light is scattered at a 90° angle, generating a side-scattered light signal. The intensity of the side-scattered light signal is mainly related to the cell density and particle size refraction.

[0086] S202, Generate a scatter plot of the test liquid based on the optical signal data.

[0087] In some embodiments, the scatter plot can be a graph / matrix drawn from at least two of the forward-scattered light signal, the side-scattered light signal, and the fluorescence signal. The scatter plot is image data characterizing the distribution characteristics of different types of particles in the test liquid.

[0088] Optionally, the scatter plot may include, but is not limited to, at least one of two-dimensional scatter plots and three-dimensional scatter plots.

[0089] Optionally, the scatter plot includes, but is not limited to, at least one of DIFF (Differential Count) scatter plot and WNR (White Blood Cell, Nucleated Red Blood Cell) scatter plot.

[0090] For example, a scatter plot of the same test liquid in different dimensions is shown in the figure; wherein, Figure 3A A scatter plot drawn based on two types of optical signal data: side-scattered light signal and fluorescence signal; Figure 3B A scatter plot drawn from two types of optical signal data: forward-scattered light signal and side-scattered light signal; Figure 3C This is a scatter plot drawn based on two types of optical signal data: forward-scattered light signal and fluorescence signal.

[0091] For example, such as Figure 4 As shown, Figure 4 This is a three-dimensional scatter plot drawn from three types of optical signal data: forward-scattered light signal, side-scattered light signal, and fluorescence signal.

[0092] In some embodiments, different axes of a scatter plot represent different measurement parameters. For example, volume can reflect cell size, commonly represented by FSC; larger cells scatter more light. Intracellular particle complexity can be measured by SSC, indicating the density of internal cell structures or particles, such as the complexity of the cell nucleus or particle content. SFL can reflect the content of fluorescent substances within the cell. For example, it reflects the content of specially fluorescently labeled RNA (ribonucleic acid) or DNA (deoxyribonucleic acid) within the cell.

[0093] S203, determine the scatter plot type and / or abnormal regions in the scatter plot based on the scatter plot; the scatter plot type includes normal type, detection abnormal type and quality abnormal type.

[0094] In this embodiment, the detection anomaly type can indicate the scatter plot type corresponding to the body fluid sample formed when the target object's organ / tissue or other structure becomes diseased or even develops a malignant tumor. The scatter plot of the test fluid for detecting the anomaly type may contain particles with abnormal distribution patterns, abnormal distribution areas, and / or particles that may exist in abnormal states, such as promyelocytes, platelet-dependent cells, malignant tumor cells, etc.

[0095] For example, the types of abnormalities detected may include, but are not limited to, abnormal types of plasma cells, abnormal types of cancer cells, or abnormal types of macrophage cells.

[0096] In this embodiment, the quality anomaly type can indicate the scatter plot type corresponding to a body fluid sample that has excessive mucus, impurities, or particulate matter due to improper storage or processing; the quality anomaly type characterizes that the corresponding body fluid sample itself has an abnormal quality, i.e., the sample quality anomaly type.

[0097] For example, such as Figure 5 As shown, Figure 5 This is a schematic diagram of a scatter plot of a normal type of test solution.

[0098] In this embodiment, the abnormal region can indicate an area where the distribution of the particle population is abnormal. For example, in a leukocyte scatter plot, the appearance of upward-pointing, non-dispersing gray scatter signals in the lymphocyte region can help indicate an abnormality in the lymphatic system, and this region can be considered an abnormal region.

[0099] In some embodiments, the sample analyzer can identify scatter plots using an AI model (e.g., a deep learning image classification model) to determine the scatter plot type and any abnormal regions. If the scatter plot type is normal, it can be determined that the body fluid sample corresponding to the test solution is a normal sample with no abnormal regions. If the scatter plot type is an abnormal detection type and / or an abnormal quality type, both of which are abnormal, it can be preliminarily determined that the body fluid sample corresponding to the test solution is abnormal and requires further assessment and processing.

[0100] S204, if the scatter plot type is not the normal type, determine at least one of the auxiliary diagnostic result and the sample quality inspection result of the body fluid sample based on the scatter plot characteristics of the abnormal region and the scatter plot type; the scatter plot characteristics are used to characterize the particle distribution of the abnormal region.

[0101] In the embodiments of this application, the scattered features may include, but are not limited to, the total number of particles, the centroid of particles, the distribution width / angle of particles in the forward scattering direction, the distribution width / angle of particles in the side scattering direction, and the distribution width / angle of particles in the fluorescence direction.

[0102] In some embodiments, when the scatter plot type is an abnormality detection type and / or a quality abnormality type, the abnormality of the test solution can be further accurately determined based on the scatter plot characteristics of the abnormal region. For example, if the scatter plot type corresponding to the first test solution is an abnormality detection type, and the number of immature granulocytes in the abnormal region is significantly increased, it may indicate abnormal proliferation in the bone marrow, which is characteristic of tumor cells, thus confirming that the first test solution is abnormal. In this way, on the one hand, normal test solutions can be initially screened out, eliminating the need to judge all test solutions to determine auxiliary diagnostic results and / or sample quality inspection results, saving resources; on the other hand, the abnormality of non-normal test solutions can be further accurately determined based on the scatter plot characteristics of the abnormal region.

[0103] The aforementioned sample analyzer and body fluid sample detection method determine at least one of the following: auxiliary diagnostic results and sample quality inspection results, by identifying the scatter plot type and abnormal regions within the scatter plot, and based on the scatter plot type and the scatter characteristics of the abnormal regions. On one hand, by classifying scatter plot types, targeted identification and judgment can be performed based on the scatter plot type, improving the sensitivity and specificity of auxiliary diagnostic results and / or sample quality inspection results. On the other hand, identifying abnormal regions within the scatter plot and judging them based on their scatter characteristics, compared to the existing technology of delineating specific cell particles through gating, can eliminate interference from other cell particles, further improving the sensitivity and specificity of body fluid detection. Furthermore, by identifying and judging only the scatter characteristics of abnormal regions, there is no need to process the scatter characteristics of all regions of the scatter plot, improving processing efficiency and resulting in more refined and accurate auxiliary diagnostic results and / or sample quality inspection results. This method is relatively low-cost and widely applicable, thus achieving rapid and low-cost body fluid detection.

[0104] In some embodiments, determining the scatter plot type and / or outlier regions in the scatter plot based on the scatter plot includes:

[0105] The scatter plot is input into an AI model for recognition to obtain the scatter plot type.

[0106] In the case where the scatter plot type is not the normal type, a distribution feature map is obtained based on the identification information of the scatter plot type in the AI ​​model;

[0107] The region of interest is selected from the distribution feature map as the abnormal region.

[0108] In this embodiment, the AI ​​(Artificial Intelligence) model can be any model that implements image classification and anomaly region localization. For example, it can be a CNN (Convolutional Neural Network) model.

[0109] In this embodiment, the region of interest (ROI) indicates the image region that the AI ​​model focuses on when making a specific classification decision. For example, the ROI could be a region in a scatter plot where anomalous particles are present.

[0110] In this embodiment of the application, the distribution feature map may include, but is not limited to, at least one of a heat map and a two-dimensional histogram.

[0111] In some embodiments, the sample analyzer inputs multiple scatter plots into an AI model for identification, obtaining multiple scatter plot types corresponding to the scatter plots. The scatter plot types may include normal types corresponding to normal body fluid samples and abnormal types corresponding to abnormal body fluid samples. Abnormal body fluid samples may include scatter plots of samples with detected abnormalities and scatter plots of samples with abnormal quality. The sample analyzer can divide and label abnormal areas in the scatter plots corresponding to abnormal body fluid samples according to tumor type, other cell abnormality types, or disease types (e.g., lung cancer, breast cancer, gastric cancer, and oral cancer).

[0112] For example, the sample analyzer inputs a scatter plot of the test liquid into the AI ​​model, as shown in the scatter plot. Figure 6 As shown; the AI ​​model infers the predicted scatter plot type; based on the scatter plot type, the Grad-CAM algorithm is used to obtain the corresponding heatmap, as shown in the image. Figure 7 As shown; the Region of Interest (ROI) is determined by heatmap calculation and outlined with a rectangular dashed box, as shown in the figure. Figure 8 As shown; determine the outlier regions in the scatter plot based on the region of interest, as shown in the figure. Figure 9 As shown.

[0113] In some embodiments, obtaining a distribution feature map based on the scatter plot type when the scatter plot type is not the normal type includes:

[0114] Based on the scatter plot type, the feature layer and predicted values ​​are obtained through forward propagation;

[0115] Backpropagation is performed on the predicted values ​​of the scatter plot type to obtain the gradient information returned to the feature layer; the gradient information characterizes the sensitivity of the channels in the feature layer to the scatter plot type;

[0116] Based on the gradient information of each channel in the feature layer, determine the weight vector corresponding to each channel;

[0117] The distribution feature map is obtained by weighting and summing the weight vector with the corresponding channel.

[0118] In some embodiments, the sample analyzer can generate a visual heatmap using the Grad-CAM (Gradient-weighted Class Activation Mapping) algorithm. The specific process includes: inputting a scatter plot of the test liquid into the AI ​​model to obtain the scatter plot type; determining the feature layer and predicted values ​​through forward propagation based on the scatter plot type; performing backpropagation on the predicted values ​​to obtain gradient information for each channel in the feature layer; performing global average pooling on the gradient information of each channel to obtain the weight vector corresponding to each channel; performing weighted summation on the feature maps of the convolutional layer using the weight vectors, and applying the ReLU activation function to obtain the Grad-CAM visual heatmap; and identifying the region of interest in the heatmap as an anomaly region.

[0119] In this embodiment, the distribution feature map can intuitively display abnormal areas in the data through changes in color or brightness, which helps to conduct targeted analysis and identification of abnormal areas, improve identification performance and processing efficiency, and make the auxiliary diagnostic results and / or sample quality inspection results more accurate.

[0120] In some embodiments, determining at least one of the auxiliary diagnostic results and sample quality inspection results of the body fluid sample based on the scatter plot characteristics of the abnormal region and the scatter plot type includes:

[0121] If the scatter plot type is the detected anomaly type, the body fluid type of the body fluid sample is obtained;

[0122] Based on the abnormal region, obtain the scatter features that match the body fluid type;

[0123] Based on the scatter plot characteristics that match the body fluid type, it is determined whether the auxiliary diagnostic result indicates an abnormality in the test result of the body fluid sample.

[0124] In this embodiment, the body fluid type may include, but is not limited to, cerebrospinal fluid, serous cavity fluid, synovial fluid, pleural effusion, ascites, and pericardial effusion.

[0125] In some embodiments, the abnormal regions and / or the scatter features within the abnormal regions in the scatter plots corresponding to different types of body fluid samples may be different. The sample analyzer can acquire scatter features in the abnormal regions that match the body fluid type of the body fluid sample, and determine whether the auxiliary diagnostic result of the body fluid sample indicates an abnormality in the test result based on the scatter features in the abnormal regions that match the body fluid type of the body fluid sample.

[0126] For example, the body fluid type is ascites. When the sample analyzer determines that the scatter plot type of the test fluid prepared based on the body fluid sample is an abnormal detection type, it can further determine whether the test result of the body fluid sample is abnormal based on the scatter plot characteristics that are compatible with ascites in the abnormal area, such as the total number of particles in the abnormal area, the distribution width and / or angle of particles in different directions, etc. If it is determined that the test result of the body fluid sample is abnormal, disease diagnosis information can be output based on the auxiliary diagnostic results to warn the operator.

[0127] In some embodiments, determining whether the auxiliary diagnostic result indicates an abnormality in the test result of the body fluid sample based on the scatter plot characteristics matching the body fluid type further includes:

[0128] Obtain particle classification information based on the scatter plot;

[0129] Based on at least one of the following: target particles matching the body fluid type in the particle classification information, the difference information between the scatter feature and the particle classification information, and the scatter feature matching the body fluid type, a preset diagnostic function is used for matching. If the matching is successful, the auxiliary diagnostic result is determined to be that the detection result is abnormal, and the disease diagnosis information corresponding to the diagnostic function is output.

[0130] In this embodiment of the application, particle classification information may include, but is not limited to, at least one of particle type information and particle count information.

[0131] In this embodiment, particle type information can indicate the cell types present in the test solution; particle type information may include, but is not limited to, lymphocytes, neutrophils, monocytes / macrophages, eosinophils, erythrocytes, basophils, and mesothelial cells.

[0132] In the embodiments of this application, particle count information can indicate the number and / or percentage of different types of cells.

[0133] In some embodiments, the sample analyzer can classify and identify each particle in the scatter plot using clustering models, classification models, or annotation software to distinguish different types of particles in the scatter plot and obtain particle classification information.

[0134] In some embodiments, after obtaining particle classification information, the sample analyzer can further label the particles of each type in the scatter plot according to the particle classification information, and display the labeled scatter plot on the display interface of the sample analyzer. Labeling methods include, but are not limited to, symbol labels, text labels, and color labels. For example, red can be used to label particles of type A, and blue to label particles of type B; or an asterisk can be used to label particles of type A, and a circle can be used to label particles of type B; or a label "A" can be placed near particles of type A, and a label "B" can be placed near particles of type B. By labeling the particles of each type in the scatter plot, users can more intuitively distinguish between the particles.

[0135] In some embodiments, the difference information between scatter plot features and particle classification information can be the difference information between all scatter plot features of the abnormal region in the scatter plot and all particle classification information of the scatter plot; or, it can be the difference information between scatter plot features in the abnormal region that match the body fluid type and all particle classification information of the scatter plot; or, it can be the difference information between all scatter plot features of the abnormal region in the scatter plot and some particle classification information of the scatter plot; or, it can also be the difference information between scatter plot features in the abnormal region that match the body fluid type and some particle classification information of the scatter plot.

[0136] In some embodiments, the step of matching at least one of the target particles matching the body fluid type in the particle classification information, the difference information between the scatter features and the particle classification information, and the scatter features matching the body fluid type with a preset diagnostic function, and determining the auxiliary diagnostic result as an abnormality in the detection result if the match is successful, includes:

[0137] Based on at least one of the following: target particles matching the body fluid type, the total number of particles in the abnormal region, the distribution width of the abnormal region, and the area occupied by particles in the abnormal region, a match is made with a preset diagnostic function. If the match is successful, the auxiliary diagnostic result is determined to be that the detection result is abnormal.

[0138] For example, if the body fluid type is cerebrospinal fluid, synovial fluid, etc., the target particle can be identified as a red blood cell.

[0139] For example, if the body fluid type is ascites, the target particle can be identified as a mononuclear cell.

[0140] In some embodiments, if the sample analyzer determines that the number of target particles matching the body fluid type is greater than a first threshold, the total number of particles in the abnormal region is greater than a second threshold, the distribution width of the abnormal region is greater than a third threshold, and the area occupied by particles in the abnormal region is greater than a fourth threshold, the auxiliary diagnostic result can be determined that the detection result of the body fluid sample is abnormal.

[0141] In one embodiment, the target particles and the first to fourth thresholds can be determined based on the body fluid type and / or scatter plot type; the target particles and the first to fourth thresholds can be different for different types of test fluids.

[0142] For example, if the body fluid type is ascites, the sample analyzer can determine the target particle as a monocyte based on the body fluid type. The diagnostic function can be Y2 = f(A, B, C, T, MON-BF%), where T indicates the body fluid type; MON-BF% indicates the percentage of monocytes; A indicates the total number of particles in the abnormal region; B indicates the distribution width of the abnormal region; and C indicates the area occupied by particles in the abnormal region. If T = ascites && MON-BF% > thD, A > thA && B > thB && C > thC, that is, the percentage of monocytes is greater than 20%, the total number of particles in the abnormal region is greater than the second threshold, the distribution width of the abnormal region is greater than the third threshold, and the total area occupied by particles in the abnormal region is greater than the fourth threshold, then the auxiliary diagnostic result of the body fluid sample can be determined to be abnormal, which may indicate the presence of adenocarcinoma cells in the body fluid sample, requiring further confirmation of cell morphology; where thD indicates the first threshold; thA indicates the second threshold; thB indicates the third threshold; and thC indicates the fourth threshold.

[0143] For example, such as Figure 10 As shown, Figure 10 The diagram shows a scatter plot of the test solution and a schematic diagram of the abnormal area; the area enclosed by the black dashed box is the abnormal area.

[0144] In this embodiment, when the scatter plot type is determined to be an abnormality detection type, effective information or features can be selected specifically for identification and judgment based on the particle classification information matching the scatter plot type and the scatter features matching the body fluid type, thereby improving the accuracy and specificity of the auxiliary diagnostic results; and, it is not necessary to use all particle classification information and scatter features for identification and judgment, thus improving processing efficiency.

[0145] In some embodiments, determining at least one of the auxiliary diagnostic results and sample quality inspection results of the body fluid sample based on the scatter plot characteristics of the abnormal region and the scatter plot type includes:

[0146] Obtain particle classification information based on the scatter plot;

[0147] When the scatter plot type is the quality anomaly type, at least one of the scatter features and particle classification information, and the scatter features that match the body fluid type, is matched with a preset quality discrimination function. If the match is successful, the sample quality inspection result is determined to be that the body fluid sample has an abnormal sample quality.

[0148] In some embodiments, the particle classification information and scatter plot characteristics of abnormal regions used to determine the sample quality inspection results may differ for different types of body fluid samples. The sample analyzer can determine the sample quality inspection results of the test fluid based on particle classification information and scatter plot characteristics adapted to the body fluid type.

[0149] For example, the body fluid type is cerebrospinal fluid. When the sample analyzer determines that the scatter plot type of the test fluid is of abnormal quality, it can further determine whether the sample quality of the body fluid sample is abnormal based on the particle type information and / or particle count information of the scatter plot, as well as the total number of particles in the abnormal region that matches cerebrospinal fluid, the distribution width and / or angle in different directions, etc. If the sample quality of the body fluid sample is abnormal, it can output a prompt message based on the sample quality inspection results to prompt the operator.

[0150] In some embodiments, when the scatter plot type is the quality anomaly type, matching at least one of the scatter features and particle classification information, and the scatter features that match the body fluid type, with a preset quality discriminant function, and determining that the sample quality inspection result indicates that the body fluid sample has an abnormal sample quality when the match is successful, includes:

[0151] Based on at least one of the following: the ratio of the total number of particles in the abnormal region to the total number of particles in the scatter plot, and the sum of the pulse widths of each type of particle in the abnormal region, a preset quality discrimination function is used for matching. If the matching is successful, the sample quality inspection result is determined to be that the sample quality of the body fluid sample is abnormal.

[0152] In some embodiments, if the sample analyzer determines that the ratio of the total number of particles in the abnormal region to the total number of particles in the scatter plot is greater than a fifth threshold, and the sum of the pulse widths of each type of particle in the abnormal region is greater than a sixth threshold, the quality inspection result is determined to be that the sample quality of the body fluid sample is abnormal.

[0153] For example, the scatter plot and thermogram of the test liquid are as follows: Figure 11As shown in the figure, the area enclosed by the black dashed box corresponds to the abnormal area. The body fluid type is bronchoalveolar lavage fluid, and the quality discrimination function can be Y1 = f(A, B); where A indicates the ratio of the total number of particles in the abnormal area to the total number of particles in the scatter plot; B indicates the sum of the pulse widths of each type of particle in the abnormal area. If the sample analyzer determines that A > thA && B > thB; where thA indicates the fifth threshold; thB indicates the sixth threshold; that is, the ratio of the total number of particles in the abnormal area to the total number of particles in the scatter plot is greater than the fifth threshold, it can be determined that there are too many particles in the abnormal area; if it is determined that the sum of the pulse widths of each type of particle in the abnormal area is greater than the sixth threshold, and it is determined that the number of particles in the abnormal area is greater than the number of particles in the effective area of ​​the scatter plot excluding the abnormal area, it can be determined that the scatter plot is abnormal. This is because the quality of the body fluid sample itself is abnormal, and therefore it can be determined that the sample quality of the body fluid sample is abnormal.

[0154] In some embodiments, the sample analyzer inputs a scatter plot of the test fluid into a deep learning image classification model. If the scatter plot type is identified as an anomaly, it can preliminarily determine that the body fluid sample may have been insufficiently pretreated or improperly stored, resulting in an abnormality in the sample quality. When the scatter plot type is identified as an anomaly, the sample analyzer can accurately determine whether the sample quality of the body fluid sample is abnormal based on information such as the total number of particles in the abnormal region, the ratio of the total number of particles in the abnormal region to the total number of particles in the scatter plot, and the pulse width of particles in the abnormal region. If the sample quality of the body fluid sample is abnormal, the analyzer can output a prompt message based on the sample quality inspection results to remind the operator to re-pretreat or replace the body fluid sample.

[0155] In this embodiment of the application, when the scatter plot type is determined to be a quality abnormality type, the difference between the scatter features of the abnormal region matching the body fluid type and the particle classification information of the scatter plot can be used to accurately determine whether the quality of the body fluid sample itself is abnormal, thereby reducing the occurrence of inaccurate auxiliary diagnostic results due to abnormal body fluid sample quality.

[0156] In some embodiments, the method further includes:

[0157] If the auxiliary diagnostic result indicates that the test result of the body fluid sample is abnormal, disease diagnostic information matching the body fluid type of the body fluid sample is output based on the auxiliary diagnostic result.

[0158] If the sample quality inspection result indicates that the body fluid sample has abnormal quality, a prompt message will be output based on the sample quality inspection result; the prompt message is used to indicate that the body fluid sample is not properly stored and / or prepared.

[0159] In some embodiments, disease diagnosis information and prompts may be output in tabular, list, graphical, or audio format.

[0160] In some embodiments, disease diagnosis information and prompts may be output in text or voice form on the display component of the sample analyzer; or, they may be sent to a third-party terminal / platform for output on the third-party terminal / platform.

[0161] In some embodiments, when the auxiliary diagnostic result indicates that the test result of the body fluid sample is abnormal, the sample analyzer can output the auxiliary diagnostic result of the body fluid sample, scatter plot, distribution feature plot, and disease diagnostic information matching the body fluid type of the body fluid sample.

[0162] For example, if the body fluid type is cerebrospinal fluid (CSF), the disease diagnostic information matching CSF could be information about potential central nervous system diseases such as infection, tumors, and / or autoimmune encephalopathy in the target individual corresponding to CSF. If the body fluid type is ascites, the disease diagnostic information matching ascites could be information about potential liver cirrhosis, heart failure, malignant tumors, etc., in the target individual corresponding to ascites.

[0163] For example, if the body fluid type is synovial fluid, and the sample analyzer determines that red blood cells are present in the synovial fluid, the disease diagnosis result can be determined as an abnormality in the test result of the body fluid sample. The disease diagnosis information matching the synovial fluid can be information such as arthritis, synovitis, or rheumatic diseases that the target subject corresponding to the synovial fluid may have.

[0164] In some embodiments, when the auxiliary diagnostic result indicates that the sample quality of the body fluid sample is abnormal, the sample analyzer can output the sample quality inspection results, scatter plot, distribution characteristic plot, and prompt information of the body fluid sample.

[0165] For example, if the body fluid type is cerebrospinal fluid (CSF) and the CSF sample quality is abnormal, it can be determined that the CSF sample may not have been pre-treated, resulting in excessive mucus, impurities, and / or particulate matter, leading to a large amount of blood shadow. The prompt message can be that the CSF sample preparation is unqualified, prompting the operator to prepare or replace the CSF sample.

[0166] In this embodiment, by outputting disease diagnosis information and / or prompt information, the auxiliary diagnostic results and / or sample quality inspection results can be determined intuitively, which makes it convenient for operators to directly determine the particle distribution, quality and abnormalities in body fluid samples, thereby enabling operators to take appropriate actions in a timely manner based on the disease diagnosis information and / or prompt information.

[0167] In some embodiments, the training process of the AI ​​model includes:

[0168] Obtain a training sample set; the training sample set includes scatter plots of normal samples, scatter plots of samples with detected anomalies, and scatter plots of samples with abnormal quality.

[0169] Based on the detected anomaly type, the scatter plots of non-normal samples in the training sample set are divided into patterns, and combined with the scatter plots of normal samples, to obtain the divided training sample set.

[0170] The initial model is trained using the partitioned training sample set. When the initial model meets the iteration termination condition, the trained AI model is obtained.

[0171] In some embodiments, the sample analyzer can divide and label abnormal regions in the scatter plots of abnormal samples, i.e., scatter plots of samples with detected abnormalities and scatter plots of samples with abnormal quality, according to tumor type or other cell abnormality type and disease type. For example, the scatter plots of abnormal samples can be divided according to lung cancer, breast cancer, gastric cancer and oral cancer, etc., to obtain scatter plots of abnormal samples with labeled abnormal types; and combined with the scatter plots of normal samples, a divided training sample set is obtained.

[0172] In some embodiments, the initial model may include, but is not limited to, a Residual Network (ResNet), a Visual Geometry Group (VGG), and a VisionTransformer (ViT). The sample analyzer trains the initial model using the partitioned training sample set until the loss function, such as the multi-class cross-entropy loss function, and the number of training iterations meet the iteration termination condition. At this point, the training is considered complete, and the trained AI model is obtained.

[0173] In this embodiment, an initial model is constructed and trained using a training sample set. During the training process, the model parameters are adjusted based on backpropagation and the loss function. When the initial model meets the iteration termination condition, an AI model is obtained. This can prevent overfitting during model training and improve the model's generalization ability on data.

[0174] In some embodiments, a method for detecting bodily fluid samples is provided, the method comprising:

[0175] Acquire optical signal data of the test liquid prepared based on a body fluid sample; the optical signal data includes at least two of forward-scattered light signals, side-scattered light signals, and fluorescence signals;

[0176] A scatter plot of the test liquid is generated based on the optical signal data;

[0177] Based on the scatter plot, identify the abnormal regions in the scatter plot;

[0178] Based on the scatter plot characteristics of the abnormal region, the auxiliary diagnostic results of the body fluid sample are determined; wherein, the scatter plot characteristics are used to characterize the particle distribution in the abnormal region.

[0179] In some embodiments, determining the outlier regions in the scatter plot based on the scatter plot includes:

[0180] The scatter plot is input into the AI ​​model for recognition to obtain the scatter plot type;

[0181] The distribution feature map is obtained based on the identification information of scatter plot type in the AI ​​model;

[0182] The regions of interest are selected as outliers in the distribution feature map.

[0183] In some embodiments, obtaining the distribution feature map based on the identification information of the scatter plot type in the AI ​​model includes:

[0184] Based on the scatter plot type, the feature layer and predicted values ​​are obtained through forward propagation;

[0185] Backpropagation is performed on the predicted values ​​of the scatter plot type to obtain the gradient information of the returned feature layer; the gradient information characterizes the sensitivity of the channels in the feature layer to the scatter plot type.

[0186] Based on the gradient information of each channel in the feature layer, determine the weight vector corresponding to each channel;

[0187] The distribution feature map is obtained by weighting and summing the weight vector with the corresponding channel.

[0188] In some embodiments, the sample analyzer can acquire scatter plot data; where the scatter plot data typically consists of a series of point coordinates, each point representing a data sample. Each data sample can contain two or more dimensions, for example, (x, y) coordinates or (x, y, z) coordinates. The sample analyzer aggregates the scatter plot data to convert the point density information into color information of a distribution feature map. Aggregation can be achieved by counting the number of points in each region or calculating the average value of points in each region. The aggregated data is then divided into small regions, forming a grid. The grid size can be adjusted according to the data distribution and visualization effect. Generally, the smaller the grid size, the richer the detail of the distribution feature map. Based on the aggregated data and grid division, various visualization tools or programming languages ​​can be used to generate distribution feature maps.

[0189] In some embodiments, determining the auxiliary diagnostic result of the body fluid sample based on the scatter plot characteristics of the abnormal region includes:

[0190] Obtain the body fluid type from the body fluid sample;

[0191] Scattered features matching body fluid type are obtained based on abnormal regions;

[0192] Based on the scatter plot characteristics matched with the body fluid type, determine whether the auxiliary diagnostic result indicates an abnormality in the test results of the body fluid sample.

[0193] In some embodiments, determining whether the auxiliary diagnostic result indicates an abnormality in the test result of the body fluid sample based on scatter features matching the body fluid type further includes:

[0194] Obtain particle classification information based on scatter plots;

[0195] Based on at least one of the following: target particles matching the body fluid type in particle classification information, differences between scatter features and particle classification information, and scatter features matching the body fluid type, a pre-defined diagnostic function is used for matching. If the matching is successful, the auxiliary diagnostic result is determined to be that the detection result is abnormal, and the disease diagnosis information corresponding to the diagnostic function is output.

[0196] The aforementioned method for detecting bodily fluid samples identifies abnormal regions in a scatter plot and determines auxiliary diagnostic results based on the scatter characteristics of these abnormal regions. On one hand, identifying and judging abnormal regions based on their scatter characteristics, compared to existing technologies that analyze specific cell particles using gating methods, eliminates interference from other cell particles, further improving the sensitivity and specificity of bodily fluid detection. On the other hand, by focusing on the scatter characteristics of abnormal regions, it eliminates the need to process the scatter characteristics of all regions in the scatter plot, improving processing efficiency and resulting in more refined and accurate auxiliary diagnostic results. This method is relatively low-cost and widely applicable, thus achieving rapid and low-cost bodily fluid detection.

[0197] This application also provides an application scenario in which the above-described scatter plot detection method is applied. Specifically, the scatter plot detection method is applied in this scenario as follows: a sample analyzer acquires optical signal data of a test liquid prepared from a body fluid sample; a scatter plot of the test liquid is generated based on the optical signal data; the scatter plot is input into an AI model for identification to determine the scatter plot type; in the case of an abnormal scatter plot type, a distribution feature map is obtained based on the identification information of the scatter plot type in the AI ​​model; and regions of interest are selected as abnormal regions from the distribution feature map.

[0198] When the scatter plot type is an anomaly detection type, the body fluid type of the body fluid sample and the particle classification information of the scatter plot are obtained; based on at least one of the following: target particles that match the body fluid type in the particle classification information, the difference information between the scatter features and the particle classification information, and scatter features that match the body fluid type, they are matched with a preset diagnostic function. If the match is successful, the auxiliary diagnostic result is determined to be that the detection result is abnormal, and the disease diagnosis information corresponding to the diagnostic function is output.

[0199] When the scatter plot type is of the quality anomaly type, at least one of the following is used: the difference information between the scatter plot features and the particle classification information, and the scatter plot features that match the body fluid type. The scatter plot features are matched with the preset quality discrimination function. If the match is successful, the sample quality inspection result is determined to be that the body fluid sample has an abnormality in quality. Based on the sample quality inspection result, a prompt message is output. The prompt message is used to indicate that the body fluid sample is not properly preserved and / or prepared.

[0200] The aforementioned sample analyzer and body fluid sample detection method determine at least one of the following: auxiliary diagnostic results and sample quality inspection results, by identifying the scatter plot type and abnormal regions within the scatter plot, and based on the scatter plot type and the scatter characteristics of the abnormal regions. On one hand, by classifying scatter plot types, targeted identification and judgment can be performed based on the scatter plot type, improving the sensitivity and specificity of auxiliary diagnostic results and / or sample quality inspection results. On the other hand, identifying abnormal regions within the scatter plot and judging them based on their scatter characteristics, compared to the existing technology of delineating specific cell particles through gating, can eliminate interference from other cell particles, further improving the sensitivity and specificity of body fluid detection. Furthermore, by identifying and judging only the scatter characteristics of abnormal regions, there is no need to process the scatter characteristics of all regions of the scatter plot, improving processing efficiency and resulting in more refined and accurate auxiliary diagnostic results and / or sample quality inspection results. This method is relatively low-cost and widely applicable, thus achieving rapid and low-cost body fluid detection.

[0201] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0202] Based on the same inventive concept, this application also provides a body fluid sample detection device for implementing the aforementioned body fluid sample detection method. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations in one or more body fluid sample detection device embodiments provided below can be found in the limitations of the body fluid sample detection method described above, and will not be repeated here.

[0203] In one embodiment, such as Figure 12 As shown, a body fluid sample detection device is provided, comprising:

[0204] The acquisition module 100 is used to acquire optical signal data of the test liquid prepared based on the body fluid sample; the optical signal data includes at least two of forward scattered light signals, side scattered light signals, and fluorescence signals;

[0205] The generation module 200 is used to generate a scatter plot of the test liquid based on the optical signal data;

[0206] Processing module 300 is used to determine the scatter plot type and / or abnormal regions in the scatter plot based on the scatter plot; the scatter plot type includes normal type, detection abnormal type and quality abnormal type;

[0207] The determination module 400 is used to determine at least one of the auxiliary diagnostic results and sample quality inspection results of the body fluid sample based on the scatter features of the abnormal region and the scatter plot type when the scatter plot type is not the normal type; the scatter features are used to characterize the particle distribution of the abnormal region.

[0208] In some embodiments, the processing module 300 includes:

[0209] The identification unit is used to input the scatter plot into the AI ​​model for identification to obtain the scatter plot type.

[0210] A generation unit is configured to obtain a distribution feature map based on the identification information of the scatter plot type in the AI ​​model when the scatter plot type is not the normal type.

[0211] A filtering unit is used to filter out regions of interest as abnormal regions in the distribution feature map.

[0212] In some embodiments, the generating unit is configured to perform the following steps:

[0213] Based on the scatter plot type, the feature layer and predicted values ​​are obtained through forward propagation;

[0214] Backpropagation is performed on the predicted values ​​of the scatter plot type to obtain the gradient information returned to the feature layer; the gradient information characterizes the sensitivity of the channels in the feature layer to the scatter plot type;

[0215] Based on the gradient information of each channel in the feature layer, determine the weight vector corresponding to each channel;

[0216] The distribution feature map is obtained by weighting and summing the weight vector with the corresponding channel.

[0217] In some embodiments, the determining module 400 includes:

[0218] The acquisition unit is used to acquire the body fluid type of the body fluid sample when the scatter plot type is the detected anomaly type;

[0219] The acquisition unit is used to acquire the scatter features that match the body fluid type based on the abnormal region;

[0220] The first determining unit is used to determine whether the auxiliary diagnostic result is an abnormality in the test result of the body fluid sample based on the scatter features that match the body fluid type.

[0221] In some embodiments, the first determining unit is configured to perform the following steps:

[0222] Obtain particle classification information based on the scatter plot;

[0223] Based on at least one of the following: target particles matching the body fluid type in the particle classification information, the difference information between the scatter feature and the particle classification information, and the scatter feature matching the body fluid type, a preset diagnostic function is used for matching. If the matching is successful, the auxiliary diagnostic result is determined to be that the detection result is abnormal, and the disease diagnosis information corresponding to the diagnostic function is output.

[0224] In some embodiments, the determining module 400 includes:

[0225] An acquisition unit is used to acquire particle classification information based on the scatter plot;

[0226] The second determining unit is used to determine, when the scatter plot type is the quality anomaly type, to match the sample quality inspection result with a preset quality discrimination function based on at least one of the difference information between the scatter features and the particle classification information, and the scatter features that match the body fluid type, and to determine that the sample quality inspection result is that the sample quality of the body fluid sample is abnormal if the matching is successful.

[0227] In some embodiments, the apparatus further includes:

[0228] The first output module is used to output disease diagnosis information that matches the body fluid type of the body fluid sample based on the auxiliary diagnosis result when the auxiliary diagnosis result indicates that the test result of the body fluid sample is abnormal.

[0229] The second output module is used to output a prompt message based on the sample quality inspection result when the sample quality inspection result indicates that the sample quality of the body fluid sample is abnormal; the prompt message is used to indicate that the storage and / or preparation of the test solution is unqualified.

[0230] In some embodiments, the training process of the AI ​​model includes:

[0231] Obtain a training sample set; the training sample set includes scatter plots of normal samples, scatter plots of samples with detected anomalies, and scatter plots of samples with abnormal quality.

[0232] Based on the anomaly type, the scatter plots of the non-normal samples in the training sample set are divided into patterns, and combined with the scatter plots of the normal samples, to obtain the divided training sample set.

[0233] The initial model is trained using the partitioned training sample set. When the initial model meets the iteration termination condition, the trained AI model is obtained.

[0234] Each module in the aforementioned body fluid sample detection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0235] Each module in the aforementioned scatter plot detection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of the computer device as software, so that the processor can call and execute the operations corresponding to each module. In one embodiment, a computer device is provided, which can be a sample analyzer, and its internal structure diagram can be as follows: Figure 13As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a scatter plot detection method. The display screen can be an LCD screen or an e-ink display screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0236] Those skilled in the art will understand that Figure 13 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0237] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps: the computer program performs the steps in the above method embodiments.

[0238] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0239] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.) and / or features involved in this application are all information, data and / or features authorized by the user or fully authorized by all parties.

[0240] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0241] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0242] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A sample analyzer, characterized in that, The sample analyzer includes: A sampling component for obtaining bodily fluid samples from a sample container; A sample preparation assembly has at least one reaction chamber and a reagent supply unit; wherein the at least one reaction chamber is used to receive the body fluid sample acquired by the sampling assembly; the reagent supply unit provides a hemolytic reagent and a fluorescent reagent to the at least one reaction chamber, thereby mixing the body fluid sample with the hemolytic reagent and the fluorescent reagent in the reaction chamber to prepare a test solution; An optical detection component is used to irradiate the test liquid flowing through the detection area with light and collect various light signal data generated by the cells due to light irradiation. A memory, wherein the memory stores a computer program; A processor, connected to the memory, is configured to perform the following steps when executing the computer program: The optical detection component acquires optical signal data of the test liquid; the optical signal data includes at least two of forward-scattered light signals, side-scattered light signals, and fluorescence signals; a scatter plot of the test liquid is generated based on the optical signal data; the scatter plot type and / or abnormal regions in the scatter plot are determined based on the scatter plot; the scatter plot type includes normal type, detection abnormal type, and quality abnormal type; if the scatter plot type is not the normal type, at least one of the auxiliary diagnostic result and sample quality inspection result of the body fluid sample is determined based on the scatter characteristics of the abnormal region and the scatter plot type; the scatter characteristics are used to characterize the particle distribution of the abnormal region.

2. A method for detecting bodily fluid samples, characterized in that, The method includes: Acquire optical signal data of the test liquid prepared based on a body fluid sample; the optical signal data includes at least two of forward-scattered light signals, side-scattered light signals, and fluorescence signals; A scatter plot of the test liquid is generated based on the optical signal data; The scatter plot type and / or abnormal regions in the scatter plot are determined based on the scatter plot; the scatter plot type includes normal type, detection abnormal type and quality abnormal type; If the scatter plot type is not the normal type, at least one of the auxiliary diagnostic result and sample quality inspection result of the body fluid sample is determined based on the scatter plot characteristics of the abnormal region and the scatter plot type; the scatter plot characteristics are used to characterize the particle distribution of the abnormal region.

3. The method according to claim 2, characterized in that, The step of determining the scatter plot type and / or the abnormal regions in the scatter plot based on the scatter plot includes: The scatter plot is input into an AI model for recognition to obtain the scatter plot type. In the case where the scatter plot type is not the normal type, a distribution feature map is obtained based on the identification information of the scatter plot type in the AI ​​model; The region of interest is selected from the distribution feature map as the abnormal region.

4. The method according to claim 3, characterized in that, When the scatter plot type is not the normal type, obtaining the distribution feature map based on the scatter plot type includes: Based on the scatter plot type, the feature layer and predicted values ​​are obtained through forward propagation; Backpropagation is performed on the predicted values ​​of the scatter plot type to obtain the gradient information returned to the feature layer; the gradient information characterizes the sensitivity of the channels in the feature layer to the scatter plot type; Based on the gradient information of each channel in the feature layer, determine the weight vector corresponding to each channel; The distribution feature map is obtained by weighting and summing the weight vector with the corresponding channel.

5. The method according to claim 2, characterized in that, The step of determining at least one of the auxiliary diagnostic results and sample quality inspection results of the body fluid sample based on the scatter plot characteristics of the abnormal region and the scatter plot type includes: If the scatter plot type is the detected anomaly type, the body fluid type of the body fluid sample is obtained; Based on the abnormal region, obtain the scatter features that match the body fluid type; Based on the scatter plot characteristics that match the body fluid type, it is determined whether the auxiliary diagnostic result indicates an abnormality in the test result of the body fluid sample.

6. The method according to claim 4, characterized in that, The step of determining whether the auxiliary diagnostic result indicates an abnormality in the test result of the body fluid sample based on the scatter plot characteristics matching the body fluid type further includes: Obtain particle classification information based on the scatter plot; Based on at least one of the following: target particles matching the body fluid type in the particle classification information, the difference information between the scatter feature and the particle classification information, and the scatter feature matching the body fluid type, a preset diagnostic function is used for matching. If the matching is successful, the auxiliary diagnostic result is determined to be that the detection result is abnormal, and the disease diagnosis information corresponding to the diagnostic function is output.

7. The method according to claim 2, characterized in that, The step of determining at least one of the auxiliary diagnostic results and sample quality inspection results of the body fluid sample based on the scatter plot characteristics of the abnormal region and the scatter plot type includes: Obtain particle classification information based on the scatter plot; When the scatter plot type is the quality anomaly type, at least one of the scatter features and particle classification information, and the scatter features that match the body fluid type, is matched with a preset quality discrimination function. If the match is successful, the sample quality inspection result is determined to be that the body fluid sample has an abnormal sample quality.

8. The method according to any one of claims 2 to 7, characterized in that, The method further includes: If the auxiliary diagnostic result indicates that the test result of the body fluid sample is abnormal, disease diagnostic information matching the body fluid type of the body fluid sample is output based on the auxiliary diagnostic result. If the sample quality inspection result indicates that the body fluid sample has abnormal quality, a prompt message will be output based on the sample quality inspection result; the prompt message is used to indicate that the body fluid sample is not properly stored and / or prepared.

9. The method according to claim 3, characterized in that, The training process of the AI ​​model includes: Obtain a training sample set; the training sample set includes scatter plots of normal samples, scatter plots of samples with detected anomalies, and scatter plots of samples with abnormal quality. Based on the detected anomaly type, the scatter plots of non-normal samples in the training sample set are divided into patterns, and combined with the scatter plots of normal samples, to obtain the divided training sample set. The initial model is trained using the partitioned training sample set. When the initial model meets the iteration termination condition, the trained AI model is obtained.

10. A method for detecting bodily fluid samples, characterized in that, The method includes: Acquire optical signal data of the test liquid prepared based on a body fluid sample; the optical signal data includes at least two of forward-scattered light signals, side-scattered light signals, and fluorescence signals; A scatter plot of the test liquid is generated based on the optical signal data; Based on the scatter plot, identify the abnormal regions in the scatter plot; Based on the scatter plot characteristics of the abnormal region, the auxiliary diagnostic results of the body fluid sample are determined; wherein, the scatter plot characteristics are used to characterize the particle distribution in the abnormal region.

Images