Its use in controls and analyzers

Abstract

An analyzer may be adapted to verify its operation using a control sample containing synthetic particles. Such an analyzer may be adapted to obtain particle representations from the sample and to obtain particle classifications that classify the particles as types specific to the control sample. The analyzer may also be adapted to perform a control-specific process on the sample and generate an analytical output for the sample based on the results of the control-specific process.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application is a PCT international application and claims the benefit of priority from U.S. Provisional Patent Application No. 63 / 318,963, entitled "Synthetic Controls and Their Use in Analyzers," filed March 11, 2023, the disclosure of which is incorporated by reference in its entirety. [Background technology]

[0002] Blood cell analysis is a widely performed medical test to provide an overview of a patient's health. A blood sample may be drawn from a patient's body and stored in a test tube with an anticoagulant to prevent clotting. A whole blood sample usually contains three major blood cell classes, including red blood cells (erythrocytes), white blood cells (leukocytes), and platelets (thrombocytes). Each class can be further divided into component subclasses. For example, the five major types or subclasses of white blood cells (WBCs) each have different shapes and functions. White blood cells may include neutrophils, lymphocytes, monocytes, eosinophils, and basophils. There are also subclasses of red blood cell types. The appearance of particles in a sample may vary depending on pathological conditions, maturity of the cells, and other causes. Subclasses of red blood cells may include reticulocytes and nucleated red blood cells.

[0003] To assess and document an analyzer's ability to effectively perform a task, such as, but not limited to, analyzing a blood sample, a control sample with known characteristics may be administered to the analyzer and the results of the analyzer's analysis may be compared to those expected based on the known characteristics of the control sample. However, current approaches to this type of assessment suffer from a number of problems. For example, control samples typically use cellular products with characteristics that mimic, as closely as possible, the characteristics of cells that may appear in a patient sample. However, such cellular products can be difficult to procure, and even if procured, they may have a limited shelf life and / or require stabilization, which can change the characteristics and add both cost and complexity.

[0004] Furthermore, existing control sample techniques are based on techniques that indirectly measure sample parameters (e.g., by fluorescence measurement, light scattering analysis, or conductivity measurement). However, recent blood analysis techniques utilize imaging (e.g., static imaging or flow imaging). Thus, there is a need for improved techniques that can be used to evaluate and document analyzer functionality to address one or more problems with current practice. Summary of the Invention

[0005] Aspects of the present disclosure may be used to verify the performance of an analyzer.

[0006] In an aspect, a method is disclosed, the method including providing a sample to an analyzer, the analyzer may be adapted to use a camera in analyzing the sample. Such a method may also include obtaining a representation of particles from the sample. Such a method may also include obtaining a visual representation or image of particles from the sample. Such a method may also include obtaining a classification of the particles, the classification classifying the particles as a type specific to the control sample. Such a method may also include performing a control-specific process on the sample. Such a method may also include generating an analytical output for the sample, the analytical output based on a result of the control-specific process.

[0007] In another aspect, a system may be provided that may include a camera, a processor, and a non-transitory computer-readable medium storing instructions operable to perform a set of operations when executed by the processor. In such a case, the set of operations may include obtaining a representation of particles from the sample and obtaining a classification of particles from the sample, where the classification classifies the particles as a type specific to the control sample. The set of operations may also include performing a control-specific process on the sample. Further, in this aspect, the set of operations may include generating an analytical output for the sample, where the analytical output is based on a result of the control-specific process. The set of operations may also include obtaining a visual representation or image of particles from the sample.

[0008] In yet another aspect, a non-transitory computer readable medium may be provided having instructions stored thereon operable, when executed by a processor, to cause an analyzer to perform a set of operations. In such an aspect, the set of operations may include obtaining a representation of particles from a sample and obtaining a classification of particles from the sample, where the classification classifies the particles as a type specific to the control sample. The set of operations may also include performing a control-specific process on the sample. Further, in such an aspect, the set of operations may include generating an analytical output for the sample, where the analytical output is based on a result of the control-specific process. The set of operations may also include obtaining a visual representation or image of particles from the sample.

[0009] In some embodiments, control samples are described that utilize control particles. In some embodiments, the control particles are composed of biological materials (e.g., cells of animal origin, such as blood cells from animals, cells of human origin, such as blood cells from humans), synthetic materials, or combinations thereof. In some embodiments, the control particles are uniquely configured for image analysis. In some embodiments, the control particles utilize various combinations of colors, shapes, surface characteristics (such as protrusions and depressions), different sizes, surface detection patterns, internal structures, or surface functionalization.

[0010] While this specification concludes with claims which particularly recite and distinctly claim the invention, it is believed the invention will be best understood from the following description of specific examples read in conjunction with the accompanying drawings, in which like reference numerals identify the same elements, and in which: [Brief description of the drawings]

[0011] [Figure 1] 1 is a schematic diagram, partially in cross section and not to scale, illustrating operational aspects of an exemplary flow cell that may be used with an analyzer configured to capture and analyze images. FIG. [Diagram 2] FIG. 2 is an illustration of an exemplary analysis process. [Diagram 3] FIG. 1 is an illustration of an exemplary classification process. [Figure 4] FIG. 1 is a diagram of an architecture that can be used to analyze images and assign them to specific classes. [Diagram 5] FIG. 1 illustrates the processing steps that may be involved in classification. [Figure 6] FIG. 1 shows exemplary images of control particles derived from blood. [Figure 7] FIG. 13 shows exemplary images of a combination of control particles, including blood-derived and synthetic control particles. [Figure 8] FIG. 1 shows examples of blood-derived control particles along with the blood cell types they can represent. [Figure 9] FIG. 1 is an illustration of a process by which control particles and sample particles can be sorted. [Figure 10] FIG. 1 is an illustration of a process by which control particles and sample particles can be sorted. [Figure 11] FIG. 1 is an illustration of a process by which control particles and sample particles can be sorted. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] The drawings are not intended to be limiting in any way, and it is contemplated that various embodiments of the invention may be embodied in a variety of other ways, including those not necessarily depicted in the drawings. The accompanying drawings, which are incorporated in and form a part of this specification, illustrate several aspects of the invention and, together with the description, serve to explain the principles of the invention. It is understood, however, that the invention is not limited to the precise configurations shown.

[0013] The present disclosure relates to articles, systems, and methods for evaluating and recording the functionality of an analyzer. In some embodiments, the analyzer may be a visual analyzer with a processor that facilitates automated conversion and / or analysis of images. Such analyzers may be useful for characterizing particles in biological fluids, such as detecting and quantifying red blood cells, reticulocytes, nucleated red blood cells, platelets, and white blood cells, including, for example, differential counting, categorization and subcategorization, and analysis. Other similar applications are also contemplated, such as characterizing blood cells from other fluids (serum, bone marrow, lavage fluid, exudate, transudate, cerebrospinal fluid, pleural fluid, peritoneal fluid, and amniotic fluid).

[0014] Turning now to the figures, FIG. 1 shows a schematic of an exemplary flow cell 22 that may be used in an analyzer to convey a sample fluid through a viewing zone 23 of a high optical resolution imaging device 24 (e.g., a camera) in a configuration for imaging microscopic particles in a sample stream 32 using digital image processing. The flow cell 22 is coupled to a source 25 of sample fluid, which may have been treated, such as contacted with a particle contrast agent composition and heated. The flow cell 22 is also coupled to one or more sources 27 of particle and / or intracellular organelle alignment liquid (PIOAL) / coating fluid, such as a clear glycerol solution having a higher viscosity than that of the sample fluid. One example of a PIOAL is disclosed in U.S. Pat. Nos. 9,316,635 and 10,451,612, the disclosures of which are incorporated herein by reference in their entireties.

[0015] Once the PIOAL flow is substantially established, the sample fluid is injected into the interior of the flow cell 22 through a flat opening at the distal end 28 of the sample supply tube 29, resulting in a stable, symmetric laminar flow of the PIOAL above and below (or on either side of) the ribbon-like sample flow. The sample and PIOAL flows may be delivered by precision metering pumps that move the PIOAL along with the injected sample fluid along the significantly narrowed flow path. The PIOAL surrounds and compresses the sample fluid in the narrowing zone 21. Thus, the reduction in the thickness of the flow path in the zone 21 may contribute to the geometric focusing of the sample flow 32. The sample flow 32 is surrounded and conveyed downstream of the narrowing zone 21 with the PIOAL, passing in front of or otherwise passing through the viewing zone 23 of the high optical resolution imaging device 24 where an image is collected using, for example, a CCD 48. The processor 18 may receive pixel data from the CCD 48 as an input. The sample fluid ribbon flows with the PIOAL to the exhaust 33.

[0016] As shown here, narrowing zone 21 can have a proximal flow path portion 21a having a proximal thickness PT and a distal flow path portion 21b having a distal thickness DT such that the distal thickness DT is less than the proximal thickness PT. Thus, sample fluid can be injected through distal end 28 of sample tube 29 at a location distal to proximal portion 21a and proximal to distal portion 21b. Thus, sample fluid can enter the PIOAL enclosure at the same time that the PIOAL flow is compressed by zone 21. Here, the sample fluid injection tube has a distal exit port through which sample fluid is injected into the flowing sheath fluid, the distal exit port being delimited by the reduction in flow cell flow path size.

[0017] A digital high optical resolution imaging device 24 with an objective lens 46 is directed along an optical axis that intersects with the ribbon-like sample flow 32. The relative distance between the objective lens 46 and the flow cell 33 can be varied by operation of a motor drive 54 to resolve and collect a focused digital image on a photosensor array. Additional information regarding the structure and operation of an exemplary flow cell such as that shown in FIG. 1 is provided in U.S. Pat. No. 9,322,752, issued Mar. 17, 2014, entitled "Flowcell Systems and Methods for Particle Analysis in Blood Samples," the disclosure of which is incorporated herein by reference in its entirety.

[0018] In operation, an analyzer, such as an analyzer incorporating a flow cell-based imaging system as shown in FIG. 1, may analyze images of particles (e.g., cells) in a sample, such as by using a method of the type shown in FIG. 2. The control particles are used to ensure that the analyzer is functioning properly and is able to validly and correctly count or otherwise determine the particles (e.g., blood cells) to be analyzed. In this manner, a control particle procedure utilizing control particles is performed (e.g., at the start of the day before the analyzer is used) to ensure that the analyzer is working properly and is able to correctly count the patient sample (e.g., blood cells). As further disclosed herein, the control particles can utilize whole blood (e.g., derived from human or animal blood), synthetic materials, or mixtures thereof.

[0019] Initially, in the process of FIG. 2, a determination 201 is made as to whether the sample is a patient sample, such as a blood sample, or a control sample, such as a sample containing control particles. In this context, it should be understood that a control particle is a particle that represents a known particle and may be used to ensure that the analyzer is functioning properly. For example, red blood cells (erythrocytes), nucleated red blood cells, platelets (thrombocytes), reticulocytes, and white blood cells (leukocytes) (including at least neutrophils, lymphocytes, monocytes, eosinophils, basophils, and optionally immature granulocytes and blasts) may each have an associated control particle that is intended to represent that particular particle. The correct identification of the control particle is used to ensure that the analyzer is functioning properly and can proceed to analyze the blood sample. In one example, if the system counts the correct number or range of control particles, it provides confirmation that the system is functioning properly and can proceed to run the blood sample.

[0020] Although the examples given herein generally describe the types of analysis that may be performed on whole blood samples, it should be understood that the disclosed techniques may be used on analyzers that analyze other types of samples. Thus, the patient sample may be any biological fluid that contains cells, and the control sample may be a control for a type of sample other than blood (both mononuclear and nucleated), such as a bone marrow control. For example, the patient sample or control sample may be whole blood, e.g., blood that has not been treated or altered except for the addition of an anticoagulant, which may be done to prevent the blood from clotting. If the blood clots, it becomes difficult to flow the blood through a flow cell for analysis. The sample may be treated, e.g., by dilution or concentration. In some circumstances, the sample may be obtained from a non-blood body fluid, such as urine, spinal fluid, saliva, bile, cerebrospinal fluid, amniotic fluid, semen, mucus, sputum, lymphatic fluid, aqueous humor, tears, vaginal fluid, pleural fluid, pericardial fluid, peritoneal fluid, etc. Similar to blood, when a non-blood body fluid is sampled, the non-blood body fluid may be processed, for example, to achieve a desired cell concentration for analysis. A possible advantage of evaluating whole blood may be that there is a relatively large number of cells available for analysis in a relatively small sample volume. A possible advantage of analyzing non-blood body fluids and / or processed blood may be that the specific cells of interest are pre-isolated and / or the number of cells is reduced due to differences in cell types and numbers that normally occur in different body fluids. A low number of cells may be useful, for example, when characterizing individual cells.

[0021] In some aspects, the control particles are included in a single control sample (e.g., a single tube with each type of control particle) such that the control sample contains multiple types of control particles. Other embodiments can distribute the control particles among various control samples (e.g., two or more control samples). In some aspects, the control particles can be provided in a carrier fluid, e.g., a carrier fluid that is isotonic and / or has the same osmolarity and / or pH as the biological sample. For example, the control particles can be provided in a carrier fluid that is isotonic with the blood sample.

[0022] In some embodiments, the control particles may be formulated to work with imaging technology (e.g., the system of FIG. 1). As such, the control particles may utilize characteristics suitable for imaging (e.g., various colors, shapes, surface characteristics (such as protrusions, depressions, or a combination thereof), different sizes, surface detection patterns, internal structures, surface functionalization) to provide an image-based representation of the target particle. In some embodiments, the control particles do not directly correlate visually with their target particles (e.g., control neutrophils may not actually visually represent neutrophils), but utilize some uniquely identifiable parameter that is sufficiently distinct from other control particle types so that the control process can properly identify the control particle type and use that information to ensure that the analyzer is functioning properly.

[0023] In some embodiments, the control particles may be composed of biological materials (e.g., blood cells from animals or humans), synthetic materials (e.g., polymers), or a mixture of biological and synthetic materials. In one example, all control particles in a control sample may utilize animal blood cells that have been modified or engineered to impart imaging characteristics (e.g., size, color, and / or shape) intended to reflect the target particles they are intended to represent. Such modification or manipulation may include utilizing various processes to shrink, enlarge, reshape, or otherwise modify the physical / imaging characteristics of the cells to reflect the target particles they are intended to represent. For example, FIG. 8 shows exemplary images of control particles for biological fluids, such as control particles corresponding to red blood cells (RBCs), platelets, lymphocytes, monocytes, neutrophils, eosinophils, nucleated red blood cells (nRBCs), and reticulocytes. In one example, all control particles in a control sample may utilize synthetic materials (e.g., polymers, beads, metals, alloys, hydrogels, combinations thereof) engineered to provide imaging properties (e.g., size, color, shape, and / or surface properties) intended to reflect the target particles they are intended to represent. For example, the synthetic materials can be engineered to have a similar size to the represented particles, or engineered to have an internal colored structure similar to a nucleus (e.g., in the case of white blood cell controls and nucleated red blood cells). In one example, some control particles, such as red blood cell control particles, may utilize animal blood cells modified (e.g., to a size or shape similar to red blood cells) to provide imaging properties intended to reflect the target particles they are intended to represent, while some control particles, such as at least a subset of white blood cell control particles, may utilize synthetic particles. As an example, some blood cell types (e.g., white blood cells) contain nuclei as described above and herein, and therefore the control particles may utilize internal structures to mimic nuclei. For example, a synthetic material may utilize an internal structure that appears like a nucleus when analyzed by an analyzer.

[0024] The control particles may be made of any suitable material. Non-limiting materials that may be used to form the control particles include, but are not limited to, cellulose, silica (silicon dioxide), polymethylmethacrylate (PMMA) / hydrogel-coated materials, melamine (melamine formaldehyde resin), cross-linked agarose, polyvinyl acetate (PVA), polystyrene, metals, hydrogels, and combinations thereof. In some embodiments, the control particles may be made of transparent or translucent materials. In this embodiment, the detection of the control particles may be based entirely or partially on the detection of the light reflectivity of the control particles. In further embodiments, the control particles may comprise a color, and the color of the control particles is used as a detectable label. In still other embodiments, the control particles may be made of a dissolvable and / or stainable material. Upon staining or dissolving the control particles, the detection of the stained and / or dissolved particles may provide further information about the sample being measured, including, but not limited to, the calibration or validation of data obtained via the disclosed methods and / or systems.

[0025] In some embodiments, the control particles may be synthetic beads, e.g., polystyrene microspheres. The control particles may be provided in a control particle composition, which includes synthetic beads of various sizes and colors. For example, as depicted in FIG. 7, beads of various colors may be used to represent different particles to be detected. For example, small red synthetic beads may represent red blood cells, large red beads may represent eosinophils, small white beads may represent platelets, medium white beads may represent lymphocytes, and large blue beads may represent basophils. In this embodiment, the size and color of the synthetic beads may be used as characteristics to represent whole blood cells. One advantage of synthetic controls is that they have a longer shelf life compared to controls derived from human or animal blood, which may require extensive quality control procedures and have a limited time frame for use before deteriorating.

[0026] In certain embodiments, the control particles may be sized from about 1 μm to about 25 μm in diameter, or from about 3 μm to about 20 μm in diameter, or from about 5 μm to about 15 μm in diameter, or from about 10 μm to about 13 μm in diameter. In certain embodiments, the control sample may include one or more control particles of different sizes. For example, the control sample may include control particles having a size equivalent to white blood cells in addition to control particles having a size equivalent to red blood cells (e.g., smaller than the white blood cell control). Similarly, the control sample may include control particles having a size that does not match any expected blood sample component in size.

[0027] In some embodiments, the control particle comprises a surface with a detection pattern. In this embodiment, the detection pattern can be detected by the instrument (e.g., by recognizing the detection pattern via imaging) to allow detection or quantification of the control particle. In some embodiments, the detection pattern can provide surface properties that affect light scattering. In this embodiment, the light scattering caused by the control pattern can be detected, which can further provide information about the sample being measured.

[0028] In further embodiments, the control particle may have surface functionalization.For example, the control particle surface may comprise functionalized particles or DNA molecules.The exemplary groups that can be used to functionalize the control particle include, but are not limited to, mercapto group, hydroxyl group, carboxyl group, disulfide group, polyvinyl alcohol group, amine group (primary and secondary ammonium), maleimide group, tertiary ammonium group, quaternary ammonium group, epoxy group, carboxyl sulfonate group, and octadecyl (C18) group.

[0029] Control particles may be used to determine if an analyzer is functioning correctly (e.g., by correctly determining the number of control particles analyzed), while calibration particles may be used to calibrate the analyzer to ensure that it is outputting the correct information (e.g., by setting certain parameters, such as the output of a numerical channel or electronics configuration). In this way, calibration particles can be used to adjust the parameters of the analyzer to ensure that it is correctly analyzing the specimen presented to it, while control particles may be used to ensure that the analyzer is working correctly. As an example, if a control sample is run and the analyzer does not read the control particles correctly (e.g., does not read the correct number or range), the data can be flagged and an operator can input calibration particles to adjust the machine's functioning and then reintroduce a control particle sample to see if the analyzer now reads the particles correctly. Although the description herein has generally focused on control particles, in some embodiments, these attributes can also be utilized for calibration particles.

[0030] In some embodiments, an analyzer system may analyze a sample in a variety of ways using a process as shown in FIG. 2. As shown in FIG. 2, in some cases, an analyzer may obtain a representation (e.g., a visual representation) of a plurality of particles (e.g., blood cells in a blood sample, or control particles in a control sample) and provide those representations to a classifier trained to recognize and classify particles into classes corresponding to either a patient sample or a control sample. In such cases, the analyzer may determine whether a sample is either a patient sample or a control sample with a variety of techniques. For example, whether a majority of the particle representations are classified as patient sample particles or control sample particles, or whether a barcode scanning process determines that the sample is a patient sample or a control sample, or whether a software process determines that a certain confidence score threshold is exceeded are all techniques that may be used to determine whether a sample is a patient sample or a control sample.

[0031] To further illustrate how a classifier-based approach may determine 201 whether a sample is a patient or control sample, consider FIG. 4, which illustrates an architecture that may be used to analyze an image and assign the image to a particular class. In the architecture of FIG. 4, an input image 401 is analyzed in a series of stages 402a-402n, each of which may be referred to as a "layer," and is illustrated in more detail in FIG. 5. As shown in FIG. 5, an input 501 (which in the first layer 502a of FIG. 4 is the input image 401, and otherwise is the output of the previous layer) is provided to layer 502, where it is processed to generate one or more transformed images 503a-503n. This processing may include convolving the input 501 with a set of filters 504a-504n, each filter identifying a certain type of feature from the underlying image that is then captured in that filter's corresponding transformed image. For example, convolving an image with the filters shown in Table 1, as a simple example, results in:

[0032] [Table 1] From the input 501, a transformed image can be generated that captures edges.

[0033] As shown in FIG. 5, in addition to generating transformed images 503a-503n, the layer may also generate a pooled image 505a-505n for each of the transformed images 503a-503n. This may be done, for example, by organizing the appropriate transformed images as a set of regions and then replacing the values ​​in the region with a single value, such as the maximum value of the region or the average of the values ​​in the region. The result is a pooled image with reduced resolution relative to the corresponding transformed image based on the size of the regions into which the transformed image was divided (e.g., if the transformed image has dimensions N×N and it is divided into 2×2 regions, then the pooled image will be of size (N / 2)×(N / 2)). These pooled images 505a-505n may then be combined into a single output image 506, with each pooled image 505a-505n treated as a separate channel in the output image 506. This output image 506 may then be provided as input to the next layer as shown in FIG. 4.

[0034] Returning to the discussion of FIG. 4, after a final output image 403 has been created through various stages of processing 402a-402n, the final output image 403 may be provided as an input to a neural network 404. This may be done, for example, by providing the values ​​of each channel of each pixel in the output image 403 to the input nodes of a tightly connected single-layer network. The output of the neural network 404 may then be treated as a classification of the image into a class indicating that it corresponds to either a control sample or a patient sample. For example, in some cases, the neural network may have multiple output nodes, each corresponding to one possible classification (e.g., red blood cell particles in a patient sample, which may be identified, for example, as rbc_blood, where the blood label subscript indicates the patient blood sample, red blood cell particles in a control sample, which may be identified, for example, as rbc_control, where the control label subscript indicates the control sample, etc.), in which case the node with the highest value may be treated as the appropriate classification for the image. Note that the _blood and _control notations are provided as example labels to indicate sample type, and any of a variety of notations may be used (e.g., _bl, _blood, _b, _control, _ctrl, _c, etc.).

[0035] Alternatively, in some cases, there may be only one output node, and the value of that output node may be treated as indicating whether the image should be treated as representing particles from a control sample or as representing particles from a patient sample (e.g., if the value of the output node is above a threshold, the image may be treated as representing particles from a control sample, and if the value is below the threshold, the image may be treated as representing particles from a patient sample). In either case, a classifier such as that shown in Figures 4 and 5 may be trained to make a decision using a training method in which it is asked to classify human-labeled images that represent the types of particles used for classification in production. In such a method, if the classifier's classification differs from the human label, an error value is determined (e.g., using cross-entropy loss if the classifier has multiple output nodes), and the error is back-propagated through the nodes of the neural network and layers of the classifier such that in future iterations the classifier's classification is expected to approach the labels assigned by the human annotators.

[0036] Of course, it should be understood that the classifiers as described above in the context of Figures 4 and 5 are merely illustrative and should not be considered to suggest limitations on how classification may be performed in various implementations of the disclosed technology. For example, in some cases, classification of a particle as indicative of a control sample or a patient sample may be based on a feature of the particle, where the feature of the particular particle is compared to a threshold or evaluated for other characteristics associated with a particular particle type. Such features may include, without limitation, the area of ​​the particle (e.g., a cell) represented in the image, the intensity value of a pixel representing the particle (which may be determined, for example, by thresholding to remove the background and taking a value for the pixel represented in L*a*b color space), the average intensity value of the pixel at the edge of the particle, the average intensity value of the pixel within a defined distance from the edge of the particle, the color of the pixel representing the particle, etc. The determination of the type of particle may be based on a combination of features, such as the result of dividing the blue and red values ​​of the pixel representing the particle in RGB color space. In cell-based / particle-based classification, as an example, each particle receives a label (e.g., rbc_blood for patient red blood cells or rbc_control for red blood cell control), and the total number of labeled cells is pooled to determine the sample type (e.g., by majority voting or assessment of a confidence score).

[0037] Additionally, in some cases, particle type determination may utilize feature-based assessment, but the determination may be made on a population basis, rather than based on properties of particles considered individually. For example, in some cases, particles may be clustered in an n-dimensional space, with each dimension being a particle property, and the particles may be classified based on the relationship of each particle's cluster to other clusters in the space (e.g., a particle may be classified as a white blood cell if it belongs to a cluster of particles with relatively high darkness and blue-redness values, and then may be further classified as a type of white blood cell based on additional properties, such as the number of dark blue particles surrounding the nucleus). In one example, cell type determination at the population level is made generally simultaneously for each cell, in which case, once the population pool and associated label of each cell are identified, a majority voting or confidence thresholding system may then be applied at the population level and used to establish the sample type (e.g., either a patient sample or a control sample). Thus, the preceding description of how particles may be classified should be understood to be merely illustrative, and should not be treated as implying a limitation on the scope of protection afforded by this document.

[0038] Information regarding feature-based and population-based classifiers is disclosed in US Provisional Application No. 63 / 434,658 and US Pat. No. 11,403,751, the disclosures of which are incorporated by reference in their entireties.

[0039] Additional examples may utilize multiple particle classifier types, where each classifier performs particle-by-particle assessments and uses them to help determine whether a sample is a patient sample or a control sample (e.g., if a majority of the particles are of the patient's blood type or control type, or if a certain confidence level for the sample type is determined). There may be a further consensus process between the multiple classifiers to determine the label of each cell type, or to determine the label of the sample itself (e.g., is a patient sample or a control sample). The particle classifiers may be feature or population-based classifiers, neural network-based classifiers, or a combination thereof. In one embodiment, a majority vote decision between the multiple classifiers may be used to assign a cell label. In one embodiment, a weight may be assigned to a particular one of the classifiers such that if the particular classifier exceeds a certain confidence score threshold for a particular label, that label is assigned. Additional information regarding voting or final label determination for multiple classifiers may be found in U.S. Provisional Application No. 63 / 434,798, the disclosure of which is incorporated by reference in its entirety.

[0040] In embodiments according to the method as shown in FIG. 2, once a sample is determined 201 as either a patient sample or a control sample, that determination may be used to perform processing corresponding to the determined sample type. For example, a determination that a sample is a control sample may trigger a specific execution 202 of a procedure (e.g., a control sample procedure or a control sample run mode) that is specific to control sample / control particle analysis. For example, if a sample is identified as a control sample based on classifying a majority of particles as control particles, the control-specific processing may include reclassifying particles that were not initially classified as control particles (e.g., reclassifying particles that were added to the class of red blood cells in the patient sample to the class of red blood cells in the control sample) using a control particle-specific classifier or using data correlating the classification of the control particles to the classification of the patient sample. Alternatively, if a sample is identified as a control sample based on classifying a majority of particles as control particles, any particles that were classified as blood particles rather than control particles may be ignored. Similarly, if the sample is determined to be a patient sample 201, the method of FIG. 2 may perform patient-specific processing 203, such as classifying particles from the sample using a patient sample-specific classifier (e.g., a convolutional neural network-based classifier as described above in the context of FIGS. 4 and 5, or a feature-based classifier as described above) or reclassifying particles originally classified as control particles in a manner similar to that described above with respect to performing control-specific processing 202.

[0041] Of course, variations on the above approach are possible: for example, in some cases, if a sample is identified as a control sample based on classifying the majority of particles as control particles using a relatively coarse classifier, the control-specific processing may be to process the particle images (which may be all particle images or only those identified as control particles with other images classified as junk or unknown) using a finer-grained classifier. For example, a cell image may first be classified into classes of particles identified as nucleated cells in a control sample, particles identified as red blood cells in the sample, particles identified as reticulocytes in a control sample, junk particles, particles identified as nucleated cells in a patient sample, particles identified as red blood cells in a patient sample, or particles identified as reticulocytes in a patient sample, and then once the sample has been classified as either a control sample or a patient sample, it may be classified using a finer grained classifier (e.g., a control particle classifier that classifies particles into classes of control eosinophils, control monocytes, control lymphocytes, control neutrophils, control nucleated red blood cells, control non-nucleated red blood cells, control reticulocytes, and junk). In some cases, other types of processing specific to patient or control samples may be performed, such as providing a report documenting verification of the functionality of the analyzer in the case of a control sample, or providing results of a test to identify a disease in the case of a patient sample. As another alternative, in some cases, all cell images may be classified with a single classifier trained to classify cells into control or patient sample classes as appropriate, and once it has been determined whether the sample was a control or patient sample, type-specific processing may be such that images added to a class that does not match the type of sample are reclassified into a junk or unknown class. Thus, the above description of classification techniques and patient / control specific processing should be understood to be merely illustrative and should not be treated as implying any limitation on the scope of protection afforded by this or other related documents.

[0042] Other uses of image-based analyzers and / or classifications are possible. For example, in some embodiments, an image-based system utilizes a classifier to classify the type of particle being analyzed (e.g., classifying blood particles as either red blood cells, nucleated red blood cells, reticulocytes, platelets, neutrophils, lymphocytes, monocytes, eosinophils, or basophils). Because neutrophils, lymphocytes, monocytes, eosinophils, and basophils are all white blood cells, the differentiation between these five groups is known as a 5-part differential, but this can be expanded to a 6-part differential by including either blasts or immature granulocytes, or to a 7-part differential by including both blasts and immature granulocytes.

[0043] Image-based systems can utilize classifiers to classify specific particles of interest. Classifiers may utilize data structures, such as decision tree classifiers, neural networks, or Bayesian classifiers, trained to recognize various subpopulations using training data that includes particle images annotated with appropriate subpopulation types. Classifiers can be used to classify or label specific particle types (e.g., using the control sample classes and patient sample classes of the types described above).

[0044] As mentioned above, a type determination may be made and then a corresponding processing on the whole sample may be applied, however, it is also possible to implement the disclosed technology to make a type determination on a particle-by-particle basis and apply an appropriate processing. An example of a method that may be used to implement this type of approach is provided in FIG. 3. Initially, in the process of FIG. 3, one or more representations of particles (e.g., cells) from the sample are received 301. This may include, for example, a processor receiving one or more images, each including a representation of multiple particles. These representations may then be separated (e.g., using a cell separation algorithm, such as an algorithm that thresholds an image captured by a flow cell-based system to identify parts of the image that represent cells and parts that do not represent cells) and stored in the analyzer memory for further processing. Alternatively, receiving the particle representations 301 may include a processor receiving multiple images, each image including a representation of only a single particle (e.g., a single cell), which may then be subjected to further image processing (e.g., thresholding to remove background parts of the image) or to entity analysis, such as classification, as described below.

[0045] In the process of FIG. 3, after one or more particle representations are received 301, each of the representations may be processed by performing steps including providing it as an input to a classifier 302. To illustrate how this may be done, consider a scenario in which an analyzer is used to perform a 5-part filter, where white blood cells in a sample are classified into subpopulations of neutrophils, lymphocytes, monocytes, eosinophils, and basophils. In such a case, to facilitate the 5-part filter analysis, prior to deployment of the analyzer, a data structure such as a decision tree classifier, neural network, or Bayesian classifier may be trained to recognize the various subpopulations using training data including particle images annotated with the appropriate subpopulation types. A copy of this trained data structure may then be stored in the analyzer's memory before being used, and the step of providing the representations to the classifier 302 may be performed using the representations as input that are processed using this data structure.

[0046] After the classifier assigns a class to a particular representation, in a process such as that shown in FIG. 3, a determination 303 may be made as to whether the classification is a particle specific to the control sample. To illustrate what this type of determination entails, consider the case of a control sample that may contain synthetic particles (control particles made from synthetic materials such as polymers). The control sample may be a composition that includes control particles, where the control particles have various combinations of predetermined sizes and shapes, detectable surface marks, colors, and internal features. In some embodiments, the above features may be used as reference marks so that the control particles can mimic blood particles that resemble them (e.g., control red blood cells having a shape similar to that of real red blood cells, control neutrophils having an internal region of a shape and color similar to that of the nucleus of a real neutrophil, control granulocytes having granular features similar to those of real granulocytes).

[0047] To accommodate synthetic control samples as described above, an analyzer may include a classifier trained to not only recognize classes of particles (e.g., blood cells / blood particles) as contained in a patient sample, but also to include classes of synthetic particles from the control sample. For example, an analyzer used to perform a 5-partition on a patient sample and that may be validated using a control that includes synthetic particles may include a classifier capable of classifying particles into neutrophils, lymphocytes, monocytes, eosinophils, and basophils (i.e., types of cells that may be expected to be present in a patient sample), or into one or more control particle types (i.e., types of synthetic particles that may be expected to be present in a control sample). In such a case, the analyzer may be configured with data indicating which of the classifications corresponded to synthetic particles not expected to be present in a control sample, and a determination 303 of whether a particle was classified using a control classification may be made by comparing the particle's classification to the analyzer's classification data.

[0048] In a process such as that shown in FIG. 3, if a particle is classified into a class that is not unique to the control (e.g., a class of a real blood type, such as red blood cells), the classification may be stored in the analyzer's memory for later processing 304. Alternatively, if a particle is classified into a class that is unique to the control (e.g., control red blood cells), the classification may be converted to a classification of the corresponding particle type expected to be present in the patient sample 305. For example, in a synthetic control sample including synthetic neutrophil control particles, the analyzer may be configured to treat particles of that class as classified as neutrophils, and a similar type of conversion may be made to each of the other types of synthetic particles expected to be present in the control sample but not in the patient sample. The converted classification may then be stored 304, such as by incrementing a counter (e.g., a neutrophil counter) corresponding to the converted classification, as if the particle had originally been classified using the class expected to be present in the patient sample. Finally, after classification of particles from the sample is complete, the results of the classification may be applied 306, such as by outputting the number of particles assigned to each associated class. This information may then be used for various purposes, such as diagnosing or treating a disease condition (if the sample was a patient sample). However, other applications are possible. For example, the results of a process such as that shown in FIG. 3 may be used to verify the functionality of an analyzer by comparing actual classification results with expected classification results if the analyzed sample was a control sample. Thus, the above description of various applications of the process of FIG. 3 and how that process may be performed using representations of particles from a patient sample or a control sample should be understood to be merely illustrative and should not be treated as limiting.

[0049] In addition to possible variations in how the output of a process such as that shown in FIG. 3 may be applied, it should be understood that variations are also possible in how a process such as that shown in FIG. 3 may be performed. For example, in some cases, storing the classification result 304 may be performed differently depending on whether the particle in question was classified into a control class or into a class of particles expected to be present in a patient sample. For example, in some cases, if a classification result is stored for a particle that was originally classified into a control class, data such as a flag or counter may be used to reflect the original classification of the particle even though it may be treated as belonging to another class after conversion 305. Similarly, while the above example describes how the disclosed technology may be used in the context of classifying a 5-part white blood cell fraction, the disclosed technology may be used in other contexts. For example, in some cases, a control sample may be used to verify aspects of an analyzer, such as its ability to properly stain particles in a sample, or to properly perform tasks other than the 5-part fraction (e.g., generate a red blood cell count, perform a 6 and / or 7-part white blood cell fraction), or to generate other types of information about the sample (e.g., report the average platelet volume). In each case, the types of techniques described above may be used on control samples that may contain particles that may be detected as different from those that would be present in a patient sample, and appropriate processing is performed based on this determination. Thus, the descriptions of the above processing variations, such as those in Figures 2 and 3 and the associated text, should be understood to be merely illustrative and should not be treated as limiting.

[0050] To further illustrate how the disclosed technology can be applied, consider FIGS. 9-11, which depict a process by which control particles and patient particles (e.g., blood) can be classified. In the process depicted in FIG. 9, a routine is performed to identify whether a sample is a control sample or a patient sample 901. This can be done in a variety of ways. In one example, a subset of images capturing particles in that sample is taken and a classifier trained for that purpose is used to classify them as control particles or patient sample particles, and the sample is identified as a control sample or a patient sample based on how the majority of particles from the subset of images were classified. In another example, a confidence score is determined upon assessment of the sample and used to determine the sample type. In another example, a barcode scan of the sample prompts the system's software to determine whether the barcode indicates a control sample or a patient sample. After this identification 901 of the sample as a control sample or a patient sample, all or a subset of the particles associated with the identified sample type (e.g., only particles associated with the control sample class if the sample was identified as a control sample, or only particles associated with the patient sample class if the sample was identified as a patient sample) are themselves identified 902 903. For example, the rest of the particles not previously classified can be classified as either control particles or patient sample particles, or the entire sample can be re-run and classified, with particles that fell into the class of the identified 901 sample type being flagged for further processing. A classifier specialized for classifying particles in control samples or samples of blood or other patient fluids can then be run 904 905 on the particles identified 902 903 as belonging to a class associated with the identified 901 sample type.9, a classifier specialized for classifying particles in a control sample may be configured to classify particles into a particular control particle type class, such as red blood cell equivalents in the control sample (e.g., rbc_control), platelet equivalents in the control sample (e.g., plt_control), neutrophil equivalents in the control sample (e.g., neutro_control), etc. Similarly, a classifier specialized for classifying particles in a blood or other bodily fluid sample may be configured to classify particles into a particular blood particle type class, such as red blood cells (e.g., rbc_blood), platelets (e.g., plt_blood), and neutrophils (e.g., neutro_blood). If any particles are misclassified in the initial step where the sample type assessment is made, they can be later reclassified or ignored (e.g., if patient particles are counted during assessment step 901 and then step 902 indicates that the sample is a control sample, the patient particles can be reclassified as control particles [e.g., rbc_blood is changed to rbc_control], or particles labeled _blood can be ignored entirely). The associated count numbers can then be provided to the user (e.g., on a screen display). However, it should be noted that the specific classes and class names shown in FIG. 9 and described above are provided for illustrative purposes only, and other names may be used for either the control particles or the patient sample particles in various embodiments implemented according to the present disclosure.

[0051] FIG. 10 depicts a process similar to that depicted in FIG. 9. Initially, in the process of FIG. 10, a representation (e.g., an image) of particles in a sample is classified 1001 with a classifier trained to place particles in a class corresponding to a control sample (control particle section) and a class corresponding to a sample of blood or other bodily fluid (blood sample section). A determination 1002 may then be made as to whether the sample is a control sample or a patient sample (e.g., whether a majority of the particles were classified into a class associated with a control sample or a class associated with a blood or other bodily fluid sample, or whether a confidence score classifies the sample as a patient sample or a control sample). As shown in FIG. 10, these classes may have labels such as neutrophil_blood and neutrophil_control, although these labels are merely exemplary and different labels may be used in different embodiments. Then, depending on whether the majority were particles expected in blood or other bodily fluids or were particles expected in a control sample, processing specific to the particles that would be identified in those types of samples may be performed 1003 1004. For example, particles identified as associated with a class of control samples may be classified with a specialized control classifier, or particles identified as associated with a class of blood or other bodily fluid samples may be classified with a specialized classifier for blood or other bodily fluid samples. In addition, particles classified by the initial classifier 1001 as belonging to a class associated with a sample type that was not in the majority may be reclassified (e.g., using a specialized classifier, or automatically reclassified, e.g., from rbc_blood to rbc_control, or from plt_blood to plt_control, if determined to be a control sample) or may be ignored. The associated counts may then be provided to the user (e.g., on a screen display).

[0052] FIG. 11 also depicts another process that can be used to classify particles in a sample. In the process of FIG. 11, a representation (e.g., an image) of particles in a sample can first be classified 1101 using a single classifier trained to classify particles in a control sample as well as particles in a sample of blood or other bodily fluid. These classifications (which may be represented by the labels shown in FIG. 11 or using other labels, depending on the implementation) can then be used to obtain a total count of all types of particles in the sample 1102 using the classes assigned in the initial classification 1101. Finally, based on the count data, a determination 1103 is made whether the sample is a control sample or a patient (e.g., blood) sample (e.g., depending on whether a majority of the particles are assigned to a class associated with a control sample or a majority of the particles are assigned to a class associated with a blood or other bodily fluid sample, or based on a confidence score associated with the labeling or count). Particles whose labels do not match the sample type call can then be relabeled (e.g., if the sample is called a control sample, rbc_blood or plt_blood are reclassified as rbc_control or plt_control) or ignored. The associated count number can then be provided to the user (e.g., on a screen display).

[0053] 9-11 illustrate process variations that may occur according to the present disclosure, but it should be understood that they are not the only variations that may exist between systems or methods implemented according to the present disclosure. As another example of the types of variations that may exist between implementations, consider the types of representations that may be used for particles in a control sample or bodily fluid sample. In some cases, rather than or in addition to a visual representation, the particle representation may include using volume, conductivity, and light scattering (VCS) properties. In such cases, the control particles may have VCS properties similar to the blood or other bodily fluid particles that they represent, rather than (or in addition to) having visual properties corresponding to blood or other bodily fluid particles. For example, control eosin particles may have VCS parameters similar (but not necessarily identical) to blood eosin sufficiently to allow imaging to differentiate the particles by imaging. In some embodiments, the control particles may be provided in a control particle composition that includes control particles engineered to have similar fluorescence or optical light scattering properties as the particles they represent. In some embodiments, the control particles may be provided in a control particle composition that includes control particles engineered to have similar properties to the depicted particles across multiple detection formats (e.g., imaging and VCS, or imaging and fluorescence / light scattering).

[0054] As another example of the type of variation that may be implemented in some cases, consider how a determination may be made whether a sample is a control sample or a sample of blood or another bodily fluid. This determination may be made using techniques such as majority voting based on the classification of the representations, but may also be determined in other ways, such as a user scanning a barcode on a control sample, which informs the system that a control sample is being run. For example, a control sample may have a specific barcode label or specific encryption that the analyzer software may identify as indicative of a control sample. In such a case, the sample may be identified as a control sample or a bodily fluid sample using a barcode or similar type of data in the manner shown in Figures 2, 3, and 9-11, and this identification may be used in processing in a similar manner to other techniques to determine the sample type as previously described. Combined techniques are also possible. For example, a system implemented according to the present disclosure may tentatively determine that a sample is a patient sample or a control sample based on a barcode scanned by the user, and then use particle classification to confirm that determination. In additional variations, such as implementations of the methods as shown in Figures 2, 3 or 9-11, thresholds other than majority vote were used to determine whether a sample is a control sample or a patient sample, such as requiring a majority (e.g., 66%, 75%, 80%, 90%, 95%, or 99%) of the particles to be classified into a class associated with a particular sample type, and flagging the sample if this majority threshold could not be met. A confidence level may also be used in determining the type of sample or particle. For example, if a particle or group of particles is classified into a particular class with less than a threshold confidence (e.g., as shown using a Softmax function on the output of a neural network-based classifier as shown in Figures 4 and 5), the particle / group of particles may be flagged for further inspection, and the sample will be classified as either a control sample or a bodily fluid sample only if the threshold requirement (e.g., majority vote classification) is met at the required confidence level.Other variations are possible, such as relative plurality scores when a majority score cannot be determined (e.g., in scenarios involving control particle labels, patient particle labels, and unidentified or junk labels). Other variations are possible and will be readily apparent to one of skill in the art in light of this disclosure. Thus, the preceding descriptions of particle and / or sample classification techniques should be understood to be merely illustrative and should not be treated as limiting.

[0055] As further illustration of possible implementations and applications of the disclosed technology, non-exhaustive examples of ways in which the teachings herein may be combined or applied are provided below. It should be understood that the following examples are not intended to limit the scope of the claims that may be presented at any time during this application or at a later time when this application is filed. No disclaimer is intended. The following examples are provided for illustrative purposes only. It is contemplated that the various teachings herein may be configured and applied in numerous other ways. It is also contemplated that some variations may omit certain features mentioned in the following examples. Thus, unless expressly indicated as such at a later date by the inventor or a successor related to the inventor, any aspect or feature mentioned below should not be considered essential. If a claim is presented in this application or a later application related to this application that includes additional features beyond those mentioned below, those additional features should not be considered added for reasons related to patentability.

[0056] Example 1 A system for cell classification comprising: a camera; a processor; and a non-transitory computer readable medium storing instructions that, when executed by the processor, are operable to perform a set of operations including obtaining representations of particles from a sample; and obtaining a classification of the particles, the classification classifying the particles as a type specific to a control sample.

[0057] Example 2 The system of Example 1, wherein the set of operations includes performing a control-specific process on the sample and generating an analytical output for the sample, the analytical output being based on a result of the control-specific process.

[0058] Example 3 The system of Example 2, wherein the control-specific processing includes obtaining a second classification of the particles based on the classification of the particles, the second classification of the particles being a classification as a type not specific to the control sample, and generating an analytical output for the sample includes generating an analytical output based on the second classification of the particles.

[0059] Example 4 The system of Example 3, wherein the type that is specific to the control sample is a synthetic particle type and the type that is not specific to the control sample is a blood cell type.

[0060] Example 5 The system of Example 4, wherein the blood cell types are selected from neutrophils, lymphocytes, monocytes, eosinophils, and basophils.

[0061] Example 6 The system of Example 5, wherein the analytical output is selected from a 5 fraction output, a 6 fraction output, and a 7 fraction output.

[0062] Example 7 The system of Example 4, wherein the blood cell types are selected from red blood cells, nucleated red blood cells, reticulocytes, and platelets.

[0063] Example 8 The system of any of Examples 1-7, wherein the representation of particles from the sample is an image of particles from the sample.

[0064] Example 9 The system of Example 8, wherein obtaining a representation of particles from a sample includes transporting a portion of the sample containing the particles through a viewing zone of a camera using a flow cell, and capturing an image of the particles using the camera when the particles are in the viewing zone of the camera.

[0065] Example 10 The system of Example 9, wherein using a flow cell to transport a portion of the sample through a viewing zone of the camera includes surrounding the portion of the sample containing the particles within a sheathing fluid, the sheathing fluid transporting the particles to the viewing zone of the camera in a flow through the narrowing zone of the flow cell.

[0066] Example 11 The system of any of Examples 1-10, wherein obtaining a classification of the particles is performed using a first classifier, the first classifier being configured to classify particles from the control sample into a type specific to the control sample and to classify particles from the patient sample into a type not specific to the control sample, and the set of operations includes, for each of a plurality of additional particles from the sample, obtaining a representation of the particle and obtaining a classification of the particle using the first classifier, and determining that the sample is a control sample based on a majority of the classifications obtained for the particles from the sample classifying the particles into a type specific to the control sample, and based on determining that the sample is a control sample, control-specific processing is performed on the sample.

[0067] Example 12 The system of any of Examples 1-11, wherein the control-specific processing includes classifying a set of representations of particles from the sample using a classifier operable only to classify particles into types specific to the control sample.

[0068] Example 13 The system of any of Examples 1-12, wherein the set of operations includes identifying the sample as a control sample based on the classification of the particles.

[0069] Example 14 1. A method for verifying performance of an analyzer, the method comprising: providing a sample to the analyzer, the analyzer adapted to use a camera in analyzing the sample; obtaining representations of particles from the sample; obtaining a classification of the particles, the classification classifying the particles as a type specific to a control sample; and performing a control-specific process on the sample.

[0070] Example 15 The method of Example 14, the method including generating an analytical output for the sample, the analytical output being based on an outcome of the control-specific processing.

[0071] Example 16 The method of Example 15, wherein the control-specific processing includes obtaining a second classification of the particles based on the classification of the particles, the second classification of the particles being a classification as a type not specific to the control sample, and generating an analytical output for the sample includes generating an analytical output based on the second classification of the particles.

[0072] Example 17 The method of Example 16, wherein the type that is specific to the control sample is a synthetic particle type and the type that is not specific to the control sample is a blood cell type.

[0073] Example 18 The method of Example 17, wherein the blood cell types are selected from neutrophils, lymphocytes, monocytes, eosinophils, and basophils.

[0074] Example 19 The method of Example 17, wherein the blood cell types are selected from red blood cells, nucleated red blood cells, reticulocytes, and platelets.

[0075] Example 20 The method of any of Examples 15-18, wherein the analytical output is selected from a 5 fraction output, a 6 fraction output, and a 7 fraction output.

[0076] Example 21 The method of any of Examples 14 to 20, wherein the sample comprises particles in a carrier fluid.

[0077] Example 22 The method of example 21, wherein the particles are configured to represent blood cell types.

[0078] Example 23 The method of Example 21, wherein the particles comprise one or more detectable characteristics selected from size, light reflectivity, color, surface detection pattern, internal structure, and surface functionalization.

[0079] Example 24 The method of any of Examples 14-23, wherein obtaining a classification of the particles is performed using a first classifier, the first classifier being configured to classify particles from the control sample into a type specific to the control sample and to classify particles from the patient sample into a type not specific to the control sample, the method including, for each of a plurality of additional particles from the sample, obtaining a representation of the particle and obtaining a classification of the particle using the first classifier, and determining that the sample is a control sample based on a majority of the classifications obtained for particles from the sample classifying the particles into a type specific to the control sample, and based on determining that the sample is a control sample, control-specific processing is performed on the sample.

[0080] Example 25 Any of the methods of Examples 15-24, wherein the control-specific processing includes classifying a set of representations of particles from the sample using a classifier operable only to classify particles into types specific to the control sample.

[0081] Example 26 The method of Examples 14-25, wherein the representation of particles from the sample is an image of particles from the sample.

[0082] Example 27 The method of any of Examples 14-26, wherein obtaining a representation of particles from the sample comprises: transporting a portion of the sample containing the particles through a viewing zone of a camera using a flow cell; and capturing an image of the particles using the camera when the particles are in the viewing zone of the camera.

[0083] Example 28 The method of Example 27, wherein using a flow cell to transport a portion of the sample through a viewing zone of the camera includes surrounding the portion of the sample containing the particles within a sheathing fluid, the sheathing fluid transporting the particles to the viewing zone of the camera in a flow through the narrowing zone of the flow cell.

[0084] Example 29 The method of any of Examples 14-28, wherein the method includes identifying the sample as a control sample based on the classification of the particles.

[0085] Example 30 A non-transitory computer readable medium having stored thereon instructions operable, when executed by a processor, to cause an analyzer to perform a method according to any one of claims 14 to 29.

[0086] Each of the calculations or operations described herein may be performed using a computer or other processor having hardware, software, and / or firmware. Various method steps may be performed by modules, which may include any of a variety of digital and / or analog data processing hardware and / or software configured to perform the method steps described herein. Modules optionally comprise data processing hardware adapted to perform one or more of the steps by having appropriate machine programming code associated with those steps, and modules for two or more steps (or portions of two or more steps) may be integrated on a single processor board or separated on different processor boards in any of a variety of integrated and / or distributed processing architectures. These methods and systems often employ tangible media embodying machine-readable code having instructions for performing the above method steps. Suitable tangible media may include memory (including volatile and / or non-volatile memory), storage media (magnetic recording on floppy disks, hard disks, tapes, etc.; optical memory such as CD, CD-R / W, CD-ROM, DVD, etc.; or any other digital or analog storage medium), etc.

[0087] All patents, patent publications, patent applications, journal articles, books, technical references, and the like mentioned in this disclosure are hereby incorporated by reference in their entirety for all purposes.

[0088] Different configurations of the components shown in the drawings or described above, as well as components and steps not shown or described, are possible. Similarly, some features and subcombinations are useful and may be employed without reference to other features and subcombinations. Although embodiments of the invention have been described for purposes of illustration and not limitation, alternative embodiments will become apparent to the reader of this patent. In some cases, method steps or operations may be performed or executed in a different order, or operations may be added, deleted, or modified. It will be understood that in certain aspects of the invention, a single component may be replaced with multiple components, and multiple components may be replaced with a single component, to provide an element or structure or to perform a given function or functions. Except where such a substitution would not be operative to implement a particular embodiment of the invention, such substitution is deemed to be within the scope of the invention. Thus, the claims should not be treated as limited to the examples, drawings, embodiments, and descriptions provided above, but should be understood to have the scope provided when the terms are given their broadest reasonable interpretation as provided by common dictionaries, and should be understood to have the meaning that they have when used in the claims, unless a term or phrase is indicated to have a specific meaning under the heading "Explicit Definition."

[0089] Explicit definition It should be understood that in the above examples and in the claims, the phrase "based on" something else should be understood to mean that it is at least partially determined by what is indicated as being based on. If something indicates that it must be completely determined based on something else, it is described as "based exclusively on" that which is completely determined by it.

[0090] It should be understood that in the above examples and in the claims, the term "set" is to be understood as one or more items grouped together. [Explanation of symbols]

[0091] 21 Zone 21a Proximal flow path section 21b Distal flow channel portion 22 Flow Cell 23 Vision Zones 24 Imaging Device 25 Source of sample fluid 27 Source of Covering Fluid 28 Distal end 29 Sample supply tube 32 Sample flow 33 Emission 46 Objective Lens 48 CCD 401 Input Image 402a~402n stages 403 Final output image 404 Neural Networks 501 Input 502 layers 502a First Layer 503a~503n Converted images 504a~504n Filters 505a~505n Pool images 506 Output Images

Claims

1. 1. A system for cell sorting, comprising: Camera and A processor; A non-transitory computer-readable medium storing instructions that, when executed by the processor, Obtaining a particle representation from a sample; and obtaining a classification of said particles, said classification classifying said particles as a type specific to a control sample; and a non-transitory computer-readable medium operable to perform a set of actions including:

2. The set of operations is subjecting said sample to a control specific treatment; generating an analytical output for said sample, said analytical output being based on an outcome of said control-specific processing; The system of claim 1 , comprising:

3. the control specific processing includes obtaining a second classification of the particles based on the classification of the particles; the second classification of the particle as a type not unique to a control sample; The system of claim 2 , wherein generating the analytical output for the sample comprises generating the analytical output based on the second classification of the particles.

4. the type specific to the control sample is a synthetic particle type; The system of claim 3 , wherein the type not specific to the control sample is a blood cell type.

5. The blood cell type is neutrophils, Lymphocytes, Monocytes, Eosinophils, and basophil The system of claim 4 , wherein the selected

6. The system of claim 5 , wherein the analysis output is selected from a 5 fraction output, a 6 fraction output, and a 7 fraction output.

7. The blood cell type is red blood cells, nucleated red blood cells, Reticulocytes, and platelet The system of claim 4 , wherein the selected

8. The system of claim 1 , wherein the representation of the particle from the sample is an image of the particle from the sample.

9. Obtaining the representation of the particles from the sample, transporting a portion of the sample containing the particles through a viewing zone of the camera using a flow cell; and capturing the image of the particle using the camera when the particle is in the viewing zone of the camera.

10. 10. The system of claim 9, wherein using the flow cell to transport the portion of the sample through the viewing zone of the camera includes surrounding the portion of the sample including the particle within a sheath fluid, the sheath fluid carrying the particle to the viewing zone of the camera in a flow through a narrowing zone of the flow cell.

11. Obtaining the classification of the particles is performed using a first classifier, the first classifier configured to classify particles from a control sample into types specific to the control sample and to classify particles from a patient sample into types not specific to the control sample; The set of operations is For each of a plurality of additional particles from the sample, Obtaining a representation of that particle, and obtaining a classification of the particle using the first classifier; determining that the sample is a control sample based on a majority of the classifications obtained for particles from the sample classifying the particles into a type unique to the control sample; The system of claim 1 , wherein the control-specific processing is performed on the sample based on determining that the sample is a control sample.

12. 2. The system of claim 1, wherein the control-specific processing comprises classifying a set of representations of particles from the control sample using a classifier operable only to classify particles into types specific to the control sample.

13. The system of claim 1 , wherein the set of actions includes identifying the sample as a control sample based on the classification of the particles.

14. 1. A method for verifying analyzer performance, comprising: providing a sample to an analyzer, the analyzer being adapted to use a camera in analyzing the sample; obtaining a representation of particles from said sample; obtaining a classification of the particles, the classification classifying the particles as a type specific to a control sample.

15. The method further comprising: subjecting said sample to a control specific treatment; 15. The method of claim 14, comprising generating an analytical output for the sample, the analytical output being based on an outcome of the control-specific processing.

16. the control specific processing includes obtaining a second classification of the particles based on the classification of the particles; the second classification of the particle as a type not unique to a control sample; 16. The method of claim 15, wherein generating the analytical output for the sample comprises generating the analytical output based on the second classification of the particles.

17. the type specific to the control sample is a synthetic particle type; 17. The method of claim 16, wherein the type not unique to the control sample is a blood cell type.

18. The blood cell type is neutrophils, Lymphocytes, Monocytes, Eosinophils, and basophil 18. The method of claim 17, wherein said compound is selected from the group consisting of

19. The blood cell type is red blood cells, nucleated red blood cells, Reticulocytes, and platelet 18. The method of claim 17, wherein said compound is selected from the group consisting of

20. 16. The method of claim 15, wherein the analysis output is selected from a 5 fraction output, a 6 fraction output, and a 7 fraction output.

21. The method of claim 14 , wherein the sample comprises the particles in a carrier fluid.

22. The method of claim 21 , wherein the particles are configured to represent blood cell types.

23. 22. The method of claim 21, wherein the particles comprise one or more detectable characteristics selected from size, light reflectivity, color, surface detection pattern, internal structure, and surface functionalization.

24. Obtaining the classification of the particles is performed using a first classifier, the first classifier configured to classify particles from a control sample into types specific to the control sample and to classify particles from a patient sample into types not specific to the control sample; The method further comprising: For each of a plurality of additional particles from the sample, Obtaining a representation of that particle, and obtaining a classification of the particle using the first classifier; determining that the sample is a control sample based on a majority of the classifications obtained for particles from the sample classifying the particles into a type unique to the control sample; 16. The method of claim 15, wherein the control-specific processing is performed on the sample based on determining that the sample is a control sample.

25. 16. The method of claim 15, wherein the control-specific processing comprises classifying a set of representations of particles from the sample using a classifier operable only to classify particles into types specific to a control sample.

26. The method of claim 14 , wherein the representation of the particles from the sample is an image of the particles from the sample.

27. Obtaining the representation of the particles from the sample, transporting a portion of the sample containing the particles through a viewing zone of the camera using a flow cell; and capturing the image of the particle using the camera when the particle is in the viewing zone of the camera.

28. 28. The method of claim 27, wherein using the flow cell to transport the portion of the sample through the viewing zone of the camera includes surrounding the portion of the sample including the particle within a sheath fluid, the sheath fluid carrying the particle to the viewing zone of the camera in a flow through a narrowing zone of the flow cell.

29. The method of claim 14 , wherein the method comprises identifying the sample as a control sample based on the classification of the particles.

30. A non-transitory computer readable medium having stored thereon instructions operable, when executed by a processor, to cause an analyzer to perform the method of any of claims 14 to 29.

Images