Identification of immature cell types using imaging

Abstract

The maturity of the cells may be detected using a computer-implemented image analysis method that includes capturing an image of the stained blood cells with a camera and then determining the maturity of the stained blood cells. The determination may include classifying the stained blood cells prior to generating maturity of the stained blood cells. Corresponding systems may also be implemented, and methods for determining maturity of cells may be used to detect other characteristics, such as infection by malaria or other parasites.

Description

Background Art

[0001] Blood cell analysis is one of the most commonly performed medical tests for providing an overview of a patient's health status. A blood sample can be extracted from a patient and stored in a test tube containing an anticoagulant to prevent clotting. The blood sample may contain a variety of particles whose identification may be clinically useful. For example, young platelets (called reticulocytes or immature platelets) containing increased levels of mRNA and rRNA compared to mature cells typically constitute 4.5% or less of the total platelets in the sample. However, an increase in the proportion of immature platelets may be an indicator of thrombopoiesis and may be an early indicator of bone marrow recovery in patients receiving chemotherapy and stem cell transplants. As another example, as red blood cells develop, they pass through an intermediate stage called reticulocytes. These reticulocytes can be classified into immature reticulocytes and mature reticulocytes, and the ratio of immature reticulocytes to total reticulocytes (called the immature reticulocyte fraction or IRF) has been found to have clinical utility, for example as described in Mitrani. et. al., The Immature Reticulocyte Fraction As an Aid in the Diagnosis and Prognosis of Parvovirus B19 Infection in Sickle Cell Disease, BLOOD (2018) 132 (Supplement 1); 3678.

[0002] The identification of particles such as platelets and reticulocytes may be difficult due to the relatively small size of the particles. Quantification between the mature and immature states of cells may be more challenging due to factors such as the size of the cells and the relatively small differences associated with specific particle types (e.g., immature platelets versus mature platelets, or immature reticulocytes versus mature reticulocytes). Such identification may be difficult in many sample analysis systems, including in image-based systems. Therefore, there is a need for a technology that can identify particles such as immature platelets and reticulocytes based on data collected by an image-based sample processing system. Summary of the Invention

[0003] Described herein are devices, systems, and methods for identifying objects, such as immature reticulocytes and platelets, in data captured by an image-based sample processing system.

[0004] An illustrative implementation of such technology involves a computer-implemented image analysis method for detecting the maturity of blood cells. As described herein, such an image analysis method can include capturing images of stained blood cells with a camera, classifying the stained blood cells, and determining the maturity of the stained blood cells after the stained blood cells have been classified.

[0005] A first aspect of the present invention relates to a computer-implemented image analysis method for detecting the maturity of blood cells, the method comprising: capturing an image of stained blood cells with a camera; classifying the stained blood cells, for example, by using the image of the stained blood cells; and determining the maturity of the stained blood cells after the stained blood cells are classified, for example, by using the classification of the stained blood cells.

[0006] Specifically, according to the present invention, classifying the stained blood cells includes determining the blood cell type of the stained blood cells. For example, the blood cell type of the stained blood cells can be at least one of white blood cells (WBCs), platelets, and reticulocytes. Alternatively or additionally, according to the present invention, classifying the stained blood cells includes selecting the blood cell type of the stained blood cells from a group of blood cell types, wherein the group of blood cell types includes at least WBCs, platelets, and reticulocytes.

[0007] Illustratively, the determination of the maturity of the stained blood cells is based on the blood cell type of the stained blood cells.

[0008] The method according to the first aspect of the present invention may further include determining whether the stained blood cell is a blood cell of interest based on the blood cell type of the stained blood cell. In this case, if the stained blood cell is a blood cell of interest, determining the maturity of the stained blood cell is performed. If the stained blood cell is not a blood cell of interest, the method may include discarding and / or removing the image of the stained blood cell. For example, if the blood cell type of the stained blood cell is a platelet or a reticulocyte, the stained blood cell is a blood cell of interest. Specifically, if the blood cell type of the stained blood cell is neither a platelet nor a reticulocyte, for example, if its blood cell type is a WBC, the stained blood cell is not a blood cell of interest.

[0009] Exemplarily, the method according to the first aspect of the present invention further includes: treating the blood sample with a lysing agent, a stain, or both; flowing the blood sample through a flow cell and past a camera; and capturing an image of stained blood cells with the camera while the blood sample is flowing through the flow cell and past the camera. Specifically, in this example, the stained blood cells are contained in the blood sample.

[0010] In some examples, classifying the stained blood cells includes classifying the stained blood cells as at least one of reticulocytes or platelets.

[0011] The method according to the first aspect of the invention further comprises determining at least one of an immature reticulocyte fraction and an immature platelet fraction of the blood sample.

[0012] Specifically, determining at least one of an immature reticulocyte fraction and an immature platelet fraction of a blood sample can be performed using an image of stained blood cells and a plurality of images, each of the plurality of images depicting a corresponding stained blood cell of the blood sample. For example, determining at least one of an immature reticulocyte fraction and an immature platelet fraction of a blood sample can include, for each of the plurality of images, classifying the corresponding stained blood cell depicted in the each image, and after the corresponding stained blood cell is classified, determining a maturity of the corresponding stained blood cell, for example, by using the classification of the corresponding stained blood cell.

[0013] For example, determining the immature reticulocyte fraction may include determining a plurality of images in the plurality of images that depict immature reticulocytes. Alternatively or additionally, determining the immature reticulocyte fraction may include determining a plurality of images in the plurality of images that depict immature platelet cells.

[0014] For example, classifying the stained blood cells may include utilizing an artificial intelligence or machine learning model. Specifically, classifying the stained blood cells may include determining the blood cell type of the stained blood cells by using a convolutional neural network and at least a portion of an image of the stained blood cells. For example, the convolutional neural network is configured, for example, trained, to process at least a portion of the image of the stained blood cells and thereby determine the blood cell type. Specifically, the convolutional neural network is a classifier that is configured, for example, trained, to assign a blood cell type from a set of blood cell types to the stained blood cells by using at least a portion of the image of the stained blood cells as input. Specifically, the portion of the image includes a set of pixels of the image. The set of pixels may be a proper subset of the pixels of the image, or may include all pixels of the image.

[0015] Additionally or alternatively, classifying the stained blood cells includes pixel analysis involving, for example, HSV space. Specifically, according to the present invention, the foreground of an image is a set of pixels in the image, specifically a portion of the image. In some examples, classifying the stained blood cells includes determining a set of image intensities, and identifying the foreground in the image based on the set of image intensities.

[0016] Illustratively, the method according to the first aspect of the present invention further includes: calculating an average red intensity of pixels in the foreground of the image and an average blue intensity of pixels in the foreground of the image; and classifying the stained blood cells based on the average red intensity and the average blue intensity. Specifically, the average blue intensity and the average red intensity are included in the aforementioned set of image intensities.

[0017] Exemplarily, determining the maturity of the stained blood cells includes utilizing pixel foreground analysis. Specifically, according to the present invention, the foreground count is the number of pixels of the image of the stained blood cells included in the foreground.

[0018] For example, determining the maturity of the stained blood cells includes determining whether the foreground count satisfies certain conditions in a set of maturity conditions (e.g., each condition in the set of maturity conditions, or a majority of the conditions in the set of maturity conditions). If the foreground count satisfies certain conditions in the set of maturity conditions, the stained blood cells are considered mature. Conversely, if the foreground count does not satisfy the maturity conditions (e.g., does not satisfy each condition, or does not satisfy a majority of the conditions), the stained blood cells are considered immature.

[0019] Alternatively, determining the maturity of the stained blood cells includes determining whether the foreground count satisfies each of a set of immaturity conditions. If the foreground count satisfies all of the set of immaturity conditions, the stained blood cells are considered immature. Conversely, if the foreground count does not satisfy at least one of the set of immaturity conditions, the stained blood cells are considered mature.

[0020] For example, determining the maturity of the stained blood cells includes determining whether the pixel foreground count exceeds a threshold amount. In this case, specifically, the set of immaturity conditions includes a condition that the pixel foreground count exceeds the threshold amount.

[0021] Illustratively, the method includes determining that the stained blood cell is immature based on determining that a pixel foreground count exceeds a threshold amount.

[0022] A second aspect of the present invention relates to an image analysis system for detecting the maturity of blood cells, the system comprising: a camera configured to capture images of stained blood cells; a processor configured to classify the stained blood cells (e.g., using the images of the stained blood cells); and, after the stained blood cells are classified, determining the maturity of the stained blood cells, e.g., using the classification of the stained blood cells. Alternatively, the maturity determination step can be part of the classification step (e.g., where immature cell types are part of an initial classification step, e.g., a primary classifier trained to distinguish immature cell types).

[0023] The image analysis system may further include a flow cell, and the camera may be configured to capture images of stained blood cells as the blood sample flows through the flow cell and past the camera. Specifically, the flow cell and the camera are arranged relative to each other such that the flow cell is configured to transport at least a portion of the sample fluid through a viewing area of ​​the camera.

[0024] Exemplarily, the processor is configured to classify the stained blood cells into at least one of reticulocytes or platelets. Alternatively or additionally, the processor is further configured to determine at least one of an immature reticulocyte fraction and an immature platelet fraction of the blood sample.

[0025] Additionally or alternatively, the processor may be configured to classify the stained blood cells using a convolutional neural network.

[0026] Additionally or alternatively, the processor may be configured to classify the stained blood cells based on identifying foreground in the image based on the intensity of each pixel in the image.

[0027] For example, the processor is further configured to calculate the average red intensity of pixels in the foreground of the image and the average blue intensity of pixels in the foreground of the image. In addition, the processor can be configured to classify stained blood cells based on the average red intensity and the average blue intensity.

[0028] Illustratively, the processor is configured to determine the maturity of the stained blood cells based on pixel foreground counts.

[0029] Specifically, the processor is configured to determine that the stained blood cells are immature based on the pixel foreground count exceeding a threshold amount.

[0030] For example, the image analysis system according to the first aspect of the present invention is configured to perform the method according to the second aspect of the present invention.

[0031] A third aspect of the present invention relates to a machine comprising: a camera; and means for classifying cells in an image captured by the camera and determining the maturity of the cells.

[0032] A fourth aspect of the present invention relates to a computer program product comprising instructions which, when executed by a computer, cause the computer to perform the method according to the first aspect of the present invention.

[0033] A fifth aspect of the invention relates to a computer-readable medium, such as a transitory computer-readable medium, comprising instructions which, when executed by a computer, cause the computer to perform the method according to the first aspect of the invention.

[0034] Additional aspects of the present invention include modules for assisting in the imaging of biological samples. The illumination module is configured to illuminate the sample to provide favorable lighting conditions for an image capture device (e.g., a camera) to capture images of the sample cells. The staining module is configured to stain biological cells (e.g., the interior or nuclear region of a blood cell) to visualize the interior region of the cell, thereby assisting in cell classification and / or determination of maturity.

[0035] It should be noted that any of the various features of the aspects disclosed herein may be included or combined in each of these aspects.

[0036] Although a number of examples are described herein, other examples of the described subject matter will become apparent to those skilled in the art from the following detailed description and accompanying drawings, which show and describe illustrative examples of the disclosed subject matter. As will be appreciated, the disclosed subject matter is capable of modifications in various aspects, all without departing from the spirit and scope of the described subject matter. Accordingly, the drawings and detailed description are to be regarded as illustrative in nature and not restrictive. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] While the specification concludes with claims which particularly point out and distinctly claim the invention, it is believed the invention will be better understood from the following description of certain examples taken in conjunction with the accompanying drawings in which like reference numerals represent like elements, and in which:

[0038] Figure 1 is a schematic illustration, partially in cross-section and not to scale, showing operational aspects of an exemplary flow cell and a high optical resolution imaging device for sample image analysis using digital image processing.

[0039] Figure 2 A slide-based visual inspection system is shown in which aspects of the disclosed technology can be used.

[0040] Figure 3 A perspective view of an exemplary illumination module is shown along with another exemplary flow cell and a high optical resolution imaging device for sample image analysis using digital image processing.

[0041] Figure 4 Shown Figure 3 A perspective view of a lighting module is shown in FIG. 1 , wherein a portion of the housing is omitted to illustrate the light emitter, focusing lens, dichroic element, and collimating lens of the lighting module.

[0042] Figure 5 Shown Figure 3 A top plan view of the lighting module, showing light traveling from each of the light emitters to the collimating lens.

[0043] Figure 6A A perspective view of an exemplary staining module for mixing a stain with a sample to form a sample mixture and for incubating the sample mixture is shown, illustrating the lamination of a ferromagnetic sheet to the housing of the staining module.

[0044] Figure 6B Shown Figure 6A A perspective view of the dyeing module, showing the ferromagnetic sheet being wrapped to the housing with tape.

[0045] Figure 6C Shown Figure 6A A perspective view of a dyeing module showing the winding of the heating coil of the dyeing module around the housing.

[0046] Figure 7A Shown is a side elevation view of an exemplary multi-chamber staining module for mixing a stain with a sample to form a sample mixture and for incubating the sample mixture.

[0047] Figure 7B Shown Figure 7A A top plan view of the multi-chamber staining module.

[0048] Figure 8 A process that can be used to stain a sample is shown.

[0049] Figure 9 A process is shown that can be used to determine the maturity of blood cells and to determine the maturity of particles in a sample.

[0050] Figure 10 A process that can be used to classify particles based on morphological features of imaged cells is shown.

[0051] Figure 11 An example machine learning model is shown.

[0052] Figure 12 Shown as an example may include Figure 11 An example of the layers in the machine learning model shown.

[0053] Figure 13 A method that can be used to generate and image the maturity of cells is shown.

[0054] Figure 14 A method of using maturity information to generate overall information of a biological sample is shown.

[0055] Figure 15 Depicted is a method for creating a maturity scale for cells as part of classifying cells.

[0056] The accompanying drawings are not intended to be limiting in any way, and it is contemplated that various embodiments of the invention may be carried out in a variety of other ways, including ways not necessarily depicted in the drawings. The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate several aspects of the invention and, together with the description, serve to explain the principles of the invention; it should be understood, however, that the invention is not limited to the precise arrangements shown. DETAILED DESCRIPTION

[0057] The present disclosure relates to devices, systems, compositions, and methods for analyzing samples containing particles. In one embodiment, the present invention relates to an automated particle imaging system comprising an analyzer, which may be, for example, a visual analyzer. In some embodiments, the visual analyzer may further comprise a processor for facilitating automated analysis of the image.

[0058] According to some aspects of the present disclosure, a system for obtaining an image of a sample including particles suspended in a liquid including a visual analyzer can be provided. Such a system can, for example, be used to characterize particles in a biological fluid, such as detecting and quantifying red blood cells, immature reticulocytes, mature reticulocytes, nucleated red blood cells, immature platelets, and / or white blood cells, including white blood cell differential counting, classification and subclassification, and analysis. Other similar uses, such as characterizing blood cells from other fluids and / or identifying parasites such as malaria, are also contemplated.

[0059] The identification of various types of particles in a blood sample is an exemplary application for which the present subject matter is particularly well suited, although other types of body fluid samples may also be used. For example, various aspects of the disclosed technology may be used to analyze non-blood body fluid samples including blood cells (e.g., white blood cells and / or red blood cells), such as serum, bone marrow, lavage fluid, exudate, transudate, cerebrospinal fluid, pleural effusion, peritoneal fluid, and amniotic fluid. The sample may also be a solid tissue sample, such as a biopsy sample that has been processed to produce a cell suspension. The sample may also be a suspension obtained by processing a fecal sample. The sample may also be a laboratory or production line sample including particles, such as a cell culture sample. The term sample may be used to refer to a sample obtained from a patient or laboratory, or any fragment, portion, or aliquot thereof. In some processes, the sample may be diluted, divided into multiple portions, or stained.

[0060] In some aspects, sample is presented, imaged and analyzed in an automated manner. In the case of a blood sample, suitable diluents or saline solutions can be used to fully dilute the sample, which reduces the extent to which the view of some cells may be hidden by other cells in the sample that is not diluted or less diluted. Reagents that enhance the contrast of certain cells can be used to process cells, for example, permeabilizing agents are used to make the cell membrane permeable, and histological stains are adhered thereto and display features such as particles and cell nuclei. In some cases, it may be desirable to dye aliquots of the sample to count and characterize the particles comprising reticulocytes, nucleated red blood cells and platelets, and to classify, characterize and analyze white blood cells. In other cases, the sample containing red blood cells can be diluted before introducing a flow trough and / or imaging in a flow trough or otherwise.

[0061] The details of the sample preparation apparatus and method for sample dilution, permeabilization, and histological staining can generally be accomplished using precision pumps and valves operated by one or more programmable controllers. Examples can be found in patents such as U.S. Patent No. 7,319,907. Similarly, techniques for distinguishing certain cell classifications and / or subclassifications by attributes such as relative size and color can be found in U.S. Patent No. 5,436,978 on white blood cells. The disclosures of these patents are incorporated herein by reference in their entirety.

[0062] I. System Overview

[0063] Turning now to the accompanying drawings, Figure 1 An exemplary flow cell 22 is schematically shown for conveying a sample fluid through an observation area 23 of a high optical resolution imaging device 24 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 subjected to processing such as contact with a particulate contrast agent, composition, and heating. The flow cell 22 is also coupled to one or more sources 27 of particles and / or intracellular organelle aligning liquid (PIOAL) (e.g., a clear glycerol solution having a viscosity greater than that of the sample fluid).

[0064] The sample fluid is injected through the flat opening at the distal end 28 of the sample feed tube 29 and enters the interior of the flow cell 22 at a point where the PIOAL flow is substantially established, resulting in a stable and symmetrical laminar flow of the PIOAL above and below (or on opposite sides) the ribbon-like sample flow. The sample flow and PIOAL flow can be supplied by a precision metering pump that moves the PIOAL along the substantially narrowed flow channel along with the injected sample fluid. The PIOAL envelops and compresses the sample fluid in the region 21 where the flow channel narrows. Thus, the reduced flow channel thickness at region 21 can aid in the geometric focusing of the sample flow 32. The sample fluid ribbon 32 is enveloping and carried along with the PIOAL downstream of the narrowed region 21, passing in front of or otherwise passing through the observation region 23 of the high-optical-resolution imaging device 24, where images are collected, for example, using a charge-coupled device (CCD) 48. In this way, flow imaging is performed in which images are collected from a flowing sample stream and the cellular material contained therein. The processor 18 may receive as input pixel data from the CCD 48. The sample fluid stream flows into the discharge port 33 along with the PIOAL.

[0065] As shown here, the narrowed region 21 can have a proximal flow channel portion 21a having a proximal thickness PT and a distal flow channel portion 21b having a distal thickness DT, with the distal thickness DT being less than the proximal thickness PT. Thus, sample liquid can be injected through the distal end 28 of the sample tube 29 at a location distal to the proximal portion 21a and proximal to the distal portion 21b. Thus, the sample liquid can enter the PIOAL envelope when the PIOAL flow is compressed by the region 21. The sample liquid injection tube has a distal outlet through which the sample liquid is injected into the flowing sheath fluid, and the distal outlet is defined by the reduction in the flow channel dimension of the flow cell.

[0066] A digital high optical resolution imaging device 24 having an objective lens 46 is oriented along an optical axis that intersects the strip-shaped sample flow 32. The relative distance between the objective lens 46 and the flow cell 22 can be varied by operation of a motor drive 54 for resolving and collecting focused digitized images on the photosensor array. A detailed description of methods such as flow cell systems and methods is provided in U.S. Patent No. 9,322,752, filed on March 17, 2014, entitled "Flow cell Systems and Methods for Particle Analysis in Blood Samples." Figure 1 The disclosure of U.S. Pat. No. 5,877,679, filed March 28, 2012, is incorporated herein by reference in its entirety for additional information regarding the construction and operation of the exemplary flow cell shown.

[0067] Aspects of the disclosed technology may also be applied to Figure 1 In contexts other than the flow cell system shown. For example, Figure 2 FIG2 shows a slide-based visual inspection system 200 in which aspects of the disclosed technology may be used. Figure 2 In the system shown, a slide 202 including a sample (e.g., a blood sample) is placed in a slide holder 204. The slide holder 204 may be adapted to hold multiple slides or only one slide, such as Figure 2 An image capture device 206 including an optical system 208 and an image sensor 210 is adapted to capture image data depicting a sample in the slide 202 .

[0068] Image data captured by the image capture device 206 can be transmitted to an image processing device 212. The image processing device 212 can be an external device, such as a personal computer, connected to the image capture device 206. Alternatively, the image processing device 212 can be incorporated into the image capture device 206. The image processing device 212 can include a processor 214 associated with a memory 216, and the processor 214 is configured to determine the changes required to determine the difference between the actual focus of the image capture device 206 and the correct focus. When the difference is determined, instructions can be transmitted to the steering motor system 218. The steering motor system 218 can change the distance z between the slide 202 and the optical system 208 based on the instructions from the image processing device 212. Descriptions of methods that can be used for focusing using this type of setup are provided in U.S. Patent 9,857,361, entitled “Flowcell, sheath fluid, and autofocus systems and methods for particle analysis in urine samples,” issued on January 2, 2018, and U.S. Patent 10,705,008, entitled “Autofocus systems and methods for particle analysis in blood samples,” issued on July 7, 2020, the disclosures of which are incorporated herein by reference in their entirety.

[0069] II. Example of Lighting Module

[0070] In the context of imaging (including the flow imaging concept discussed for bioimaging), proper illumination is important to enable proper visualization of biological material (e.g., blood cells). Illumination is an important criterion for image capture devices (e.g., cameras) to capture clear and well-illuminated images of samples, for example, to enable algorithms to properly identify cell types.

[0071] In such Figure 1 or Figure 2 In the system shown, Figure 3 The illustrated lighting module (also referred to as a lighting system or lighting device) 300 can be used to illuminate an image captured by a camera (e.g., Figure 1 High optical resolution imaging device 24 or Figure 2 For example, the lighting module 300 may replace the image sensor 210 of the cell. Figure 1 The light source 42 is shown in combination. Figure 3 Illumination module 300 is shown along with an exemplary flow cell 302 that can be imaged, and a high optical resolution imaging device 304. Figure 1As shown in the flow cell 22, the high optical resolution imaging device 304 can be configured and operated as Figure 1 As shown, illumination module 300 is located on a side of flow cell 302 opposite high optical resolution imaging device 304 and is configured and operated like high optical resolution imaging device 24. As shown, illumination module 300 is located on an opposite side of flow cell 302 from high optical resolution imaging device 304 and is used to illuminate an analysis region, such as an observation region (also referred to as an image capture region or imaging region), of flow cell 302 as the sample moves through the analysis region to facilitate capturing an image of the sample by high optical resolution imaging device 304. Cells of the sample may be stained prior to moving through flow cell 302 via a staining module, such as staining modules 400 and 500 described below.

[0072] In the example shown, the lighting module 300 includes a housing 310, a plurality of light emitters 312a, 312b, 312c, a plurality of focusing lenses 314a, 314b, 314c, a plurality of dichroic elements 316a, 316b, 316c, and a collimating lens 318. Each of the light emitters 312a, 312b, 312c can be any suitable light source, including, for example, an arc lamp, a light emitting diode (LED), or any other suitable light emitter for providing pulsed or continuous illumination. In some embodiments, each of the light emitters 312a, 312b, 312c can be configured to emit light of a different color than the other light emitters 312a, 312b, 312c. For example, the first light emitter 312a may include a red LED configured to emit red light having a wavelength between about 600 nanometers and about 650 nanometers, for example, about 620 nanometers; the second light emitter 312b may include a green LED configured to emit green light having a wavelength between about 470 nanometers and about 600 nanometers, for example, about 525 nanometers; and / or the third light emitter 312c may include a blue LED configured to emit blue light having a wavelength between about 400 nanometers and about 470 nanometers, for example, about 450 nanometers.

[0073] The light emitters 312a, 312b, 312c are each mounted to a side of the housing 310 in a row that extends generally parallel to the optical axis of the high optical resolution imaging device 304, such that light emitted by each light emitter 312a, 312b, 312c may be initially projected into the interior of the housing 310 in a direction generally perpendicular to the optical axis of the high optical resolution imaging device 304. As shown, each focusing lens 314a, 314b, 314c is mounted within the housing 310 and axially aligned with a corresponding one of the light emitters 312a, 312b, 312c for focusing the light emitted by the corresponding light emitter 312a, 312b, 312c.

[0074] Each dichroic element 316a, 316b, 316c is mounted within the housing 310 in alignment with a corresponding one of the light emitters 312a, 312b, 312c for reflecting and / or filtering light emitted from one or more of the light emitters 312a, 312b, 312c (and focused by the corresponding focusing lens 314a, 314b, 314c). In this regard, each dichroic element 316a, 316b, 316c of the present example includes a corresponding reflective side 320a, 320b, 320c and a corresponding filtering side 322a, 322b, 322c. Each dichroic element 316a, 316b, 316c is oriented obliquely relative to the optical axis of the high optical resolution imaging device 304 and relative to the light received from the corresponding focusing lens 314a, 314b, 314c. For example, each dichroic element 316a, 316b, 316c may be oriented at an angle of approximately 45 degrees relative to the optical axis of the high optical resolution imaging device 304. More specifically, each dichroic element 316a, 316b, 316c is oriented such that the corresponding reflective side 320a, 320b, 320c generally faces toward both the corresponding focusing lens 314a, 314b, 314c and the high optical resolution imaging device 304, while the corresponding filtering side 322a, 322b, 322c generally faces away from both the corresponding focusing lens 314a, 314b, 314c and the high optical resolution imaging device 304.

[0075] In this way, the reflective side 320a, 320b, 320c of each dichroic element 316a, 316b, 316c can be configured to reflect light emitted from the corresponding light emitter 312a, 312b, 312c (and focused by the corresponding focusing lens 314a, 314b, 314c) and traveling generally perpendicular to the optical axis of the high optical resolution imaging device 304, so that the reflected light travels generally parallel to the optical axis of the high optical resolution imaging device 304. For example, the reflective side 320a of the first dichroic element 316a can be configured to reflect light emitted from the first light emitter 312a (and focused by the first focusing lens 314a) such that the reflected light travels generally parallel to the optical axis of the high optical resolution imaging device 304; the reflective side 320b of the second dichroic element 316b can be configured to reflect light emitted from the second light emitter 312b (and focused by the second focusing lens 314b) such that the reflected light travels generally parallel to the optical axis of the high optical resolution imaging device 304; and / or the reflective side 320c of the third dichroic element 316c can be configured to reflect light emitted from the third light emitter 312c (and focused by the third focusing lens 314c) such that the reflected light travels generally parallel to the optical axis of the high optical resolution imaging device 304.

[0076] Additionally, the filter sides 322a, 322b, 322c of at least some of the dichroic elements 316a, 316b, 316c can be configured to filter light received from one or more of the other dichroic elements 316a, 316b, 316c. For example, the filter side 322b of the second dichroic element 316b can be configured to filter light reflected from the first dichroic element 316a; and / or the filter side 322c of the third dichroic element 316c can be configured to filter light reflected from the second dichroic element 316b, and / or filter light reflected from the first dichroic element 316a (and filtered by the second dichroic element 316b). In this regard, the filter side 322a, 322b, 322c of each dichroic element 316a, 316b, 316c can be configured to suppress the passage of light having a wavelength below a corresponding predetermined threshold. For example, the filter side 322b of the second dichroic element 316b can be configured to reject light having a wavelength below a predetermined threshold of approximately 596 nanometers, and / or the filter side 322c of the third dichroic element 316c can be configured to reject light having a wavelength below a predetermined threshold of approximately 484 nanometers. In some cases, the filter sides 322b, 322c of the second and third dichroic elements 316b, 316c can be configured to allow approximately 95% of light having a wavelength above the corresponding predetermined threshold to pass through and / or reject approximately 99% of light having a wavelength below the corresponding predetermined threshold to pass through.

[0077] Thus, light emitted by the first light emitter 312a can be focused by the first focusing lens 314a, reflected by the reflective side 320a of the first dichroic element 316a, filtered by the filtering side 322b of the second dichroic element 316b, and filtered by the filtering side 322c of the third dichroic element 316c; light emitted by the second light emitter 312b can be focused by the second focusing lens 314b, reflected by the reflective side 320b of the second dichroic element 316b, and filtered by the filtering side 322c of the third dichroic element 316c; and / or light emitted by the third light emitter 312c can be focused by the third focusing lens 314c and reflected by the reflective side 320c of the third dichroic element 316c. In this way, the light emitted by the light emitters 312a, 312b, 312c can be tuned via the dichroic elements 316a, 316b, 316c to increase the whiteness of the light before being focused together by the collimating lens 318 into a single collimated white light beam, which can then be transmitted out of the housing 310 to the flow cell 302.

[0078] It should be understood that where the first light emitter 312a includes a red LED, the red light emitted by the first light emitter 312a may be substantially unaffected by the filtering sides 322b, 322c of the second and third dichroic elements 316b, 316c because the wavelength of red light is relatively high, which may be greater than the threshold of either filtering side 322b, 322c.

[0079] The collimated white light beam formed by the collimating lens 318 can be transmitted to the flow cell 302 via a light pipe (also known as a lighting column or light guide) (e.g., a hexagonal light pipe). The light pipe can be configured to collect the collimated white light beam, randomize the collimated white light beam, and / or converge the collimated white light beam onto the flow cell 302 (e.g., at an observation area of ​​the flow cell 302). The light pipe can be positioned relative to the collimating lens 318 so that the collimated light beam converges to a point (e.g., phase) at the entrance of the light pipe. The light pipe can be mounted to a flow cell holder (not shown) that holds the flow cell 302 to securely fix the outlet of the light pipe relative to the flow cell 302, such as the flow cell holder 700 described below. In this way, the distance between the light pipe and the flow cell 302 is fixed because the light pipe and the flow cell 302 are operably connected by the flow cell holder 700.

[0080] In some embodiments, light emitters 312a, 312b, and 312c can be configured to provide pulsed illumination in synchronization with one another (e.g., simultaneously) to capture a still image of a sample cell moving through flow cell 302. For example, the objective lens of high-optical-resolution imaging device 304 can be turned on; light emitters 312a, 312b, and 312c can then simultaneously emit light pulses; an image of the sample cell can then be captured by high-optical-resolution imaging device 304; and the objective lens of high-optical-resolution imaging device 304 can then be turned off. This process can be repeated for any suitable number of iterations. In some cases, the duration of each pulse can be between approximately 1 microsecond and approximately 3 microseconds, such as approximately 2 microseconds. The pulse frequency can depend on the camera frame acquisition frequency, which can be between approximately 220 frames per second and approximately 300 frames per second. For example, 220 frames per second can correspond to one image approximately every 4.5 milliseconds. The objective lens can be turned on for approximately 100 microseconds, which may be sufficient to capture a single image. In some embodiments, higher frame rates can be used, such as with a reduced field of view. Depending on the specific application, the type of camera used, etc., an increased speed may be used. It will be appreciated that an increase in speed may provide more data in less time (e.g., more images of more sample cells), while a smaller field of view may remove visual artifacts (e.g., "black bars") that may be used for focusing. It will also be appreciated that pixel rate may contribute to the amount of data provided. For example, increasing the pixel rate may compensate for reducing the frame rate to provide the same amount of data.

[0081] In some embodiments, light emitters 312a, 312b, 312c can be configured to sequentially emit light pulses to operate in a diagnostic mode. For example, a time delay can be provided between each flash to capture three separate images of a particular sample cell at three different moments along the path of the sample cell. The pixel representation of the distance can then be used to calculate the velocity of the sample cell to determine whether the sample cell is accelerating or decelerating and / or to determine whether the flow is too fast to obtain reliable data. This diagnostic mode can be selectively entered and exited. For example, after operating in the diagnostic mode, light emitters 312a, 312b, 312c can be configured to emit light pulses simultaneously, as described above in the main operating mode.

[0082] III. Example of a staining module

[0083] In some embodiments, a flow imaging system incorporates a stain and an associated staining module to enhance visualization of biological material (e.g., blood cells). The stain can be used, for example, to stain the inner cell region of a leukocyte to visualize the inner nuclear structure, thereby facilitating identification of cell types (e.g., subsets of leukocyte types, e.g., identified as neutrophils, lymphocytes, monocytes, eosinophils, or basophils). In some examples, a stain is applied to the outer surface of a cell to enhance visualization of the cell (e.g., an external stain for red blood cells or platelets).

[0084] A. Example of a single chamber staining module

[0085] In such Figure 1 or Figure 2 In the system shown, Figures 6A to 6C The staining module (also referred to as a staining device) 400 shown can be used to mix a sample with a stain and cells in the sample mixture are imaged by a camera (e.g., Figure 1 High optical resolution imaging device 24 or Figure 2 The sample mixture is incubated by heating before imaging with the image sensor 210. For example, the staining module 400 can replace Figure 1 The source 25 shown is combined or combined with Figure 1Between the source 25 and the sample feed tube 29 shown, to facilitate mixing of the sample with the stain and incubation of the sample mixture before capturing an image of the sample by the high optical resolution imaging device 24. The stain can include any suitable composition. For example, the stain can be composed according to any one or more of the teachings of: U.S. Patent No. 9,279,750, entitled "Method and Composition for Staining and Sample Processing," issued on March 8, 2016, the disclosure of which is incorporated herein by reference in its entirety; and / or U.S. Patent No. 9,322,753, entitled "Method and Composition for Staining and Processing a Urine Sample," issued on April 26, 2016, the disclosure of which is incorporated herein by reference in its entirety; and / or U.S. Publication No. 2021 / 0108994, entitled "Method and Composition for Staining and Sample Processing," published on April 15, 2021, the disclosure of which is incorporated herein by reference in its entirety.

[0086] In the illustrated embodiment, the dyeing module 400 includes a housing 410, a pair of ferromagnetic sheets 412, and a heater in the form of a heating coil 414 ( Figure 6C In various embodiments, the heating coil 414 may include a resistance coil, or alternatively an induction coil. Figure 6A As best shown, the housing 410 includes a plurality of (e.g., four) side walls 420 that collectively define an interior chamber 422 (also referred to as a sample reservoir) for receiving a stain and a sample, mixing the stain and sample to form a sample mixture, and culturing the sample mixture. The housing 410 also includes a top wall 424 and a port 426 extending through the top wall 424 to the interior chamber 422. The port 426 can allow a stain dispenser (not shown) to deliver the stain to the interior chamber 422 and / or can allow a sample dispenser (not shown) to deliver the sample to the interior chamber 422 to add to the stain.

[0087] In some embodiments, the housing 410 can include a metallic material having a relatively high thermal conductivity, such as aluminum, to promote uniform heating of the housing 410 and, likewise, the contents of the interior chamber 422. In the illustrated embodiment, the side walls 420 of the housing 410 are laminated with respective ferromagnetic sheets 412 to improve the efficiency of the heating (e.g., resistive heating, or, alternatively, inductive heating) performed by the dyeing module 400 (e.g., due to the relatively low ferromagnetic properties of aluminum). More specifically, each ferromagnetic sheet 412 is secured to the outer surface of a corresponding pair of side walls 420. It will be appreciated that any suitable number of ferromagnetic sheets 412 can be used to laminate the side walls 420. In the illustrated embodiment, a thermally conductive compound 430 is deposited on the outer surface of the side walls 420 for adhering the ferromagnetic sheets 412 to the side walls 420. As shown Figure 6B As best shown, the tape 432 is tightly wrapped around the ferromagnetic sheet 412 to securely engage the inner surface of the ferromagnetic sheet 412 with the outer surface of the sidewall 420 .

[0088] like Figure 6C As best shown, the heating coil 414 includes a wire 440 wound around the sidewall 420 of the housing 410 (and around the ferromagnetic sheet 412). The wire 440 can include a metal material with relatively high electrical conductivity, such as copper. The wire 440 can have any suitable cross-sectional area and / or thickness and can be wound to define any suitable number of turns of the heating coil 414. In one embodiment, the heating coil 414 acts as an inductor or induction coil and is operably coupled to a power unit 450, which can be configured to drive the heating coil 414 to a frequency at which the heating coil 414 behaves as a resonant circuit. At this frequency, the heating coil 414 generates an alternating current when excited, thereby generating an alternating magnetic field at or near the heating coil 414. This field can generate an electromagnetic field (EMF) on the outer surface of the sidewall 420, which in turn can induce the alternating current. This current, together with the resistivity of the housing 410, can generate power dissipation and heat the outer surface of the sidewall 420. In one embodiment, the induction heating element 414 is heated to a temperature of 100 ℃ and / or more and / or more than 100 ℃.Such heat can be transferred to the content of chamber 422, for example staining agent and / or sample.Should be understood that such induction heating can use relatively low input power to carry out and / or can realize the homogeneous heating of the content of chamber 422, and therefore improves staining and / or cracking performance.In this respect, can deliver maximum power with resonant frequency excitation circuit, and can regulate power delivery effectively with the frequency excitation circuit that increases.Alternative embodiment can utilize resistance heater / resistance heating coil as heating coil 414.

[0089] In some embodiments, a temperature sensor such as a thermistor (not shown) can be configured to continuously sense the temperature of the contents of chamber 422. The temperature sensor can be configured to send a feedback signal indicating the sensed temperature to a controller (not shown), which in turn can be configured to send a control signal to the power unit 450 for selectively driving the heating coil 414. In this way, the controller can stop heating the contents of chamber 422 when a threshold temperature is reached. In one example, the controller utilizes a heating control algorithm, and the feedback signal is incorporated into an element of the algorithm or computer-driven instructions provided to the power unit 450 and / or the heating coil 414 to optimally regulate the temperature. In some embodiments, the controller can be configured to send a control signal to a maintenance heater (not shown) for maintaining the contents of chamber 422 at a threshold temperature.

[0090] In one embodiment, multiple dyeing modules are envisioned, each utilizing Figures 6A to 6C The structure (that is, multiple structural elements 400) of the present invention. In this way, multiple samples can be dyed, cultured or otherwise prepared at similar times. In one example, each dyeing module has its own unique heating element. In one example, the dyeing module has multiple chambers 422 (each chamber 422 can receive samples) and a public heating structure (for example, a single housing 410 with multiple chambers 422 and a public heating coil 414 around the housing 410) that is connected to the entire module.

[0091] B. Example of a multi-chamber staining module

[0092] In such Figure 1 or Figure 2 In the system shown, Figure 7A and Figure 7B The illustrated multi-chamber staining module (also referred to as a staining apparatus) 500 can be used to mix a sample with a stain and cells in the sample mixture are imaged by a camera (e.g., Figure 1 High optical resolution imaging device 24 or Figure 2 The sample mixture is incubated via induction heating before imaging with the image sensor 210). For example, the staining module 500 may replace Figure 1 The source 25 shown is combined or combined with Figure 1Between the source 25 and the sample feed tube 29 shown, to facilitate mixing of the sample with the stain and incubation of the sample mixture before capturing an image of the sample by the high optical resolution imaging device 24. The stain can include any suitable composition. For example, the stain can be composed according to any one or more of the teachings of: U.S. Patent No. 9,279,750, entitled "Method and Composition for Staining and Sample Processing," issued on March 8, 2016, the disclosure of which is incorporated herein by reference in its entirety; and / or U.S. Patent No. 9,322,753, entitled "Method and Composition for Staining and Processing a Urine Sample," issued on April 26, 2016, the disclosure of which is incorporated herein by reference in its entirety; and / or U.S. Publication No. 2021 / 0108994, entitled "Method and Composition for Staining and Sample Processing," published on April 15, 2021, the disclosure of which is incorporated herein by reference in its entirety. Note that these stains generally describe a stain containing a lysing agent that lyses red blood cells, a penetrant that permeates white blood cells, a staining element that stains the inner contents of the white blood cells, and a repairing element that repairs the white blood cells so that the stain does not escape.

[0093] In the embodiment shown, the staining module 500 includes a housing 510, a support (also referred to as a sleeve) 512, and a heater in the form of a heating coil 514. The housing 510 includes a plurality of inner chambers 522a, 522b, 522c, 522d (also referred to as sample reservoirs) for receiving a stain and a sample, mixing the stain and the sample to form a sample mixture, and culturing the sample mixture. The housing 510 also includes a top wall 524 and a plurality of ports 526a, 526b, 526c, 526d extending through the top wall 524 to the corresponding inner chambers 522a, 522b, 522c, 522d. Ports 526a, 526b, 526c, 526d can allow a dye dispenser (not shown) to deliver dye to the corresponding inner chambers 522a, 522b, 522c, 522d and / or can allow a sample dispenser (not shown) to deliver a sample to the corresponding inner chambers 522a, 522b, 522c, 522d to be added to the dye. Although four inner chambers 522a, 522b, 522c, 522d and corresponding ports 526a, 526b, 526c, 526d are shown, it should be understood that any suitable number of inner chambers 522a, 522b, 522c, 522d and corresponding ports 526a, 526b, 526c, 526d, such as two, three, or more than four inner chambers 522a, 522b, 522c, 522d and corresponding ports 526a, 526b, 526c, 526d, can be used. In some versions, the first and second inner chambers 522a, 522b can be configured to function as white blood cell (WBC) chambers 522a, 522b, while the third inner chamber 522c can be configured to function as a red blood cell (RBC) chamber 522c.

[0094] In some embodiments, the housing 510 may comprise a metallic material having a relatively high thermal conductivity, such as aluminum, to promote uniform heating of the housing 510 and likewise the contents of the internal chambers 522a, 522b, 522c, 522d.

[0095] In the illustrated embodiment, a bracket or sleeve 512 is positioned about the housing 510. The bracket 512 includes an inner bore 530 that is sized and configured to receive at least a portion of the housing 510. In some embodiments, the inner bore 530 can be sized and configured to slidably receive a portion of the housing 510 such that the portion of the housing 510 can be selectively inserted into and / or removed from the inner bore 530. The bracket 512 also includes an upper lip 532 and a lower lip 534 that define a recessed area 536 therebetween. The recessed area 536 is sized and configured to accommodate at least a portion of the heating coil 514.

[0096] The heating coil 514 includes a wire 540 wound around the bracket 512 on the recessed area 536. The wire 540 can include a metal material with relatively high electrical conductivity, such as copper. The wire 540 can have any suitable cross-sectional area and / or thickness and can be wound to define any suitable number of turns of the heating coil 514. The heating coil 514 is operably coupled to a power unit 550, which can be configured to drive the heating coil 514 to a frequency at which the heating coil 514 behaves as a resonant circuit. At this frequency, the heating coil 514 generates an alternating current under excitation, thereby generating an alternating magnetic field at or near the heating coil 514. In this way, the heating coil 514 can act as an inductor and be configured as an induction coil or induction heating coil. This field can generate an electromagnetic field (EMF) on the outer surface of the housing 510, which in turn can induce an alternating current. This current, together with the resistivity of the housing 510, can generate power dissipation and heat the outer surface of the housing 510. Such heat can be transferred to the content of one or more chambers 522a, 522b, 522c, 522d, such as staining agent and / or sample. It should be understood that such induction heating can be performed using relatively low input power and / or can achieve homogeneous heating of the content of one or more chambers 522a, 522b, 522c, 522d, and therefore improve staining and / or cracking performance. In this regard, the maximum power can be delivered with the resonant frequency excitation circuit, and the power delivery can be effectively regulated with the increased frequency excitation circuit.

[0097] In some embodiments, the bracket or sleeve 512 is composed of a conductive material (e.g., a metal material such as aluminum) to increase heat transfer to the housing and the interior surface that receives the blood sample and stain. In some embodiments, the bracket or sleeve 512 is composed of a ferromagnetic material. In some embodiments, the bracket or sleeve 512 is not used, and the heater / heating coil 514 is directly mounted to the exterior surface of the housing 510.

[0098] In some embodiments, a temperature sensor, such as a thermistor (not shown), can be configured to continuously sense the temperature of the contents of one or more chambers 522a, 522b, 522c, 522d. The temperature sensor can be configured to send a feedback signal indicating the sensed temperature to a controller (not shown), which in turn can be configured to send a control signal to the power unit 550 for selectively driving the heating coil 514. In this way, the controller can stop heating the contents of the chamber 522 when a threshold temperature is reached. In some embodiments, the controller can be configured to send a control signal to a maintenance heater (not shown) for maintaining the contents of the one or more chambers 522a, 522b, 522c, 522d at a threshold temperature.

[0099] C. Example of Sample Preparation Process

[0100] In such Figure 1 or Figure 2 In the system shown, Figure 8 The process shown can be used to visualize the cells being imaged by a camera (e.g. Figure 1 High optical resolution imaging device 24 or Figure 2 Sample preparation is performed before imaging with the image sensor 210). Initially, Figure 8 During the process, at step 601, a stain may be delivered to a chamber, e.g. Figures 6A to 6C Chamber 422 of the staining module 400 is shown or 7A to 7B One or two WBC chambers 522a, 522b of the staining module 500 shown. This can include, for example, delivering the stain to the chambers 422, 522a, 522b via a stain dispenser through corresponding ports 426, 526a, 526b. Then, at step 602, the stain can be preheated within the chambers 422, 522a, 522b, such as via induction heating. Next, at step 603, the sample can be delivered to the chambers 422, 522a, 522b. This can include, for example, delivering the sample to the chambers 422, 522a, 522b via corresponding ports 426, 526a, 526b via a sample dispenser to add to the stain. In some embodiments, delivering the sample to the chambers 422, 522a, 522b can include mixing the sample with the preheated stain within the chambers 422, 522a, 522b. Figure 8 During the process, then, at step 604, a homogeneous sample mixture can be formed in chambers 422, 522a, 522b. This can include, for example, using fluid energy to mix the sample with a dye, such as by cyclically removing the sample from chambers 422, 522a, 522b via the corresponding tangential ports of housing 410, 510 and pushing the sample back into chambers 422, 522a, 522b to perform backflow mixing. Alternatively, this can include using a magnet to drive a spherical ferromagnetic ball placed in chambers 422, 522a, 522b to perform stirring mixing. As another example, this can include introducing one or more bubbles to generate a vortex at the bottom of chambers 422, 522a, 522b.

[0101] Then, at step 605, the homogenized sample mixture can be heated in chambers 422, 522a, 522b, for example, via induction heating or resistance heating. In some embodiments, the homogenized sample mixture can be heated to a threshold temperature via induction heating or resistance heating and then maintained at the threshold temperature via a maintenance heater. In embodiments using a multi-chamber staining module 500, it will be understood that the sample mixture in both WBC chambers 522a, 522b can be heated simultaneously via induction heating, and any sample in RBC chamber 522c can also be heated via induction heating (even if such heating of the sample in RBC chamber 522c may not be required prior to imaging), while the fourth chamber 522d can remain empty. In some other embodiments using a multi-chamber staining module 500, one or more chambers 522a, 522b, 522c, 522d, such as the first WBC chamber 522a, can be used to heat the sample mixture while one or more of the other chambers 522a, 522b, 522c, 522d, such as the RBC chamber 522c, is simultaneously flushed; in such a case, increased power can be provided to the heating coil 514 to offset any cooling effect that the flushing of the RBC chamber 522c might otherwise have on the heating of the sample mixture within the first WBC chamber 522a (e.g., by increasing the same amount of energy lost by such cooling effect).

[0102] After the homogenized sample mixture reaches a threshold temperature, the sample mixture can be delivered to a flow cell such as Figure 1 The flow tank 22 is used for the camera e.g. Figure 1 The sample mixture can be imaged by a high optical resolution imaging device 24. For example, the homogenized sample mixture can be delivered directly from chambers 422, 522a, 522b to the flow cell 22 (e.g., without undergoing other preparation). It will be appreciated that heating the sample mixture in the same chamber 422, 522a, 522b as the chamber in which the sample mixture is formed can improve the throughput of the staining process.

[0103] While the formation and inductive heating of the sample mixture have been described as occurring within the chambers 422, 522 of the housings 410, 510, it will be appreciated that an alternative arrangement may include a tube having a lumen (not shown) in which the sample mixture may be formed and inductively heated in a manner similar to that described above. Additionally, or alternatively, any one or more of the teachings herein may be combined with any one or more of the teachings disclosed in U.S. Patent No. 9,429,524, issued August 30, 2016, entitled “Systems and Methods for Imaging Fluid Samples,” the disclosure of which is incorporated herein by reference in its entirety.

[0104] In some embodiments, the addition of diluent is part of the preparation step, wherein the diluent is added to each chamber 522a to 522d during, before, after, or both before and after the addition of the blood sample to each chamber. For example, the RBC chamber may receive diluent as the primary or sole preparation reagent, while the WBC chamber may receive both diluent and stain.

[0105] It should be understood that the preparation steps for the RBC chamber can be different from those for the WBC chamber. For example, the RBC chamber may utilize a preparation step that involves: a) receiving a diluent, then a blood sample; b) receiving a blood sample, then a diluent; or c) receiving a diluent, then a blood sample, then additional diluent; but not receiving a stain. In this way, the preparation time for the RBC chamber can be shortened, and the workflow can involve running the RBC sample through the imaging process while the WBC sample is still being prepared.

[0106] In some embodiments, the staining reagent utilizes both a lysing agent (to lyse red blood cells) and a staining agent (to penetrate the remaining white blood cells, stain the interior area, and fix the white blood cells so that the staining agent does not escape). In this way, a single staining reagent can be used to treat certain types of cells (e.g., white blood cells) to eliminate red blood cells and stain the remaining white blood cells. Other embodiments can utilize multiple compositions, such as a first lysing reagent to lyse red blood cells and a second staining reagent to stain white blood cells, wherein the workflow will involve a chamber (e.g., a WBC chamber) receiving a separate lysing reagent and a separate staining reagent to prepare a WBC sample for visualization.

[0107] In some embodiments, various chambers (e.g., 522a to 522d) are not intended to strictly prepare dedicated cell types, or in other words, cell types can be rotated. For example, these chambers can be alternately used for RBC preparation and WBC preparation. In this way, once the sample in the chamber is prepared for imaging, a cleaning cycle can be utilized to clean the chamber before receiving subsequent blood samples (e.g., chamber 522a can first be configured to prepare the WBC for a certain amount of identical preparation operation, then prepare the RBC for a certain amount of sample preparation operation, e.g., 1 WBC preparation, then 1 RBC preparation, or 2 WBC preparations, then 1 RBC preparation, then 2 more WBC preparations, etc.). A cleaning agent such as a diluent or detergent can be used between sample operations to eliminate carryover. Even when a specific chamber is only used for a specific cell type (e.g., 522a is only used as the WBC chamber), a cleaning step can be run after sample preparation and imaging to eliminate carryover.

[0108] Other embodiments may still utilize multiple stains as part of the preparation process. For example, a first stain is configured to stain white blood cells in the manner described herein, and a second stain is configured to stain at least one of platelets or reticulocytes. These staining compositions can be uniquely used in various workflows. For example, a first chamber of housing 410, 510 can be used to prepare a white blood cell sample, including receiving at least a WBC stain and a lysis reagent, while a second chamber of housing 410, 510 can be used to prepare a platelet sample—this chamber will receive at least a platelet reagent that is different from the WBC stain and lysis reagent.

[0109] Note that although the terms white blood cell (WBC) chamber and red blood cell (RBC) chamber are used to refer to the sample preparation chambers for imaging, the samples imaged as a result of the preparation process can allow for biological imaging of multiple cell types. For example, the WBC chamber utilizes a lysing agent to eliminate red blood cells, however, the lysing agent may still retain platelets and reticulocytes, and thus, a sample prepared in the WBC chamber can still image at least, for example, white blood cells, platelets, and reticulocytes. Similarly, the RBC chamber can receive a different preparation procedure than the WBC chamber (e.g., no lysis or no staining / lysis combination reagents), but a sample prepared in the RBC chamber can still visualize multiple cell types, such as red blood cells, as well as one or more of white blood cells, platelets, and reticulocytes.

[0110] Additional information regarding the illumination module and the dye module can be found in U.S. Patent Application Serial Nos. 18 / 224,937, 18 / 224,947, and 18 / 224,953, the contents of which are incorporated herein by reference in their entirety.

[0111] IV. Example of Maturity Determination

[0112] Figure 1 or Figure 2The system generally shows an imaging system for obtaining images of particles (e.g., blood cells). For certain cells, such as reticulocytes and platelets, these cells start out as immature versions with relatively high RNA content. Being able to quantify cell maturity, such as quantifying immature platelets or immature reticulocytes, can provide detailed information about patient health and therefore can be valuable. In some embodiments herein, image analysis techniques can be used to analyze images of cell types, for example, analyzing the nuclear region of a cell to correlate with RNA content assessment, and classifying or quantifying cell maturity (e.g., identifying a cell as an immature platelet or immature reticulocyte, and identifying the number of such cells in a blood sample). In various embodiments, these techniques can be used to generally identify and / or enumerate immature cells in a blood sample (e.g., not just immature platelets or immature reticulocytes).

[0113] In such Figure 1 or Figure 2 In the system shown, the maturity of particles in a sample can be determined using a process of acquiring an image of the particles and then using the image to determine their maturity. Figure 9 An illustration of such a process is provided in , which not only shows acquiring 901 a particle image and determining 902 its maturity, but also shows certain actions that may be performed in the image acquisition. Figure 9 As shown, this may optionally include one or more sample preparation steps. These may include, for example, treating 903 the sample with a lysing agent that will remove one or more types of particles that are not of interest. For example, if the sample is a blood sample, and Figure 9 If a process is used to determine the maturity of reticulocytes or platelets in a sample, the sample can be treated 903 with a lysing agent (e.g., saponin) that will destroy mature red blood cells in the sample. Alternatively (or in addition), sample preparation can include depositing 904 a staining composition (e.g., new methylene blue (NMB), which will stain RNA within cells in the sample and thus can distinguish between immature platelets and reticulocytes, which have relatively more RNA than mature platelets or reticulocytes), and depositing 905 the sample into a chamber where they will be mixed 906. In embodiments where these types of preparation steps are performed, after the sample is prepared, an image of the stained blood cells can be captured 907. This image capture 907 can occur in the context of additional activities, such as flowing 908 the sample through a flow cell (e.g., when using a flow cell such as Figure 1 flow imaging system as shown).

[0114] It should be noted that although Figure 9 Depicts sample preparation steps that may be performed in some embodiments, but Figure 9These steps and their depiction are intended to be illustrative only and are not intended to imply limitations on how sample preparation may be performed (in embodiments where sample preparation is performed at all). Figure 9 Lysis and staining are shown as essentially separate processes, but in some cases, a single composition can include both a lysis component that lyses mature red blood cells and a staining composition that stains the remaining cells. For example, step 904 can include depositing a composition containing both a lysis compound (e.g., saponin) and a staining compound (e.g., new methylene blue) as a single step. In some examples, multiple reagents or compositions can be used (e.g., a lysis composition and a separate staining composition added after the lysis composition). It is also possible that Figure 9 The various steps shown can be performed in a different order, for example, by depositing 905 the biological sample simultaneously with, before, or after depositing 904 the staining composition. Similarly, in some embodiments, preparation can involve only staining the blood sample without lysing the blood sample, i.e., leaving mature red blood cells intact. Descriptions of staining and staining / lysing compounds can be found in U.S. Patent No. 9,279,750 and U.S. Publication No. 2021 / 0108994, the contents of which are incorporated herein by reference in their entirety.

[0115] However, image acquisition 901 is performed (e.g., via Figure 1 Flow imaging as shown or as Figure 2 ), once an image depicting blood cells has been captured, the image can be used to determine 902 the maturity of the cells. This can include, for example, classifying 1000 the depicted cells, such as using a method for performing classification based on morphological features. When performing this type of classification, cells whose maturity is not of interest can be identified, and a maturity determination for these cells may not be completed (although data about these cells may be retained for other purposes). For example, where a blood sample is expected to include white blood cells, platelets, and reticulocytes (e.g., because mature red blood cells have been removed by treatment with a lysing agent), and where a maturity determination is to be performed in order to calculate an immature platelet and reticulocyte fraction, the maturity determination may be omitted for cells that are classified as white blood cells, while the remaining cells may have their maturity determined based on their classification as reticulocytes or platelets. An illustration of how this type of classification 1000 may be performed based on morphological features is provided below. Figure 10 .

[0116] Initially, in Figure 10In the method, an intensity image 1001 is generated based on an image captured by the system. Specifically, the intensity image includes pixels of the image captured by the system and associates each of these pixels with a corresponding intensity associated therewith. For example, if the system is configured to capture RGB images, the intensity image 1001 can be generated by associating each pixel of the image captured by the system with the average value of the R, G, and B color channels of each pixel.

[0117] After step 1001, foreground and background portions can be identified 1002 in these intensity images. This can be achieved, for example, by thresholding, where darker pixels representing blood cells are separated from lighter pixels representing the background—for example, by thresholding at a specific intensity value, where pixels exceeding the intensity value are classified as foreground, while pixels below the intensity value are classified as background, or vice versa, depending on the encoding of the intensity value. For example, in some embodiments, each channel of each pixel of the image captured by the system has a value between 0 and 255. In this case, for example, pixels associated with an intensity value greater than or equal to approximately 204 are considered to belong to the foreground, while pixels associated with an intensity value less than approximately 204 are considered to belong to the background. These intensity values ​​can depend on the conditions of the imaging system (by way of example, these can include: lighting conditions, staining quality, camera imaging characteristics, cell velocity at the time the cell image was captured). In some examples, depending on these system qualities, the pixel intensity threshold can have various values, such as numbers in the range of approximately 170 to 230, 190 to 210, or 200 to 207.

[0118] With the foreground and background pixels identified, the intensity of the foreground pixels can be calculated 1003. This can be accomplished, for example, by calculating 1005 the average red intensity of the foreground pixels and calculating 1006 the average blue intensity of the foreground pixels from the original RGB image. Other methods of calculating 1003 the intensity of the foreground pixels are also possible and can be used instead of or in addition to the RGB-based calculations. For example, where an image is encoded using a hue, saturation, value color model (i.e., an HSV image), calculating 1003 the intensity may simply require treating the V value of the pixel from the HSV image as the intensity. Therefore, the description of calculating 1003 the intensity should be understood as merely illustrative and should not be considered limiting.

[0119] Once calculated 1003, the intensity can be used to classify the type of imaged cells 1004. For example, if the average intensity of the red channel of the foreground pixels of an image (e.g., denoted as MASK_R_MEAN) is less than a specific white blood cell (WBC) red channel intensity cutoff, and the average intensity of the blue channel of its foreground pixels (e.g., denoted as MASK_B_MEAN) is less than a specific white blood cell (WBC) blue channel intensity cutoff, then the image can be considered to depict white blood cells (e.g., denoted as WBC). Similar approaches can be taken when identifying other types of cells, such as platelets and reticulocytes, although more complex classifications are also possible. For example, in some cases, rather than simply comparing the foreground pixel red channel mean intensity (e.g., represented as MASK_R_MEAN) and the foreground pixel blue channel mean intensity (e.g., represented as MASK_B_MEAN) to a cutoff value, the classification may include one or more of: (i) comparing the foreground pixel red channel mean intensity (e.g., represented as MASK_R_MEAN) to the foreground pixel blue channel mean intensity (e.g., represented as MASK_B_MEAN), (ii) comparing a function of the foreground pixel red channel mean intensity (e.g., represented as MASK_R_MEAN) to a function of the foreground pixel blue channel mean intensity (e.g., represented as MASK_B_MEAN), and / or (iii) comparing a function of the foreground pixel red channel mean intensity (e.g., represented as MASK_R_MEAN) and the foreground pixel blue channel mean intensity (e.g., represented as MASK_B_MEAN) to a cutoff value.

[0120] To illustrate Figure 10 The classification 1004 in the method of FIG. 1 is set forth in Table 1 below for WBC, platelet, and reticulocyte classification criteria for images captured using a blood cell imaging analysis device (in one example, a blood cell imaging analysis device utilizing flow imaging, wherein cells are imaged as they continue past a camera in a flow cell and are thresholded using intensity values ​​for separating foreground pixels from background pixels (e.g., by a threshold of approximately 204)). For purposes of the following table, MASK_B_MEAN refers to the foreground pixel blue channel mean intensity, and MASK_R_MEAN refers to the foreground pixel red channel mean intensity.

[0121]

[0122] Table 1: Example classification criteria

[0123] It should be understood that the classification criteria of Table 1 (and the thresholds used to separate foreground from background 1003) are intended to be exemplary only. Figure 10 Different methods can be used to implement this classification than those described in the context of Figure 10 In addition to the morphology-based classification described in the context of , some embodiments may use a machine learning model to classify cells 1000. Examples of such machine learning models are Figure 11 When using Figure 11 When using the model 1100, the input image 1101 will be analyzed in a series of stages 1102a to 1102n, each of which can be called a "layer" and Figure 12 As shown in more detail in Figure 12 As shown, input 1201 (which is in Figure 11 The initial layer 1102a (which would be the input image 1101 otherwise, or the output of the previous layer) is provided to layer 1202, where the input 1201 is processed to generate one or more transformed images 1203a through 1203n. This processing may include convolving the input 1201 with a set of filters 1204a through 1204n, each of which identifies a type of feature from the underlying image that will then be captured in the corresponding transformed image for that filter. For example, as a simple example, convolving the image with the filters shown in Table 2 can generate a transformed image that captures edges from the input image 1201.

[0124]

[0125] Table 2

[0126] like Figure 12 As shown, in addition to generating the transformed images 1203a to 1203n, the layer can also generate a pooled image 1205a to 1205n for each of the transformed images 1203a to 1203n. This can be achieved, for example, by organizing the appropriate transformed images into a set of regions and then replacing the values ​​in each of the regions with a single value (e.g., the maximum value for that region or the average of the values ​​for that region). The result will be a pooled image whose resolution will be reduced relative to its corresponding transformed image based on the size of the regions into which it is partitioned (e.g., if the transformed image has NxN dimensions and it is partitioned into 2x2 regions, the size of the pooled image will be (N / 2)x(N / 2)). These pooled images 1205a to 1205n can then be combined into a single output image 1206, where each of the pooled images 1205a to 1205n is treated as a separate channel in the output image 1206. This output image 1206 can then be provided as input to the next layer, as shown in FIG. Figure 11 shown.

[0127] Return to Figure 11As discussed above, after a final output image 1103 has been created through the various stages of processing 1102a to 1102n, the final output image 1103 can be provided as input to a neural network 1104. This can be achieved, for example, by providing the value of each channel of each pixel in the output image 1103 to the input nodes of a densely connected single layer network. The output of the neural network 1104 can then be considered a classification of the original input image 1101. For example, in a network such as Figure 11 In the case shown, where the neural network 1104 has multiple output nodes, each of these output nodes can be considered to correspond to a cell class (e.g., one output node corresponding to WBC, one output node corresponding to reticulocyte, and one output node corresponding to platelet), and the corresponding class of the output node having the highest value can be considered to be the class of the cell depicted in the input image that resulted in that value being achieved.

[0128] Can train such as Figure 11 and Figure 12 The machine learning model shown is used to classify cells using blood cell images with known categories, thereby minimizing the cross entropy loss among the output nodes of the neural network 1104. Such blood cell images can be obtained by manually annotating images generated during normal operation of the analyzer (e.g., manually inspecting the images and then labeling them with cell categories), but they can also be obtained in other ways. For example, morphology-based classification can be used to classify images, such as the one described above in Figure 10 These classified images can then be used to train machine learning models, e.g. Figure 11 and Figure 12 as shown. The training can include splitting the classified images into multiple subsets or folds, and then training and evaluating the model multiple times, each time retaining a different fold of the training images as a validation set (i.e., K-fold cross validation). In this way, performance metrics from each training instance can be averaged to verify the generalization performance of the model, and assuming that the performance is acceptable, the final trained version of the model (e.g., whichever trained model has the best individual performance) can be used for inference in production (i.e., classifying cell images). Additional types of machine learning models, such as models with a single output, are described in Patent Cooperation Treaty application PCT / US22 / 52702 filed on December 13, 2022, the disclosure of which is incorporated herein by reference in its entirety, and the additional types of machine learning models can be used instead of Figure 11(e.g., a single output machine learning model can be trained to provide different output values ​​for different types of cells (e.g., 0 for WBCs, 1 for reticulocytes, 2 for platelets), and whichever of these values ​​is closest to the output value generated by the model based on a particular input image will be considered the classification of the cell depicted in that input image). Similarly, other types of cell classifications can also be used, such as those described in U.S. Patent No. 11,403,751, which is incorporated herein in its entirety. Thus, the above Figure 11 and Figure 12 The discussion of using machine learning models for classification set forth in the context of should be understood as illustrative only and should not be regarded as limiting.

[0129] In embodiments that include classifying 1000 the imaged cells, after the classification is complete, the imaged cells may be sorted, for example, using Figure 13 The method shown generates maturity. Figure 13 As shown, generating 1300 a maturity metric for an imaged cell may include utilizing 1301 a count of pixels included in a foreground portion of an image. For example, in the case of generating 1300 a maturity metric to calculate an immature platelet fraction and an immature reticulocyte fraction, utilizing 1301 a foreground pixel count may include determining 1302 whether a threshold pixel count is greater than a threshold value. If the number of pixels in the foreground of the image is above the pixel threshold, the depicted cell may be classified as immature, while if the pixel count is below the threshold, it may be classified as mature. In embodiments employing this method, the pixel threshold used in this determination 1302 may depend on how the cells for which the maturity metric is being generated are classified. For example, in an image where each pixel corresponds to a square with a side length of 0.14 μm, a threshold of 490 pixels may be used to separate immature platelets from mature platelets, where an image where platelets cover more than 490 pixels is considered immature, while an image where platelets cover 490 or fewer pixels is considered mature. Alternatively or additionally, in images captured using the same device, a threshold of 700 pixels can be used to separate immature reticulocytes from mature reticulocytes, where reticulocytes covering more than 700 pixels in the image are considered immature, while reticulocytes covering 700 pixels or less in the image are considered mature. Different thresholds can be used in images captured using devices with different pixel sizes, or for other types of cells of interest, where the particular threshold used in a particular context is identified using methods known in the art, such as using a clustering algorithm to determine a best-fit line that separates labeled images of immature and mature cells of a particular type.

[0130] Other methods of utilizing 1301 the count of foreground pixels in generating 1300 the maturity of the imaged cells are also possible. Figure 13 As shown, the count of foreground pixels 1301 can be utilized when calculating the average parameter value for the cells 1303. This type of approach can take advantage of the fact that the maturity of a cell can be correlated with a particular measurable parameter. For example, because more immature reticulocytes or platelets tend to have more RNA than more mature ones, or in other words, as these cells mature, the RNA content decreases, the staining process can result in measurable color differences associated with (im)maturity (e.g., more immature cells may tend to have bluer pixels than less mature cells). Therefore, the values ​​of color channels that differ measurably depending on immaturity can be averaged across the foreground pixels, thereby providing the maturity of the cell based on the average value of that parameter (e.g., the average "blueness" value). Other variations are possible and can be implemented by those skilled in the art based on the present disclosure without undue experimentation. For example, rather than making a binary distinction between "mature" and "immature" or treating maturity as a continuous value like a parameter average, cells can be placed into one of a set of maturity bins based on where the average parameter value (e.g., blueness) falls on a continuum of possible values. Similarly, in some cases, other methods may be used, such as calculating maturity based on the number of foreground pixels in the cell nucleus versus the number of foreground pixels in the cell in question, based on the perimeter of the cell nucleus versus the area of ​​the cell in question, or based on a combination of two or more of the previously listed (or other) factors. Accordingly, the discussion of binary mature / immature classification using 1301 foreground pixel counts should be understood as merely illustrative and should not be taken as implying a limitation on the protection afforded by this document or any other document claiming the benefit of this document.

[0131] Regardless of how it is generated, once it is available, the maturity data for the cells depicted in the image being analyzed can be stored 1304, for example, for combination with maturity data for other cells to provide an overall picture of the biological sample. Figure 14 , which illustrates a method in which maturity information, such as can be generated using the disclosed techniques, is used to generate overall information about a biological sample. Figure 14 When using the method shown, a plurality of additional images are obtained 1401, for example by capturing images of additional cells included in the biological sample being analyzed. Once these additional images are obtained, the maturity of a cell depicted in one of these images can be determined 1402 (e.g., using the methods described above as possible for use in Figure 9The process may then continue 1403 to the next image as long as there are pictures of cells whose maturity has not yet been determined. Finally, once the maturity of all depicted cells of interest (e.g., platelets and reticulocytes) has been determined 1402, an analysis output may be generated 1404 based on the maturity data for these cells. Figure 14 As shown, generating 1404 the analysis output can be performed in a variety of ways. For example, where the disclosed technology is used to separate reticulocytes and / or platelets into "mature" and "immature" classifications, then generating 1404 the analysis output can include determining 1405 at least one of an immature reticulocyte fraction (i.e., the number of reticulocytes classified as "immature" divided by the total number of reticulocytes, abbreviated as IRF) and an immature platelet fraction (i.e., the number of platelets classified as "immature" divided by the total number of platelets, abbreviated as IPF). Alternatively, where the disclosed technology is used to generate maturity values ​​along a continuum (e.g., using a method such as previously described in Figure 13 Generating 1404 analysis output may include generating 1406 a graph showing the distribution of reticulocyte maturity values ​​within a range of reticulocyte maturity values ​​and / or generating a graph showing the distribution of platelet maturity values ​​within a range of platelet maturity values. Other types of analysis output may also be generated 1404 (e.g., a histogram showing the distribution of reticulocytes and / or platelets in the case of grouping imaged cells based on maturity), and thus Figure 14 The examples shown should not be construed as imposing limitations on the types of output that may be generated by implementations of the disclosed technology.

[0132] In addition to the types of outputs that can be generated based on maturity data, other aspects may also vary. For example, while in some embodiments, determining 902 cell maturity may include classifying 1000 cells and then generating 1300 cell maturity, in other embodiments, determining 902 cell maturity may be performed without separate classification 1000 and generation 1300 steps. As an illustration of this, Figure 15 A method is depicted in which a cell maturity level is created as part of classifying a cell. Initially, in the method, a cell is presented 1501 to a machine learning model, for example by providing an image of the cell as a Figure 11 and Figure 12 The machine learning model can then use the convolutional neural network 1502 to classify cells. The machine learning model can be used in the same way as above. Figure 11 and Figure 12The classification is generated in a manner similar to that discussed in the context of , but if the cell being classified is of a type whose maturity is of interest, then the maturity level may also be provided 1503. For example, where the maturity of reticulocytes and platelets is determined, then when a cell is classified 1504 as a reticulocyte class and when a cell is classified 1505 as a platelet class, the machine learning model may also provide the maturity level of the cell 1503. This may be accomplished by constructing a dense layer 1104 to include both a set of type nodes (e.g., a node corresponding to a WBC, a node corresponding to a reticulocyte, and a node corresponding to a platelet) and a maturity node, which is trained to provide the maturity level and can be polled if the type node indicates that the cell is a reticulocyte or a platelet. Alternatively, where maturity would be a binary determination of mature or immature, the machine learning model could include not only a single platelet class and a single reticulocyte class, but also an immature reticulocyte class, a mature reticulocyte class, an immature platelet class, and a mature platelet class, and which type of platelet / reticulocyte class a cell is classified into could be considered to provide the cell's maturity.

[0133] However, in Figure 15 In the method, 1503 maturity is provided, after the maturity is provided 1503 by the machine learning model, the maturity from the model can be regarded as the maturity of the cell being analyzed 1506. For example, in following Figure 11 Where a separate maturity node is present in the dense layer 1104 of the machine learning model of the architecture, the value at that output node can be simply considered 1506 as the maturity of the cell whose presentation 1501 to the machine learning model resulted in that value being generated. Alternatively, where there are output nodes corresponding to immature reticulocytes, mature reticulocytes, immature platelets, and mature platelets, the cell can be considered 1506 immature if it is classified into the immature platelet or immature reticulocyte class, or can be considered 1506 mature if it is classified into the mature reticulocyte or mature platelet class. Thus, the discussion of an embodiment in which a cell maturity is generated 1300 after the cell is classified 1000 should be understood as merely illustrative and should not be considered limiting.

[0134] Other variations in the potential implementation of the disclosed technology are also possible and will be immediately apparent to those skilled in the art in light of this disclosure. Figure 11 and Figure 12The machine learning model shown may simply take as input an image depicting a cell, but it is also possible that other information may be provided as input to such a model. For example, in addition to or as an alternative to the cell image, derived information may be provided to the model, such as the number of foreground pixels, the ratio of the number of pixels in the nucleus to the number of pixels in the cell, and / or other parameters described herein as being used to determine cell maturity (whether as part of cell classification or otherwise). As another example, while a machine learning model (or other classification function, such as in Figure 10 The morphology-based functions described in the context of can be designed to classify cells in samples that have previously been lysed to remove mature red blood cells (RBCs), but it is also possible that the disclosed techniques can be implemented to classify samples where RBCs may be present (e.g., by adding an additional output node for RBCs to the Figure 11 1104 of the machine learning model shown). Variations are also possible with respect to the physical devices that may be used for different implementations. For example, in some cases, analysis of images associated with determination of the maturity of depicted cells may use methods for, for example, Figure 1 and Figure 2 The system shown is for capturing the images themselves, performed by a local processor. However, in other cases, the analysis and collection of data may be separate, such as in a system where the captured images are transmitted via a network connection to a remote processor, which analyzes them and then processes the results (e.g., as in Figure 14 The analysis outputs generated 1404 as described in the context of ) are sent back to their source (or another endpoint) for viewing.

[0135] Other applications are also possible. For example, although the above description focuses on determining maturity, various aspects of the disclosed technology can also be used for other types of analysis, such as detecting cells infected with malaria or other parasites. This is because infected cells such as immature platelets and reticulocytes will have relatively more RNA and can therefore be distinguished from other cells using techniques similar to those described above for identifying reticulocyte and / or platelet maturity (e.g., using thresholding based on the number of dark pixels in a cell image). As further illustration of potential implementations and applications of the disclosed technology, the following examples are provided that are non-exhaustive examples of ways in which the teachings herein can be combined or applied. It should be understood that the following examples are not intended to limit the scope of any claims that may be filed at any time in this application or subsequent filings of this application. No disclaimer is required. The following examples are provided for illustrative purposes only. It is contemplated that the various teachings herein can be arranged and applied in many other ways. It is also contemplated that some variations may omit certain features mentioned in the following examples. Therefore, any aspect or feature mentioned below should not be considered critical unless otherwise explicitly indicated by the inventor or the inventor's successor in interest at a later date. If any claims are presented in this application or in a subsequent filing related to this application that include additional features beyond those mentioned below, those additional features should not be presumed to have been added for any reason related to patentability.

[0136] Example 1

[0137] A computer-implemented image analysis method for detecting the maturity of blood cells comprises: depositing a staining composition into a chamber; depositing a biological sample into the chamber; mixing the biological sample and the staining composition in the chamber; capturing an image of the stained blood cells from the biological sample with a camera; and determining the maturity of the stained blood cells based on the image of the stained blood cells captured with the camera.

[0138] Example 2

[0139] An image analysis method according to Example 1, wherein: the method further comprises: treating the biological sample with a lysing agent; and allowing the biological sample to flow through a flow tank and past the camera; and while the biological sample is flowing through the flow tank and past the camera, capturing the image of the stained blood cells with the camera.

[0140] Example 3

[0141] The image analysis method according to Example 2, wherein: determining the maturity of the stained blood cells includes classifying the stained blood cells; and classifying the stained blood cells includes classifying the stained blood cells into at least one of reticulocytes or platelets.

[0142] Example 4

[0143] An image analysis method according to Example 3, wherein the method further comprises: obtaining a plurality of additional images, wherein each image from the plurality of additional images depicts additional stained blood cells corresponding to the image; for each image from the plurality of additional images, determining the maturity of the additional stained blood cells corresponding to the image; and generating an analysis output based on a set of maturity data, wherein the set of maturity data includes the maturity of the stained blood cells and the maturity of the stained blood cells depicted in the plurality of additional images.

[0144] Example 5

[0145] The image analysis method of Example 4, wherein generating the analysis output comprises determining at least one of an immature reticulocyte fraction and an immature platelet fraction of the biological sample.

[0146] Example 6

[0147] An image analysis method according to any one of Examples 4 to 5, wherein generating the analysis output includes generating at least one of: a curve graph showing the distribution of reticulocytes in the biological sample within the reticulocyte maturity range; and a curve graph showing the distribution of platelets in the biological sample within the platelet maturity range.

[0148] Example 7

[0149] The image analysis method of any one of Examples 3 to 6, wherein classifying the stained blood cells comprises utilizing a convolutional neural network.

[0150] Example 8

[0151] An image analysis method according to Example 7, wherein: the convolutional neural network is included by a machine learning model, and the machine learning model is configured to, when presented with input stained blood cells: classify the input stained blood cells as reticulocytes by classifying the input stained blood cells into a reticulocyte category; provide the maturity of the input stained blood cells when the input stained blood cells are classified into the reticulocyte category; classify the input stained blood cells as platelets by classifying the input stained blood cells into a platelet category; and provide the maturity of the input stained blood cells when the input stained blood cells are classified into the platelet category; and determining the maturity of the stained blood cells includes: presenting the stained blood cells as the input stained blood cells to the machine learning model; and regarding the maturity provided by the machine learning model as the maturity of the stained blood cells.

[0152] Example 9

[0153] An image analysis method according to Example 8, wherein: the machine learning model is configured to: classify the input stained blood cells into the platelet category, and when the input stained blood cells are classified into the platelet category, provide the maturity of the input stained blood cells by classifying the input stained blood cells into a category selected from the group including the following categories: an immature platelet category; and a mature platelet category; and the machine learning model is configured to: classify the input stained blood cells into the reticulocyte category, and when the input stained blood cells are classified into the reticulocyte category, provide the maturity of the input stained blood cells by classifying the input stained blood cells into a category selected from the group including the following categories: an immature reticulocyte category; and a mature reticulocyte category.

[0154] Example 10

[0155] The image analysis method according to any one of Examples 3 to 7, wherein determining the maturity of the stained blood cells includes generating the maturity of the stained blood cells after the stained blood cells are classified.

[0156] Example 11

[0157] The image analysis method of any one of Examples 3 to 7, wherein classifying the stained blood cells comprises identifying a foreground in the image based on the intensity of each pixel in the image.

[0158] Example 12

[0159] The image analysis method according to Example 11 includes: calculating: the average red intensity of pixels in the foreground of the image; and the average blue intensity of pixels in the foreground of the image; and classifying the stained blood cells based on the average red intensity and the average blue intensity.

[0160] Example 13

[0161] The image analysis method according to any one of Examples 1 to 12, wherein determining the maturity of the stained blood cells includes utilizing pixel foreground counting.

[0162] Example 14

[0163] The image analysis method of any one of Examples 1 to 7 or 10 to 12, wherein determining the maturity of the stained blood cells comprises determining whether a pixel foreground count exceeds a threshold amount.

[0164] Example 15

[0165] The image analysis method according to any one of Examples 1 to 14, wherein the method includes: preheating the staining composition in the chamber before depositing the biological sample in the chamber; and heating the biological sample and the staining composition in the chamber.

[0166] Example 16

[0167] The image analysis method according to Example 15, wherein both preheating the staining composition in the chamber and heating the biological sample and the staining composition in the chamber include: increasing the chamber temperature using a heating coil.

[0168] Example 17

[0169] The image analysis method of Example 16, wherein the heating coil is mounted on a mounting bracket connected to the chamber and one or more other mixing chambers.

[0170] Example 18

[0171] A computer-implemented image analysis method for detecting the maturity of blood cells, comprising: flowing a biological sample through a flow cell and past a camera; capturing images of stained blood cells with the camera while the biological sample is flowing through the flow cell and past the camera; and determining the maturity of the stained blood cells based on the images of the stained blood cells captured with the camera.

[0172] Example 19

[0173] The image analysis method according to Example 18 further comprises: treating the biological sample with a lysis agent.

[0174] Example 20

[0175] The image analysis method according to Example 19, wherein: determining the maturity of the stained blood cells includes classifying the stained blood cells; and classifying the stained blood cells includes classifying the stained blood cells into at least one of reticulocytes or platelets.

[0176] Example 21

[0177] An image analysis method according to Example 20, wherein the method further comprises: obtaining a plurality of additional images, wherein each image from the plurality of additional images depicts additional stained blood cells corresponding to the image; for each image from the plurality of additional images, determining the maturity of the additional stained blood cells corresponding to the image; and generating an analysis output based on a set of maturity data, wherein the set of maturity data includes the maturity of the stained blood cells and the maturity of the stained blood cells depicted in the plurality of additional images.

[0178] Example 22

[0179] The image analysis method of Example 21, wherein generating the analysis output comprises determining at least one of an immature reticulocyte fraction and an immature platelet fraction of the biological sample.

[0180] Example 23

[0181] An image analysis method according to any one of Examples 21 to 22, wherein generating the analysis output includes generating at least one of: a curve graph showing the distribution of reticulocytes in the biological sample within the reticulocyte maturity range; and a curve graph showing the distribution of platelets in the biological sample within the platelet maturity range.

[0182] Example 24

[0183] The image analysis method of any one of Examples 20 to 23, wherein classifying the stained blood cells comprises utilizing a convolutional neural network.

[0184] Example 25

[0185] An image analysis method according to Example 24, wherein: the convolutional neural network is included by a machine learning model, and the machine learning model is configured to, when presented with input stained blood cells: classify the input stained blood cells as reticulocytes by classifying the input stained blood cells into a reticulocyte category; provide the maturity of the input stained blood cells when the input stained blood cells are classified into the reticulocyte category; classify the input stained blood cells as platelets by classifying the input stained blood cells into a platelet category; and provide the maturity of the input stained blood cells when the input stained blood cells are classified into the platelet category; and determining the maturity of the stained blood cells includes: presenting the stained blood cells as the input stained blood cells to the machine learning model; and treating the maturity provided by the machine learning model as the maturity of the stained blood cells.

[0186] Example 26

[0187] An image analysis method according to Example 25, wherein: the machine learning model is configured to: classify the input stained blood cells into the platelet category, and when the input stained blood cells are classified into the platelet category, provide the maturity of the input stained blood cells by classifying the input stained blood cells into a category selected from the group including the following categories: an immature platelet category; and a mature platelet category; and the machine learning model is configured to: classify the input stained blood cells into the reticulocyte category, and when the input stained blood cells are classified into the reticulocyte category, provide the maturity of the input stained blood cells by classifying the input stained blood cells into a category selected from the group including the following categories: an immature reticulocyte category; and a mature reticulocyte category.

[0188] Example 27

[0189] The image analysis method according to any one of Examples 20 to 24, wherein determining the maturity of the stained blood cells includes generating the maturity of the stained blood cells after the stained blood cells are classified.

[0190] Example 28

[0191] The image analysis method of any one of Examples 20 to 23, wherein classifying the stained blood cells comprises identifying a foreground in the image based on the intensity of each pixel in the image.

[0192] Example 29

[0193] The image analysis method according to Example 28 also includes: calculating: the average red intensity of pixels in the foreground of the image; and the average blue intensity of pixels in the foreground of the image; and classifying the stained blood cells based on the average red intensity and the average blue intensity.

[0194] Example 30

[0195] The image analysis method of any one of Examples 18 to 29, wherein determining the maturity of the stained blood cells comprises utilizing pixel foreground counting.

[0196] Example 31

[0197] The image analysis method of any one of Examples 18 to 24 or 27 to 29, wherein determining the maturity of the stained blood cells comprises determining whether a pixel foreground count exceeds a threshold amount.

[0198] Example 32

[0199] The image analysis method of any one of Examples 18 to 31, wherein the method comprises: depositing a staining composition into a chamber; depositing the biological sample into the chamber; and mixing the biological sample and the staining composition in the chamber.

[0200] Example 33

[0201] The image analysis method according to Example 32 further comprises: preheating the staining composition in the chamber before depositing the biological sample in the chamber; and heating the biological sample and the staining composition in the chamber.

[0202] Example 34

[0203] The image analysis method according to Example 33, wherein both preheating the staining composition in the chamber and heating the biological sample and the staining composition in the chamber include: using a heating coil to increase the chamber temperature.

[0204] Example 35

[0205] An image analysis method according to Example 34, wherein the heating coil is mounted on a mounting bracket connected to the chamber and one or more other mixing chambers.

[0206] Example 36

[0207] An image analysis system for detecting the maturity of blood cells comprises: one or more chambers configured to receive a biological sample and a staining composition; a camera configured to capture images of the stained blood cells; and one or more processors configured to determine the maturity of the stained blood cells based on the images of the stained blood cells captured by the camera.

[0208] Example 37

[0209] The image analysis system of Example 36, wherein determining the maturity of the stained blood cells comprises classifying the stained blood cells as at least one of reticulocytes or platelets.

[0210] Example 38

[0211] An image analysis system according to Example 37, wherein the one or more processors are further configured to: obtain a plurality of additional images, wherein each image from the plurality of additional images depicts additional stained blood cells corresponding to the image; for each image from the plurality of additional images, determine the maturity of the additional stained blood cells corresponding to the image; and generate an analysis output based on a set of maturity data, wherein the set of maturity data includes the maturity of the stained blood cells and the maturity of the stained blood cells depicted in the plurality of additional images.

[0212] Example 39

[0213] The image analysis system of Example 38, wherein generating the analysis output comprises determining at least one of an immature reticulocyte fraction and an immature platelet fraction of the biological sample.

[0214] Example 40

[0215] An image analysis system according to any one of Examples 38 to 39, wherein generating the analysis output includes generating at least one of: a curve graph showing the distribution of reticulocytes in the biological sample within a range of reticulocyte maturity; and a curve graph showing the distribution of platelets in the biological sample within a range of platelet maturity.

[0216] Example 41

[0217] The system of any of Examples 37 to 40, wherein the one or more processors are configured to classify the stained blood cells using a convolutional neural network.

[0218] Example 42

[0219] A system according to Example 41, wherein: the convolutional neural network is included by a machine learning model, and the machine learning model is configured to, when presented with input stained blood cells: classify the input stained blood cells as reticulocytes by classifying the input stained blood cells into a reticulocyte category; provide the maturity of the input stained blood cells when the input stained blood cells are classified into the reticulocyte category; classify the input stained blood cells as platelets by classifying the input stained blood cells into a platelet category; and provide the maturity of the input stained blood cells when the input stained blood cells are classified into the platelet category; and the one or more processors are configured to determine the maturity of the stained blood cells by performing actions, the actions comprising: presenting the stained blood cells as the input stained blood cells to the machine learning model; and treating the maturity provided by the machine learning model as the maturity of the stained blood cells.

[0220] Example 43

[0221] A system according to Example 42, wherein: the machine learning model is configured to: classify the input stained blood cells into the platelet category, and when the input stained blood cells are classified into the platelet category, provide the maturity of the input stained blood cells by classifying the input stained blood cells into a category selected from the group consisting of: an immature platelet category; and a mature platelet category; and the machine learning model is configured to: classify the input stained blood cells into the reticulocyte category, and when the input stained blood cells are classified into the reticulocyte category, provide the maturity of the input stained blood cells by classifying the input stained blood cells into a category selected from the group consisting of: an immature reticulocyte category; and a mature reticulocyte category.

[0222] Example 44

[0223] A system according to any of Examples 37 to 41, wherein the one or more processors are configured to determine the maturity of the stained blood cells by performing actions including generating the maturity of the stained blood cells after the stained blood cells are classified.

[0224] Example 45

[0225] The system of any of Examples 37 to 41 or 44, wherein the one or more processors are configured to classify the stained blood cells based on identifying a foreground in the image based on an intensity of each pixel in the image.

[0226] Example 46

[0227] A system according to Example 45, wherein: the one or more processors are further configured to calculate: an average red intensity of pixels in the foreground of the image; and an average blue intensity of pixels in the foreground of the image; and the one or more processors are configured to classify the stained blood cells based on the average red intensity and the average blue intensity.

[0228] Example 47

[0229] The system of any of Examples 36 to 41 or 44 to 46, wherein the one or more processors are configured to determine the maturity of the stained blood cells based on pixel foreground counts.

[0230] Example 48

[0231] The system of any of Examples 36 to 41 or 44 to 47, wherein the one or more processors are configured to determine that the stained blood cells are immature based on a pixel foreground count exceeding a threshold amount.

[0232] Example 49

[0233] The system of any of Examples 36 to 48, wherein the one or more chambers are further configured to receive a lysis composition.

[0234] Example 50

[0235] The system of any of Examples 36 to 48, wherein the staining composition further comprises a lytic compound.

[0236] Example 51

[0237] The system of any one of Examples 36 to 50, wherein the system further comprises a heater configured to heat the biological sample and the staining composition.

[0238] Example 52

[0239] The system of Example 51 wherein the heater comprises an induction coil.

[0240] Example 53

[0241] The system of any of Examples 51 to 52, wherein the one or more chambers include two or more chambers, and the heater is linked to all of the two or more chambers.

[0242] Example 54

[0243] A system according to Example 53, wherein the heater includes a mounting bracket connected to all of the two or more chambers, and a heating element mounted to the mounting bracket.

[0244] Example 55

[0245] An image analysis system for detecting the maturity of cells in a biological sample comprises: a flow cell configured to flow stained cells therethrough; a camera configured to capture images of the stained blood cells as they flow past the camera in an imaging region of the flow cell; and one or more processors configured to determine the maturity of the stained blood cells based on the images of the stained blood cells captured by the camera.

[0246] Example 56

[0247] An image analysis system according to Example 55, wherein: determining the maturity of the stained blood cells includes classifying the stained blood cells; and classifying the stained blood cells includes classifying the stained blood cells as at least one of reticulocytes or platelets.

[0248] Example 57

[0249] An image analysis system according to Example 56, wherein the one or more processors are configured to: obtain a plurality of additional images, wherein each image from the plurality of additional images depicts additional stained blood cells corresponding to the image; for each image from the plurality of additional images, determine the maturity of the additional stained blood cells corresponding to the image; and generate an analysis output based on a set of maturity data, wherein the set of maturity data includes the maturity of the stained blood cells and the maturity of the stained blood cells depicted in the plurality of additional images.

[0250] Example 58

[0251] The image analysis system of Example 57, wherein generating the analysis output comprises determining at least one of an immature reticulocyte fraction and an immature platelet fraction of the biological sample.

[0252] Example 59

[0253] An image analysis system according to any one of Examples 57 to 58, wherein generating the analysis output includes generating at least one of: a curve graph showing the distribution of reticulocytes in the biological sample within a range of reticulocyte maturity; and a curve graph showing the distribution of platelets in the biological sample within a range of platelet maturity.

[0254] Example 60

[0255] An image analysis system according to any one of Examples 56 to 59, wherein classifying the stained blood cells includes utilizing a convolutional neural network.

[0256] Example 61

[0257] An image analysis system according to Example 60, wherein: the convolutional neural network is included by a machine learning model, and the machine learning model is configured to, when presented with input stained blood cells: classify the input stained blood cells as reticulocytes by classifying the input stained blood cells into a reticulocyte category; provide the maturity of the input stained blood cells when the input stained blood cells are classified into the reticulocyte category; classify the input stained blood cells as platelets by classifying the input stained blood cells into a platelet category; and provide the maturity of the input stained blood cells when the input stained blood cells are classified into the platelet category; and determining the maturity of the stained blood cells includes: presenting the stained blood cells as the input stained blood cells to the machine learning model; and treating the maturity provided by the machine learning model as the maturity of the stained blood cells.

[0258] Example 62

[0259] An image analysis system according to Example 61, wherein: the machine learning model is configured to: classify the input stained blood cells into the platelet category, and when the input stained blood cells are classified into the platelet category, provide the maturity of the input stained blood cells by classifying the input stained blood cells into a category selected from the group including the following categories: an immature platelet category; and a mature platelet category; and the machine learning model is configured to: classify the input stained blood cells into the reticulocyte category, and when the input stained blood cells are classified into the reticulocyte category, provide the maturity of the input stained blood cells by classifying the input stained blood cells into a category selected from the group including the following categories: an immature reticulocyte category; and a mature reticulocyte category.

[0260] Example 63

[0261] An image analysis system according to any one of Examples 56 to 60, wherein determining the maturity of the stained blood cells includes generating the maturity of the stained blood cells after the stained blood cells are classified.

[0262] Example 64

[0263] An image analysis system according to any one of Examples 56 to 59, wherein classifying the stained blood cells includes identifying a foreground in the image based on the intensity of each pixel in the image.

[0264] Example 65

[0265] The image analysis system according to Example 64 also includes: calculating: the average red intensity of pixels in the foreground of the image; and the average blue intensity of pixels in the foreground of the image; and classifying the stained blood cells based on the average red intensity and the average blue intensity.

[0266] Example 66

[0267] An image analysis system according to any one of Examples 55 to 60, wherein determining the maturity of the stained blood cells includes utilizing pixel foreground counting.

[0268] Example 67

[0269] The image analysis system of any of Examples 55 to 60 or 63 to 65, wherein determining the maturity of the stained blood cells comprises determining whether a pixel foreground count exceeds a threshold amount.

[0270] Example 68

[0271] The image analysis system of any one of Examples 55 to 67, wherein the system further comprises one or more chambers adapted to receive the biological sample and the lysis composition.

[0272] Example 69

[0273] The image analysis system of any one of Examples 55 to 68, wherein the system comprises one or more chambers configured to receive the biological sample and the staining composition.

[0274] Example 70

[0275] The image analysis system of Example 69, wherein the system further comprises a heater configured to heat the biological sample and the staining composition.

[0276] Example 71

[0277] An image analysis system according to Example 71, wherein the heater includes an induction coil.

[0278] Example 72

[0279] An image analysis system according to any one of Examples 70 to 71, wherein the one or more chambers include two or more chambers, and the heater is linked to all of the two or more chambers.

[0280] Example 73

[0281] An image analysis system according to Example 72, wherein the heater includes a mounting bracket connected to all of the two or more chambers, and a heating element mounted to the mounting bracket.

[0282] Example 74

[0283] A machine comprising: a camera; and means for determining the maturity of cells in an image captured by the camera.

[0284] A computer or other processor with hardware, software and / or firmware can be used to perform each of the calculations or operations described herein. Various method steps can be performed by modules, and the modules can include any of various digital and / or analog data processing hardware and / or software arranged to perform the method steps described herein. The module optionally includes data processing hardware that is suitable for performing one or more of these steps by having an appropriate machine programming code associated therewith, and the modules for two or more steps (or a part of two or more steps) are integrated into a single processor board or separated into different processor boards in any of various integrated and / or distributed processing architectures. These methods and systems will generally adopt tangible media that implement machine-readable code with instructions for performing the above-mentioned method steps. Suitable tangible media can include memory (including volatile memory and / or non-volatile memory), storage media (such as magnetic recording on floppy disks, hard disks, magnetic tapes, etc.; magnetic recording on optical storage devices such as CDs, CD-R / Ws, CD-ROMs, DVDs, etc.; or any other digital or analog storage media), etc.

[0285] All patents, patent publications, patent applications, journal articles, books, technical references, etc. discussed in the instant disclosure are incorporated herein by reference in their entirety for all purposes.

[0286] Different arrangements of components depicted in the figures or described above, as well as components and steps not shown or described, are possible. Similarly, some features and subcombinations are useful and can be employed without reference to other features and subcombinations. The embodiments of the present invention have been described for illustrative, not limiting, purposes, and alternative embodiments will become apparent to readers of this patent. In some cases, method steps or operations may be performed or implemented in a different order, or operations may be added, deleted, or modified. It is understood that in certain aspects of the present invention, a single component may be replaced by multiple components, and multiple components may be replaced by a single component to provide an element or structure or to perform a given function or functions. Except where such substitutions would not be operable to practice certain embodiments of the present invention, such substitutions are considered within the scope of the present invention. Accordingly, the claims should not be construed as limited to the examples, figures, embodiments, and description provided above, but should be construed as having the scope provided when their terms are given the broadest reasonable interpretations provided by a commonly used dictionary. Except where a term or phrase is indicated under the heading "Defined" as having a specific meaning, it should be understood to have that meaning when used in the claims.

[0287] Clear definition

[0288] It should be understood that in the above examples and claims, the statement that something is "based on" something else should be understood to mean that it is at least partially determined by the thing it is indicated as being based on. In order to indicate that something must be completely determined based on something else, it is described as being "completely based on" whatever it must be completely determined by.

[0289] It should be understood that in the above examples and claims, the phrase "means for determining the maturity of cells in an image captured by a camera" is a means-plus-function limitation as provided in 35 U.S.C. § 112(f), where the function is "determining the maturity of cells in an image captured by a camera," and the corresponding structure is a computer that uses the Figure 9 Determine 902 maturity as shown, Figure 10 The classification 1000 of cells depicted in (in some embodiments described as being included in determining 902 maturity), as Figure 13 As shown in the figure, 1300 maturity data are generated. Figure 14 Determine 1402 the maturity of the cells in the attached image, as shown, and Figure 15 The algorithm depicted in the context of utilizing 1502 a convolutional neural network and considering the maturity provided by the machine learning model as 1506 the maturity of the cell determines the maturity of the cell.

[0290] It should be understood that in the above examples and claims, the term "group" should be understood as one or more things grouped together.

Claims

1. A computer-implemented image analysis method for detecting the maturity of blood cells, comprising: depositing a staining composition into the chamber; depositing a biological sample into the chamber; mixing the biological sample and the staining composition in the chamber; capturing an image of stained blood cells from the biological sample with a camera; as well as The maturity of the stained blood cells is determined based on the image of the stained blood cells captured by the camera.

2. The image analysis method according to claim 1, wherein: The method further comprises: treating the biological sample with a lysing agent; and flowing the biological sample through a flow cell and past the camera; and Capturing the image of the stained blood cells with the camera is performed while the biological specimen is flowing through the flow cell and past the camera.

3. The image analysis method according to claim 2, wherein: Determining the maturity of the stained blood cells includes classifying the stained blood cells; and Classifying the stained blood cells includes classifying the stained blood cells into at least one of reticulocytes or platelets.

4. The image analysis method according to claim 3, wherein: The method further comprises: obtaining a plurality of additional images, wherein each image from the plurality of additional images depicts additional stained blood cells corresponding to the image; for each image from the plurality of additional images, determining a maturity of the additional stained blood cells corresponding to the image; and An analysis output is generated based on a set of maturity data including the maturity of the stained blood cells and the maturity of the stained blood cells depicted in the plurality of additional images.

5. The image analysis method according to claim 4, wherein: Generating the analysis output includes determining at least one of an immature reticulocyte fraction and an immature platelet fraction of the biological sample.

6. The image analysis method according to any one of claims 4 to 5, wherein: Generating the analysis output includes generating at least one of: a graph showing the distribution of reticulocytes in the biological sample over a range of reticulocyte maturity; and A graph showing the distribution of platelets in the biological sample across a range of platelet maturity.

7. The image analysis method according to any one of claims 3 to 6, wherein: Classifying the stained blood cells includes utilizing a convolutional neural network.

8. The image analysis method according to claim 7, wherein: The convolutional neural network is comprised of a machine learning model configured to, when presented with input stained blood cells: classifying the input stained blood cells as reticulocytes by classifying the input stained blood cells into a reticulocyte class; providing a maturity of the input stained blood cells when the input stained blood cells are classified into the reticulocyte class; classifying the input stained blood cells as platelets by classifying the input stained blood cells into a platelet class; as well as providing a maturity of the input stained blood cells when the input stained blood cells are classified into the platelet class; and Determining the maturity of the stained blood cells includes: presenting the stained blood cells as the input stained blood cells to the machine learning model; as well as The maturity provided by the machine learning model is regarded as the maturity of the stained blood cells.

9. The image analysis method according to claim 8, wherein: The machine learning model is configured to classify the input stained blood cells into the platelet class, and when the input stained blood cells are classified into the platelet class, provide a maturity of the input stained blood cells by classifying the input stained blood cells into a class selected from the group consisting of: Immature platelet categories; as well as mature platelet categories; and The machine learning model is configured to classify the input stained blood cells into the reticulocyte class, and when the input stained blood cells are classified into the reticulocyte class, provide a maturity of the input stained blood cells by classifying the input stained blood cells into a class selected from the group consisting of: Immature reticulocyte category; as well as Mature reticulocyte class.

10. The image analysis method according to any one of claims 3 to 7, wherein: Determining the maturity of the stained blood cells includes generating the maturity of the stained blood cells after the stained blood cells are classified.

11. The image analysis method according to any one of claims 3 to 7, wherein: Classifying the stained blood cells includes identifying foreground in the image based on the intensity of each pixel in the image.

12. The image analysis method according to claim 11, comprising: calculate: an average red intensity of pixels in the foreground of the image; as well as an average blue intensity of pixels in the foreground of the image; as well as The stained blood cells are classified based on the mean red intensity and the mean blue intensity.

13. The image analysis method according to any one of claims 1 to 12, wherein: Determining the maturity of the stained blood cells includes utilizing pixel foreground counting.

14. The image analysis method according to any one of claims 1 to 7 or 10 to 12, wherein: Determining the maturity of the stained blood cells includes determining whether a pixel foreground count exceeds a threshold amount.

15. The image analysis method according to any one of claims 1 to 14, wherein: The method comprises: preheating the staining composition in the chamber prior to depositing the biological specimen in the chamber; and The biological sample and the staining composition are heated in the chamber.

16. The image analysis method according to claim 15, wherein: Both preheating the staining composition in the chamber and heating the biological sample and the staining composition in the chamber include increasing the chamber temperature using a heating coil.

17. The image analysis method according to claim 16, wherein: The heating coil is mounted on a mounting bracket connected to the chamber and one or more other mixing chambers.

18. A computer-implemented image analysis method for detecting the maturity of blood cells, comprising: flowing the biological sample through the flow cell and past the camera; capturing an image of stained blood cells with the camera while the biological sample is flowing through the flow cell and past the camera; as well as The maturity of the stained blood cells is determined based on the image of the stained blood cells captured by the camera.

19. The image analysis method according to claim 18, wherein: The method further includes treating the biological sample with a lysing agent.

20. The image analysis method according to claim 19, wherein: Determining the maturity of the stained blood cells includes classifying the stained blood cells; and Classifying the stained blood cells includes classifying the stained blood cells into at least one of reticulocytes or platelets.

21. The image analysis method according to claim 20, wherein: The method further comprises: obtaining a plurality of additional images, wherein each image from the plurality of additional images depicts additional stained blood cells corresponding to the image; for each image from the plurality of additional images, determining a maturity of the additional stained blood cells corresponding to the image; and An analysis output is generated based on a set of maturity data including the maturity of the stained blood cells and the maturity of the stained blood cells depicted in the plurality of additional images.

22. The image analysis method according to claim 21, wherein: Generating the analysis output includes determining at least one of an immature reticulocyte fraction and an immature platelet fraction of the biological sample.

23. The image analysis method according to any one of claims 21 to 22, wherein: Generating the analysis output includes generating at least one of: a graph showing the distribution of reticulocytes in the biological sample over a range of reticulocyte maturity; and A graph showing the distribution of platelets in the biological sample across a range of platelet maturity.

24. The image analysis method according to any one of claims 20 to 23, wherein: Classifying the stained blood cells includes utilizing a convolutional neural network.

25. The image analysis method according to claim 24, wherein: The convolutional neural network is comprised of a machine learning model configured to, when presented with input stained blood cells: classifying the input stained blood cells as reticulocytes by classifying the input stained blood cells into a reticulocyte class; providing a maturity of the input stained blood cells when the input stained blood cells are classified into the reticulocyte class; classifying the input stained blood cells as platelets by classifying the input stained blood cells into a platelet class; as well as providing a maturity of the input stained blood cells when the input stained blood cells are classified into the platelet class; and Determining the maturity of the stained blood cells includes: presenting the stained blood cells as the input stained blood cells to the machine learning model; as well as The maturity provided by the machine learning model is regarded as the maturity of the stained blood cells.

26. The image analysis method according to claim 25, wherein: The machine learning model is configured to classify the input stained blood cells into the platelet class, and when the input stained blood cells are classified into the platelet class, provide a maturity of the input stained blood cells by classifying the input stained blood cells into a class selected from the group consisting of: Immature platelet categories; as well as mature platelet categories; and The machine learning model is configured to classify the input stained blood cells into the reticulocyte class, and when the input stained blood cells are classified into the reticulocyte class, provide a maturity of the input stained blood cells by classifying the input stained blood cells into a class selected from the group consisting of: Immature reticulocyte category; as well as Mature reticulocyte class.

27. The image analysis method according to any one of claims 20 to 24, wherein: Determining the maturity of the stained blood cells includes generating the maturity of the stained blood cells after the stained blood cells are classified.

28. The image analysis method according to any one of claims 20 to 23, wherein: Classifying the stained blood cells includes identifying foreground in the image based on the intensity of each pixel in the image.

29. The image analysis method according to claim 28, further comprising: calculate: an average red intensity of pixels in the foreground of the image; as well as an average blue intensity of pixels in the foreground of the image; as well as The stained blood cells are classified based on the mean red intensity and the mean blue intensity.

30. The image analysis method according to any one of claims 18 to 29, wherein: Determining the maturity of the stained blood cells includes utilizing pixel foreground counting.

31. The image analysis method according to any one of claims 18 to 24 or 27 to 29, wherein: Determining the maturity of the stained blood cells includes determining whether a pixel foreground count exceeds a threshold amount.

32. The image analysis method according to any one of claims 18 to 31, wherein: The method comprises: depositing a staining composition into the chamber; depositing the biological sample into the chamber; and The biological sample and the staining composition are mixed in the chamber.

33. The image analysis method according to claim 32, wherein: The method further comprises: preheating the staining composition in the chamber prior to depositing the biological specimen in the chamber; and The biological sample and the staining composition are heated in the chamber.

34. The image analysis method according to claim 33, wherein: Both preheating the staining composition in the chamber and heating the biological sample and the staining composition in the chamber include increasing the chamber temperature using a heating coil.

35. The image analysis method according to claim 34, wherein: The heating coil is mounted on a mounting bracket connected to the chamber and one or more other mixing chambers.

36. An image analysis system for detecting the maturity of blood cells, comprising: one or more chambers configured to receive a biological sample and a staining composition; a camera configured to capture images of stained blood cells; as well as One or more processors configured to determine a maturity of the stained blood cells based on the images of the stained blood cells captured by the camera.

37. The image analysis system according to claim 36, wherein: Determining the maturity of the stained blood cells includes classifying the stained blood cells as at least one of reticulocytes or platelets.

38. The image analysis system according to claim 37, wherein: The one or more processors are further configured to: obtaining a plurality of additional images, wherein each image from the plurality of additional images depicts additional stained blood cells corresponding to the image; for each image from the plurality of additional images, determining a maturity of the additional stained blood cells corresponding to the image; and An analysis output is generated based on a set of maturity data including the maturity of the stained blood cells and the maturity of the stained blood cells depicted in the plurality of additional images.

39. The image analysis system according to claim 38, wherein: Generating the analysis output includes determining at least one of an immature reticulocyte fraction and an immature platelet fraction of the biological sample.

40. An image analysis system according to any one of claims 38 to 39, wherein: Generating the analysis output includes generating at least one of: a graph showing the distribution of reticulocytes in the biological sample over a range of reticulocyte maturity; and A graph showing the distribution of platelets in the biological sample across a range of platelet maturity.

41. A system according to any one of claims 37 to 40, wherein: The one or more processors are configured to classify the stained blood cells using a convolutional neural network.

42. The system of claim 41, wherein: The convolutional neural network is comprised of a machine learning model configured to, when presented with input stained blood cells: classifying the input stained blood cells as reticulocytes by classifying the input stained blood cells into a reticulocyte class; providing a maturity of the input stained blood cells when the input stained blood cells are classified into the reticulocyte class; classifying the input stained blood cells as platelets by classifying the input stained blood cells into a platelet class; as well as providing a maturity of the input stained blood cells when the input stained blood cells are classified into the platelet class; and The one or more processors are configured to determine the maturity of the stained blood cells by performing actions comprising: presenting the stained blood cells as the input stained blood cells to the machine learning model; as well as The maturity provided by the machine learning model is regarded as the maturity of the stained blood cells.

43. The system of claim 42, wherein: The machine learning model is configured to classify the input stained blood cells into the platelet class, and when the input stained blood cells are classified into the platelet class, provide a maturity of the input stained blood cells by classifying the input stained blood cells into a class selected from the group consisting of: Immature platelet categories; as well as mature platelet categories; and The machine learning model is configured to classify the input stained blood cells into the reticulocyte class, and when the input stained blood cells are classified into the reticulocyte class, provide a maturity of the input stained blood cells by classifying the input stained blood cells into a class selected from the group consisting of: Immature reticulocyte category; as well as Mature reticulocyte class.

44. A system according to any one of claims 37 to 41, wherein: The one or more processors are configured to determine the maturity of the stained blood cells by performing actions including generating the maturity of the stained blood cells after the stained blood cells are classified.

45. A system according to any one of claims 37 to 41 or 44, wherein: The one or more processors are configured to classify the stained blood cells based on identifying a foreground in the image based on an intensity of each pixel in the image.

46. ​​The system of claim 45, wherein: The one or more processors are further configured to compute: The average red intensity of pixels in the foreground of the image; and an average blue intensity of pixels in the foreground of the image; and The one or more processors are configured to classify the stained blood cells based on the mean red intensity and the mean blue intensity.

47. A system according to any one of claims 36 to 41 or 44 to 46, wherein: The one or more processors are configured to determine a maturity of the stained blood cells based on the pixel foreground counts.

48. A system according to any one of claims 36 to 41 or 44 to 47, wherein: The one or more processors are configured to determine that the stained blood cells are immature based on a pixel foreground count exceeding a threshold amount.

49. A system according to any one of claims 36 to 48, wherein: The one or more chambers are also configured to receive a lysis composition.

50. A system according to any one of claims 36 to 48, wherein: The dyeing composition further comprises a lytic compound.

51. A system according to any one of claims 36 to 50, wherein: The system also includes a heater configured to heat the biological sample and the staining composition.

52. The system of claim 51, wherein: The heater includes an induction coil.

53. A system according to any one of claims 51 to 52, wherein: The one or more chambers include two or more chambers, and the heater is linked to all of the two or more chambers.

54. The system of claim 53, wherein: The heater includes a mounting bracket connected to all of the two or more chambers, and a heating element mounted to the mounting bracket.

55. An image analysis system for detecting the maturity of cells in a biological sample, comprising: a flow cell configured to flow the stained cells through; a camera configured to capture images of the stained blood cells as they flow past the camera in an imaging region of the flow cell; as well as One or more processors configured to determine a maturity of the stained blood cells based on the images of the stained blood cells captured by the camera.

56. The image analysis system of claim 55, wherein: Determining the maturity of the stained blood cells includes classifying the stained blood cells; and Classifying the stained blood cells includes classifying the stained blood cells into at least one of reticulocytes or platelets.

57. The image analysis system according to claim 56, wherein: The one or more processors are configured to: obtaining a plurality of additional images, wherein each image from the plurality of additional images depicts additional stained blood cells corresponding to the image; for each image from the plurality of additional images, determining a maturity of the additional stained blood cells corresponding to the image; and An analysis output is generated based on a set of maturity data including the maturity of the stained blood cells and the maturity of the stained blood cells depicted in the plurality of additional images.

58. The image analysis system according to claim 57, wherein: Generating the analysis output includes determining at least one of an immature reticulocyte fraction and an immature platelet fraction of the biological sample.

59. An image analysis system according to any one of claims 57 to 58, wherein: Generating the analysis output includes generating at least one of: a graph showing the distribution of reticulocytes in the biological sample over a range of reticulocyte maturity; and A graph showing the distribution of platelets in the biological sample across a range of platelet maturity.

60. An image analysis system according to any one of claims 56 to 59, wherein: Classifying the stained blood cells includes utilizing a convolutional neural network.

61. The image analysis system of claim 60, wherein: The convolutional neural network is comprised of a machine learning model configured to, when presented with input stained blood cells: classifying the input stained blood cells as reticulocytes by classifying the input stained blood cells into a reticulocyte class; providing a maturity of the input stained blood cells when the input stained blood cells are classified into the reticulocyte class; classifying the input stained blood cells as platelets by classifying the input stained blood cells into a platelet class; as well as providing a maturity of the input stained blood cells when the input stained blood cells are classified into the platelet class; and Determining the maturity of the stained blood cells includes: presenting the stained blood cells as the input stained blood cells to the machine learning model; as well as The maturity provided by the machine learning model is regarded as the maturity of the stained blood cells.

62. The image analysis system of claim 61, wherein: The machine learning model is configured to classify the input stained blood cells into the platelet class, and when the input stained blood cells are classified into the platelet class, provide a maturity of the input stained blood cells by classifying the input stained blood cells into a class selected from the group consisting of: Immature platelet categories; as well as mature platelet categories; and The machine learning model is configured to classify the input stained blood cells into the reticulocyte class, and when the input stained blood cells are classified into the reticulocyte class, provide a maturity of the input stained blood cells by classifying the input stained blood cells into a class selected from the group consisting of: Immature reticulocyte category; as well as Mature reticulocyte class.

63. An image analysis system according to any one of claims 56 to 60, wherein: Determining the maturity of the stained blood cells includes generating the maturity of the stained blood cells after the stained blood cells are classified.

64. An image analysis system according to any one of claims 56 to 59, wherein: Classifying the stained blood cells includes identifying foreground in the image based on the intensity of each pixel in the image.

65. The image analysis system of claim 64, further comprising: calculate: an average red intensity of pixels in the foreground of the image; as well as an average blue intensity of pixels in the foreground of the image; as well as The stained blood cells are classified based on the mean red intensity and the mean blue intensity.

66. An image analysis system according to any one of claims 55 to 60, wherein: Determining the maturity of the stained blood cells includes utilizing pixel foreground counting.

67. An image analysis system according to any one of claims 55 to 60 or 63 to 65, wherein: Determining the maturity of the stained blood cells includes determining whether a pixel foreground count exceeds a threshold amount.

68. An image analysis system according to any one of claims 55 to 67, wherein: The system also includes one or more chambers adapted to receive the biological sample and the lysis composition.

69. An image analysis system according to any one of claims 55 to 68, wherein: The system includes one or more chambers configured to receive the biological sample and a staining composition.

70. The image analysis system of claim 69, wherein: The system also includes a heater configured to heat the biological sample and the staining composition.

71. The image analysis system of claim 70, wherein: The heater includes an induction coil.

72. An image analysis system according to any one of claims 70 to 71, wherein: The one or more chambers include two or more chambers, and the heater is linked to all of the two or more chambers.

73. The image analysis system of claim 72, wherein: The heater includes a mounting bracket connected to all of the two or more chambers, and a heating element mounted to the mounting bracket.

74. A machine comprising: camera; as well as Means for determining the maturity of cells in images captured by said camera.

Images