Maturity classification of stained reticulocytes using optical microscopy

Abstract

To enable maturity classification of stained reticulocytes using optical microscopy.SOLUTION: The present invention relates to a method of performing maturity classification of reticulocytes from a whole blood sample, the method comprising: staining the sample with a supravital agglutinating dyeing reagent or a fluorescent agglutinating dye; illuminating the stained sample with a light beam to detect reticulocytes; determining, for each reticulocyte, parameters of (i) a fraction (Λ) of reticulum area (Ar) to the whole cell area (Ac), and (ii) a fraction (Γ) of a perimeter of the reticulum (Ur) to the reticulum area (Ar); and maturity-classifying each reticulocyte into one of four major maturity classes according to the determined values of Λ and Γ.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a method for classifying reticulocytes from a whole blood sample by maturity, comprising the steps of staining the sample with an ultrasound aggregation staining reagent or a fluorescent aggregation dye; illuminating the stained sample with a light beam to detect reticulocytes; determining, for each reticulocyte, the parameters of (i) the ratio (Λ) of the reticular area (Ar) to the total cell area (Ac); and (ii) the ratio (Γ) of the peripheral part (Ur) of the reticulum to the reticular area (Ar); and classifying the reticulocytes into one of four main maturity classes according to the values determined for Λ and Γ.

Background Art

[0002] Monitoring of erythropoiesis is essential for determining the health status of patients and their response to various treatments, for example, in iron deficiency anemia or recovery from chemotherapy. During the process of erythropoiesis, when late erythroblasts lose their nuclei, the developed cells are then called reticulocytes, which contain a reticulum, a filamentous RNA network. Reticulocytes obtained from peripheral blood are readily available and indicate the rate of red blood cell (RBC) regeneration by determining the cell count and maturity (Non-Patent Document 1). Fluorescent aggregation staining or ultrasound staining of the reticulum, which are two main methods for detecting reticulocytes, are generally used. Both staining methods aggregate thin filamentous RNA strands into network-like filaments, which can be observed under a microscope.

[0003] Automated hematological analyzers determine reticulocyte parameters more precisely, accurately, and reproducibly than cells counted manually with an optical microscope. However, each analyzer uses various reagents, which exhibit varying sensitivities to binding to RNA and other cellular components. Therefore, there is no universal classification for determining the maturity of reticulocytes using various staining methods (Non-Patent Literature 2). A method to unify the classification was attempted by determining the immature reticulocyte fraction (IRF) using a fluorescence method. IRF is the ratio of immature reticulocytes to the total number of reticulocytes (Non-Patent Literature 3). However, no consensus has been found for each class, and IRF values ​​also differ depending on the analyzer (Non-Patent Literature 4). Another classification method, based on the work of Ludwig Heilmeyer in the 1930s, defines four maturity classes, including immature reticulocytes with a dense reticular structure (Class 1), reticulocytes with a broad but loose reticular network (Class 2), reticulocytes with a scattered reticular network (Class 3), and mature reticulocytes with scattered reticular granules (Class 4) (Non-Patent Literature 4).

[0004] Due to differences among various automated hematological analyzers, manual counting under an optical microscope remains the absolute standard for the diagnostic determination of reticulocyte maturity class. Furthermore, there is growing interest in devices that optically mimic an optical microscope and can determine all parameters of whole blood counts.

[0005] Therefore, there is a need for standard metrological methods for analyzing stained reticulocytes, which may be linked to manual, and especially automated, microscopic analysis, as well as the resulting diagnostic detections. [Prior art documents] [Patent Documents]

[0006] [Non-Patent Document 1] Piva et al., 2015, Clinics in Laboratory Medicine, 35, 133–163 [Non-Patent Document 2] Van Den Bossche et al., 2002, Clin. Chem. Lab. Med., 40(1), 69–73 [Non-Patent Document 3] Heimpel et al., 2010, Med. Klin., 105, 538–543 [Non-Patent Document 4] Riley et al., 2001, J. Clin. Labor. Anal., 15, 267–294 [Overview of the project] [Means for solving the problem]

[0007] The present invention addresses this need and provides a method for classifying reticulocytes from a whole blood sample by maturity, comprising: (a) staining the sample with a hyperbiotic agglutination reagent or a fluorescent agglutination dye; (b) illuminating the stained sample with a light beam preferably in the wavelength range of 200 nm to 780 nm using a photodetector, preferably a microscope, to detect reticulocytes; (c) determining the parameters of each reticulocyte: (i) the ratio of reticular area (Ar) to total cell area (Ac) (Λ); and (ii) the ratio of reticular periphery (Ur) to reticular area (Ar) (Γ); and (d) classifying the reticulocytes by maturity into one of four major maturity classes according to the values ​​determined for Λ and Γ. The method advantageously allows for the classification of reticulocytes according to a standard metrological method that is largely independent of the platform or device used, and thus enables a uniform allocation of reticulocytes according to the well-established Heilmeyer classification scheme, which is familiar to most practitioners. The novel methodology according to the present invention further enables improved diagnostic readings based on time-dependent changes in reticulocyte counts after treatment or during the course of disease.

[0008] In a preferred embodiment of the present invention, the method further includes, as a final step, a step of enumerating the reticulocytes. It is particularly preferred that the enumeration be by class and by sample.

[0009] In a further preferred embodiment, staining is performed with a hyperbiotic agglutination staining reagent selected from NMB (new methylene blue), brilliant cresyl blue, crystal violet, methyl violet, and Nile blue. In yet another preferred embodiment, staining is performed with a fluorescent agglutination dye selected from acridine orange, auramine O, D-methyl oxacarbocyanide, ethidine bromide, pyronin Y, thioflavin-T, and thiazole orange.

[0010] The staining may further include a step of chemically crosslinking the nucleic acids, preferably with nitrogen mustard, cis-diaminedichloroplatinum(II) or a derivative, or chloroethylnitrosourea (CENU).

[0011] In another embodiment, step (c) described above is performed in an apparatus including an imaging module. The imaging module is preferably designed to perform morphological segmentation operations.

[0012] Further embodiments relate to the method defined above, further comprising, as step (e), a step of morphologically comparing each stained reticulocyte with an image repository of reticulocytes independently classified by an expert.

[0013] In a preferred embodiment, the morphological comparison includes applying images of stained reticulocytes to a machine learning-based method trained on images from the aforementioned reticulocyte image repository.

[0014] In one embodiment of the method according to the present invention, maturity classification includes assigning reticulocytes to classes 1, 2, 3, or 4 according to their Λ / Γ ratio. In a preferred embodiment, a Λ / Γ value of about >2.5 indicates class 1, a Λ / Γ value of about 1 to about 2.5 indicates class 2, a Λ / Γ value of about 0.35 to about 1 indicates class 3, and a Λ / Γ value of about <0.35 indicates class 4.

[0015] In a further embodiment, the present invention relates to a computer implementation method for maturing reticulocytes from one or more images obtained from a whole blood sample, comprising the steps of: determining the parameters of each reticulocyte in the image: (i) the ratio of reticular area (Ar) to total cell area (Ac) (Λ); and (ii) the ratio of reticular periphery (Ur) to reticular area (Ar) (Γ); and maturing the reticulocytes into four main classes according to the values ​​of Λ and Γ. The method preferably further comprises the step of enumerating the reticulocytes.

[0016] In another embodiment, the present invention relates to an in vitro method for monitoring and determining the health status of a subject and / or the subject's response to treatment, the method comprising performing the above-defined method of maturity classification of reticulocytes on one or more whole blood samples obtained from the subject.

[0017] In a preferred embodiment of the in vitro method for monitoring and determining the health status of a subject and / or the subject's response to treatment, when an initial sample is compared to a second or subsequent sample taken from the subject after a period of 2, 3, 4, 5, 6, 7 days or more, an increase in class 2, 3, or 4, preferably class 3 or 4, more preferably class 4, reticulocyte counts indicate improvement in the subject's health and / or a positive response to treatment in hyporeticoplasmosis, or deterioration in the subject's health and / or a negative response to treatment in hyperreticoplasmosis.

[0018] In a further embodiment, the present invention relates to a method for classifying reticulocytes as defined above in terms of maturity, or to an apparatus comprising means for performing an in vitro method for monitoring and determining the health status and / or response of an object to treatment as defined above.

[0019] In another aspect, the present invention relates to a data processing apparatus including means for performing a computer-implemented method for classifying reticulocytes by maturity from one or more of the images defined above.

[0020] In a last aspect, the present invention relates to a computer program including instructions for causing a computer to execute a method for classifying reticulocytes by maturity including the morphological comparison step defined above, or a computer-implemented method as defined herein, when the program is executed by the computer.

[0021] It should be understood that the features mentioned above and the following features not yet described can be used not only in the combinations shown, but also in other combinations or alone, without departing from the scope of the present invention.

Brief Description of the Drawings

[0022] [Figure 1] It is a graph showing the quantitative maturity classification of reticulocytes recorded in bright-field transmission mode according to the present invention. The classification is based on the ratio (Λ) of the area of the stained reticular body within the cell compared to the total cell area, and on the ratio (Γ) of the perimeter of that area to the area of all RNA fragments. Four classes of reticulocytes proposed by Highmeyer are shown. Class 1 (closed circle (1)) is the most immature class, and an exemplary cell is shown in (2). Class 2 (open circle with dots (3)) corresponds to the second youngest group of reticulocytes (an exemplary cell is shown in (4)), followed by Class 3 (open circle (5)) of an exemplary cell (6), and Class 4 (dotted circle (7); exemplary cell (8)) is the most mature group. [Figure 2]This is an example of a bright-field transmission color (9) image divided into separate RGB (R: red (10), G: green (11), B: blue (12)) 8-bit images of stained reticulocytes. The cell size is determined using the Otsu threshold for the red channel. For the determination of the reticulum area, since the cell wall appears exclusively in the blue channel, a mask of stained RNA is generated by subtracting the green channel from the blue channel. Further, by applying the automatic threshold of ImageJ to the generated image, a mask of the stained reticulum is obtained (13). [Figure 3] It is a different descriptive diagram of the quantitative maturity classification of reticulocytes recorded in the bright-field transmission mode of the present invention shown in FIG. 1. In this figure, the Λ / Γ ratio is shown in comparison with the cell number (14). The figure shows class 1 (closed circle (1)), class 2 (open circle with dots (3)), class 3 (open circle (5)) and class 4 (dotted circle (7)) according to the scheme of Highmeyer.

Mode for Carrying Out the Invention

[0023] The present invention is described with respect to specific embodiments, but this description should not be construed in a limiting sense.

[0024] Before describing the exemplary embodiments of the present invention in detail, important definitions are presented for understanding the present invention.

[0025] As used in this specification and the appended claims, the singular forms "a" and "an" include each plural form unless the context clearly dictates otherwise.

[0026] In the context of the present invention, the term "about" indicates an interval of accuracy that a person skilled in the art would understand to still guarantee the technical effect of the feature in question. The term typically indicates a deviation from the indicated numerical value of ±25%. In certain embodiments, the term may also indicate a deviation from the indicated numerical value of ±15%, ±10%, ±5%, ±3%, ±2%, ±1% or ±0.5%.

[0027] It should be understood that the term "comprising" is not limiting. For the purposes of this invention, the terms "consisting of" or "essentially consisting of" are interpreted as preferred embodiments of the term "comprising of." Hereinafter, where a group is defined to include at least a certain number of embodiments, this also means that it preferably includes a group consisting only of these embodiments.

[0028] Furthermore, terms such as "(i)", "(ii)", "(iii)", or "(a)", "(b)", "(c)", "(d)", or "first", "second", "third", etc., in the description or claims are used to distinguish similar elements and are not necessarily used to describe a consecutive or chronological order.

[0029] It should be understood that the terms used herein are interchangeable under appropriate circumstances, and that embodiments of the invention described herein may operate in any order other than those described or illustrated herein. Where terms relate to methods, procedures, or steps of use, unless otherwise indicated, there is no consistency in time or time intervals between steps; that is, steps may be performed simultaneously, or there may be time intervals of seconds, minutes, hours, days, weeks, etc., between such steps.

[0030] It should be understood that the present invention is not limited to the specific methodologies, protocols, etc., described herein, as these can be modified. It should also be understood that the technical terms used herein are for the purpose of describing only specific embodiments and are not intended to limit the scope of the present invention, which is limited only by the appended claims.

[0031] Drawings should be considered schematic diagrams, and the elements illustrated in them are not necessarily shown to scale. Rather, various elements are shown in a way that makes their function and general purpose clear to those skilled in the art.

[0032] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art.

[0033] As described above, in one embodiment, the present invention relates to a method for classifying the maturity of reticulocytes from a whole blood sample, comprising: (a) staining the sample with a hyperbiotic agglutination reagent or a fluorescent agglutination dye; (b) illuminating the stained sample with a light beam preferably in the wavelength range of 200 nm to 780 nm using a photodetector, preferably a microscope, to detect the reticulocytes; (c) determining the parameters of each reticulocyte: (i) the ratio of the reticular area (Ar) to the total cell area (Ac) (Λ); and (ii) the ratio of the peripheral area of ​​the reticular structure (Ur) to the reticular area (Ar) (Γ); and (d) classifying the reticulocytes into one of four major maturity classes according to the values ​​determined for Λ and Γ.

[0034] As used herein, the term “whole blood sample” refers to a blood sample of a mammal, preferably a human, obtained from a subject by a preferred method known to those skilled in the art. Samples used in the context of the present invention should preferably be collected in a clinically acceptable manner, more preferably in a manner that preserves nucleic acids, particularly RNA. “Whole blood” essentially comprises red blood cells and progenitor cells, leukemia and progenitor cells, and platelets suspended in plasma. In certain embodiments, the sample can be pooled.

[0035] The present invention preferably assumes the use of non-pooled samples. In certain embodiments of the present invention, the contents of a whole blood sample may be subjected to certain processing steps. For example, the sample may be diluted or concentrated. Furthermore, nucleic acid stabilizers or degradation inhibitors may be added. In particularly preferred embodiments, the use of an anticoagulant such as EDTA is assumed. In even more specific embodiments, the whole blood sample may be subjected to an initial cell sorting or cell separation step before staining. Such steps are assumed to concentrate and / or purify erythrocytes (including progenitor cells such as reticulocytes), or concentrate and / or purify reticulocytes.

[0036] The method according to the present invention assumes a first step of “staining” the whole blood sample with the whole blood sample defined above. This step can be carried out with any suitable staining agent capable of showing the area and periphery of the reticular structure. Thus, the staining agent can at least partially stabilize nucleic acid structures, particularly RNA structures, together with the cells, and preferably, when illuminated with light, make these structures visible under suitable optical conditions. As used herein, the term “reticular structure” refers to a mesh-like network of nucleic acids, particularly RNA, typically ribosomal RNA, which becomes visible under staining conditions. The reticular structure is a specific structure that distinguishes reticulocytes from other blood cells. As used herein, “reticulocyte” refers to immature red blood cells that do not have a cell nucleus. During erythrogenesis, reticulocytes develop and mature in the bone marrow, and then circulate in the bloodstream for about 1 day before developing into mature red blood cells. In healthy subjects, the percentage of reticulocytes in the blood is typically around 0.5%–2.5% in adults and 2%–6% in infants. The reticulocyte count in whole blood samples is typically used as an indicator of bone marrow activity because it reflects recent erythropoiesis events.

[0037] Since nucleic acid morphology is present in the reticular structure, which allows reticulocytes to be distinguished from mature red blood cells and other cells, staining is preferably carried out with a staining agent that produces an agglutination effect and allows for suitable contrast in optical detection procedures. Such a staining method is based on an agglutination staining reagent in a preferred embodiment of the present invention.

[0038] As used herein, the term "aggregation staining reagent" refers to a staining agent that is presumed to bind to intracellular nucleic acid-containing structures, particularly ribosomes, and cause them to aggregate. In preferred embodiments, the present invention envisions the use of any suitable agglutination staining reagent having this function or capability.

[0039] In a preferred series of embodiments, the agglutination staining reagent is a supravilatory agglutination staining agent. "Superbiotic" dyes are typically used for staining living cells removed from an organism. Preferred examples of supravilatory agglutination staining agents include NMB (new methylene blue), brilliant cresyl blue, crystal violet, methyl violet, and Nile blue. The use of any suitable derivative or functional equivalent is further anticipated. The use of azur B or any suitable derivative is also anticipated. The present invention further anticipates the use of supravilatory agglutination staining reagents that may not yet be developed and that satisfy the above functions. The use of NMB (new methylene blue) is particularly preferred. Further information can be obtained from preferred literature sources such as, for example, Samuel M. Rapoport, 2019, The reticulocyte, 1st edition, CRC Press.

[0040] In a more preferred series of embodiments, the agglutination staining reagent is a fluorescent agglutination staining agent. Such dyes also interact with nucleic acid-containing structures, particularly ribosomes, and exert a fluorescent effect when exposed to suitable excitation light. Examples of such staining reagents include acridine orange, auramine O, D-methyloxacarbocyanide, ethidine bromide, pyronin Y, thioflavin-T, and thiazole orange. The use of any suitable derivative or functional equivalent thereof is further anticipated. The present invention further anticipates the use of fluorescent agglutination staining reagents that may not yet be developed and that satisfy the above functions. Further information can be obtained from suitable literature sources, for example, Samuel M. Rapoport, 2019, The reticulocyte, 1st edition, CRC Press.

[0041] Staining may be carried out according to a preferred procedure known to those skilled in the art. In certain embodiments, staining is carried out according to a procedure suggested by the manufacturer of the staining reagent. For example, a typical staining procedure may include adding the staining reagent to a mixture of whole blood samples as defined herein, and optionally a buffer, such as PBS, at a concentration of, for example, about 1%. The mixture is then incubated for a certain period, for example, 1 minute, 2 minutes, etc. The incubation time may be adapted to the dye used. For example, for staining methods using fluorescent dyes, the incubation time may be extended by several minutes, preferably according to the manufacturer's instructions.

[0042] For subsequent analysis, the stained sample may be provided in any preferred form. For example, the sample may be analyzed in a liquid or solution immediately after staining. Alternatively, the sample may be stored for a period of time before analysis. For such storage, the stained sample may be covered with a coverslide, for example, or mounted in an aqueous or non-aqueous mounting medium. When a mounting medium is used, the sample may be covered with a coverslip, especially if the sample needs to be stored for a long period of time. This method typically allows for stable and substantially permanent storage of the sample. Examples of preferred aqueous mounting mediums include Aquatex, gelatin, glycerin, Kaiser's glycerin gelatin, and Sorbitol F Solution E420. Preferred examples of non-aqueous mounting mediums include DPX, Entellan Rapid Mounting Medium, M-Glas Liquid Coverslip, and Neomount Anhydrous Mounting Medium. In further embodiments, cells may be fixed with any preferred fixation method and reagents. For example, glutaraldehyde can be used.

[0043] Further information can be obtained from suitable literature sources such as the Clinical Laboratory Standards Association document H44-A2 “Methods for Reticulocyte Counting (Automated Blood Cell Counters, Flow Cytometry, and Supravital Dyes); Approval Guidelines, 2nd Edition” or any further editions.

[0044] After staining and, if applicable, storage, the mixture may be placed on a suitable carrier for subsequent analytical steps, subjected to in situ analysis, or transferred to an analytical instrument. Alternatively, the staining may be performed in the same analytical instrument in which the subsequent analysis is performed.

[0045] In very specific embodiments, the staining procedure according to the present invention may be performed simultaneously with or prior to the crosslinking of a separate nucleic acid. Such crosslinking can be performed, for example, with any suitable crosslinking reagent known to those skilled in the art. Examples of such reagents include nitrogen mustard, i.e., alkylating agents having a bis-(2-ethylchloro)amine core structure with a variable R group, such as cyclophosphamide, chlorambucyl, uramustine, melphalan, or bendamustine. Further examples include cis-diaminedichloroplatinum(II), i.e., cisplatin, which can form intrachain or interchain crosslinks. Further variants or derivatives of cisplatin are conceivable. Another example is chloroethylnitrosourea (CENU), in particular carmustine (BCNU). Further crosslinking agents such as psoralen or mitomycin C are also conceivable. The staining step may follow or be performed simultaneously with the use of the crosslinking step. In further embodiments, the use of the crosslinking reagent can also be combined with staining with non-aggregating dyes, such as non-aggregating supravilative dyes or non-fluorescent agglutinating dyes.

[0046] After the staining process, the stained sample is illuminated. Illumination is performed using the light beam of a photodetector. It is assumed that light in the range of 200 nm to 780 nm is used. The wavelength of the light can be adapted to one or more factors, such as the properties of the dye and its excitation wavelength, the form of the photodetector and its functional spectrum. As used herein, the term “photodetector” refers to any optical system capable of detecting and visualizing light reflected from a sample, particularly cells in the sample such as reticulocytes. In a preferred embodiment, the photodetector is a microscope or microscope system capable of visualizing and / or fluorescently characterizing cells. It may include a light source, or be coupled to a light source, which may be either a laser or a light source for visual detection. The laser may be, in particular, a laser capable of stimulating a fluorescent staining reagent, preferably a fluorescent staining reagent as referred herein. Thus, the microscope system may be a system that enables fluorescence microscopy. The microscope system may receive a fluorescent response from a sample, e.g., cells being analyzed, to some kind of visual reflection and / or stimulation. The microscope may further include elements known to those skilled in the art, such as a focusing optical element that can be designed as a lens, and / or a diaphragm. The microscope system may further be connected to evaluation modules, image acquisition modules, AI modules or neural networks, computer systems, computer networks or interfaces, databases, image repositories or laboratory or hospital systems. In certain embodiments, the microscope system may include a system that includes a flow cytometer or flow cytometry function, or may be built essentially on such a system, particularly when fluorescent dyes are used. Examples of flow cytometry systems envisioned in the present invention include the Siemens ADVIA 2120i, Sysmex XN series, Sysmex XE series, Abbott Diagnostics Cell-DYN Sapphire, and Beckman Coulter HmX. The use of the Siemens ADVIA 2120i system is particularly preferred.In further embodiments, these systems can be combined with other systems, components or units of other systems, or individual additional components or units.

[0047] When stained samples within reticulocytes are illuminated, the reticular or reticular network structure becomes detectable. This structure can have various morphologies, area sizes, peripheral regions, and optical densities depending on the developmental stage of the reticulocyte. Typically, reticulocytes can be classified according to the following four classes (Heilmeyer): Class 1 = immature reticulocytes with a dense reticular structure; Class 2 = reticulocytes with a broad but loose reticular network; Class 3 = reticulocytes with a scattered reticular network; and Class 4 = mature reticulocytes with scattered reticular granules. Examples of Heilmeyer's classes 1-4 are shown in Figure 1.

[0048] In a further central step of the method of the present invention, parameters for each reticulocyte are determined, which allows for a quantitative and unbiased determination of the class, and thus the developmental state of the reticulocyte. These parameters are the ratio (Λ) of the reticular area (Ar) to the total cell area (Ac) and the ratio (Γ) of the reticular periphery (Ur) to the reticular area (Ar).

[0049] The ratio (Λ) of reticular area (Ar) to total cell area (Ac) can be determined as follows:

number

[0050] The ratio (Γ) of the peripheral area (Ur) of the reticular structure to the total reticular area (Ar) can be determined as follows:

number

[0051] The values ​​obtained are stored, for example, in a computer system, evaluation module, and / or database, for subsequent comparison or evaluation processes.

[0052] Typically, reticular or network portions smaller than approximately 200 nm may not be detectable in the context of the present invention, particularly when using the microscopy techniques described herein. Therefore, the detectable limit for reticular area is approximately 0.15 μm. 2 It can be set to this.

[0053] In subsequent processes, reticulocytes are classified into four major maturity classes. This classification is carried out according to the values ​​determined for Λ and Γ as defined above.

[0054] Advantageously, calculating the peripheral Ur of all nucleic acid fragments, particularly RNA fragments, and dividing it by the reticular area Ar, significantly improves the accuracy of classifying reticulocytes into the above-defined classes 1-4. Furthermore, the use of peripherals within the formula for maturity classification advantageously allows for differentiation between cells with a larger reticular area and fewer particles compared to cells with a larger reticular area and more particles. Thus, this novel method translates Heilmeyer's morphological findings into an automabilizable algorithm, providing for the first time a suitable quantitative method that enables highly accurate maturity classification of reticulocytes. In conclusion, while the present invention provides a quantitative analytical method, morphological methods are qualitative and therefore transmit a strong subjective bias, thus reducing their comparability.

[0055] In a particularly preferred embodiment, the maturity classification of reticulocytes is based on the use of the Λ / Γ value. This ratio allows for a combination of the values ​​of both proportions to be translated into specific values ​​obtained from Figure 3. Thus, the maturity classification involves assigning reticulocytes to Heilmeyer's class 1, 2, 3, or 4 according to the Λ / Γ ratio.

[0056] The Λ / Γ ratio can have different values, which are influenced by several factors such as the photodetector system used, the staining protocol used, the quality and age of the test cells, and any possible pretreatment steps. Such differences can be corrected by suitable calibration methods known to those skilled in the art. Calibration may include, for example, the use of a predetermined number of reticulocytes and standard staining conditions for analysis with various photodetector systems. In certain embodiments, commercially available calibration solutions may be used for calibration and reference. Examples of reasonably assumed calibration solutions include Cal-Chex, Cal-Chex A Plus, or Retic-Chex manufactured by Streck, Inc.

[0057] In a more preferred embodiment, the Λ / Γ ratio can be used to classify reticulocytes according to the following values:

[0058] A Λ / Γ ratio of approximately >2.5 indicates Class 1 reticulocytes, i.e., immature reticulocytes with a high density of reticular structure. Therefore, a value of approximately 2.5 constitutes a boundary value between Class 1 and Class 2, which exhibits a Λ / Γ ratio of less than approximately 2.5.

[0059] A Λ / Γ ratio of approximately 1 to 2.5 indicates class 2 reticulocytes, i.e., reticulocytes with a broad but loose reticular network. Therefore, a value of approximately 1 constitutes a boundary value between class 2 and class 3, which exhibits a Λ / Γ ratio of less than approximately 1.

[0060] A Λ / Γ ratio of approximately 0.35 to 1 indicates Class 3 reticulocytes, i.e., reticulocytes with a scattered network of reticular structures. Therefore, a value of approximately 0.35 constitutes a boundary value between Class 3 and Class 4, which exhibits a Λ / Γ ratio of less than approximately 0.35.

[0061] A Λ / Γ value of approximately <0.35 indicates class 4 reticulocytes, i.e., mature reticulocytes with scattered reticular granules.

[0062] The indicated boundary values ​​may vary slightly by a tolerance of ±25%, preferably ±15%, ±10%, more preferably ±5%, ±3%, ±2%, ±1%, or ±0.5%, due to, for example, tolerances of optical measurement of different analytical instruments, staining differences, or differences in reticular or cellular segmentation based on different programs or algorithms.

[0063] The present invention further envisions the provision of boundary classes between classes 1 and 2; 2 and 3; and 3 and 4. These boundaries may include reticulocytes whose classification into class 1 or 2; 2 or 3; or 3 or 4 is impossible or unclear due to the Λ / Γ values ​​corresponding to the boundary values ​​mentioned above. Boundary classes may be further established based on the application of the tolerance factors mentioned above to the class definitions provided above. It is further envisioned that the recombination of boundary classes with classes 1-4 can be performed using suitable calibration factors, for example, after calibration or comparative experiments and calculations, taking into account the optical and chemical differences in staining and detection of reticulocytes. Further information can be obtained from suitable literature sources such as Samuel M. Rapoport, 2019, The reticulocyte, 1st edition, CRC Press.

[0064] The present invention further relates to a method comprising the step of enumerating reticulocytes. As used herein, “enumerate” means counting and summing the number of reticulocytes per defined area, volume, time or other preferred unit, preferably per defined volume, e.g., per sample volume. In certain embodiments, the enumeration may be performed for each class of reticulocytes as defined herein. For example, all reticulocytes of class 1, 2, 3 and / or 4 in a sample may be counted. In further specific embodiments, the enumeration may further include counting non-reticulocytes, preferably red blood cells, in the sample. The corresponding numbers may be further compared to reticulocyte counts, class 1, 2, 3 or 4 reticulocyte counts and / or previous counting results, e.g., the same subject, different samples of the same subject, reference values ​​from a database, calibration criteria referred herein, reference values ​​from text or other literature sources, reference values ​​from independently determined healthy or diseased subjects. A good example of a suitable online source providing further details can be found at https: / / apps.who.int / iris / handle / 10665 / 61756 (last visited September 28, 2020).

[0065] In a more particular preferred embodiment, the present invention assumes the method defined above, wherein step (c), namely the determination of the parameters of each reticulocyte, namely (i) the ratio of the reticular area (Ar) to the total cell area (Ac) (Λ); and (ii) the ratio of the peripheral area (Ur) of the reticular region (Γ) to the reticular area (Ar) (Γ), is performed by an apparatus including an imaging module.

[0066] As used herein, the term “imaging module” refers to a unit capable of performing image processing procedures. Therefore, the present invention envisions obtaining images of reticulocytes or other cellular components present in a sample described herein, preferably stained reticulocytes present in the sample described herein. This image acquisition may further include preprocessing or scaling operations.

[0067] Furthermore, the present invention specifically envisions image processing of the acquired images. As used herein, “image processing” refers to general methods of converting an image into a digital format and performing calculations thereon to improve the image and / or extract useful or desired information. The output of image processing may be a modified image or characteristics or values ​​associated with the image.

[0068] In some embodiments, the imaging module according to the present invention is designed to enable image processing in conjunction with other modules, programs, databases, image repositories, or networks.

[0069] Image processing may include one or more of the following tasks, functions, or procedures: image enhancement, including brightness or contrast adjustment; wavelet and multiresolution processing, including image subdivision and pyramidal representation; compression, including techniques to reduce the storage bandwidth required to store an image or to transmit an image; morphological processing, including extraction of image components necessary for displaying or describing the shape, area information, or peripheral information of a cell or cellular component; segmentation, i.e., dividing an image into components or objects; displaying the data obtained in the segmentation process; extracting attributes from segmented data, including description, i.e., providing quantitative data that enables the distinction between one class of object and another; and object recognition, i.e., assigning labels to objects based on their description. These tasks, functions, or procedures can preferably be performed automatically or programmed, for example, based on the use of a suitable computer program or AI module. It is particularly preferable that the imaging module be designed to perform morphological segmentation operations, i.e., extraction of image components and their division into components or objects, such as reticular structures, reticular network structures, cell periphery, cell area, reticular area, reticular periphery, staining intensity within the reticular structure, and differences in staining intensity within the reticular structure.

[0070] In a more preferred embodiment of the present invention, the method described above includes, as an additional step (e), a morphological comparison of each stained reticulocyte with an image repository of reticulocytes independently classified by an expert. As used herein, “morphological comparison” refers to matching and contrasting cells with one or more features related to the reticulocytes, or more preferably intracellular portions, such as the presence of structures, shapes, forms, sizes, patterns, optical / visual displays, contrast, color, or staining density, particularly those related to the reticulocytes, or more preferably intracellular portions, in the sample, such as the reticular or reticular network. The comparison involves matching and contrasting images, preferably images processed with an imaging module as described herein, with images or data previously or alternatively classified by an expert into classes 1-4 according to Heilmeyer and stored in an image repository along with classification information. These stored images may further undergo image processing similar to or identical to the image processing performed on the images of reticulocytes obtained according to the method of the present invention.

[0071] The morphological comparison described above preferably involves applying images of stained reticulocytes to a machine learning-based method. The concept of “machine learning” as used in the context of the present invention typically relies on a two-stage method consisting of: first, a training phase; and second, a prediction phase. In the training phase, the values ​​of one or more parameters of a machine learning model (MLM) are set using training techniques and training data. In the prediction phase, the trained MLM operates based on measurement data. Exemplary parameters of an MLM include: the neuron weights in a given layer of an artificial neural network (ANN), such as a convolutional neural network (CNN); the kernel value of a classifier kernel, and so on.

[0072] The construction of an MLM may include a training phase in which parameter values ​​are determined. It is particularly preferable to perform the training using images from a reticulocyte image repository. As mentioned above, the images in the repository are advantageously independently classified by experts. The experts may be, for example, histologists or hematologists. In a further embodiment, the experts may be a group of experts who adjust or match subjective classifications. The corresponding results, i.e., the labeling of reticulocyte images to class 1, 2, 3, or 4, are then stored in the image repository, for example, along with the images. This information is retrievalable and can also be used by the MLM as a training set.

[0073] Building an MLM generally involves determining the values ​​of one or more hyperparameters. Typically, the values ​​of one or more hyperparameters in the MLM are set and remain unchanged during the training phase. Therefore, the values ​​of hyperparameters can be changed in the outer loop iterations; on the other hand, the values ​​of the MLM parameters can be changed in the inner loop iterations. Sometimes, there may be multiple training phases to allow testing or even optimization of multiple values ​​for one or more hyperparameters. The performance and accuracy of most MLMs are highly dependent on the values ​​of the hyperparameters.

[0074] Examples of hyperparameters include: the number of layers in a convolutional neural network; the kernel size of the classifier kernel; the input neurons of the ANN; the output neurons of the ANN; the number of neurons per layer; and the learning speed.

[0075] Various types and varieties of MLMs can be used in the context of this invention. For example, novelty detector MLMs / anomalousness detector MLMs, or classifier MLMs, such as binary classifiers, can be used. For example, deep learning (DL) MLMs can be used: here, the features detected by the DL MLM are not predefined but can be set by the values ​​of each parameter of the model that can be learned during training.

[0076] Several techniques can be used to build an MLM. For example, the type of training may differ depending on the type of MLM. Furthermore, the type of training used may differ in various implementations. For instance, iterative optimization may be used, which employs an optimization function defined for one or more error signals.

[0077] Next, the results of the morphological comparison calculation for each of the stained reticulocytes described above are compared with the classification results according to the Λ and Γ metric methods according to the present invention as defined above. If both classification methods agree, no specific tagging, warning, or finding is required. Consistency can be stored in certain embodiments as "metric classification confirmed by morphology," etc. If the results of the Λ and Γ metric methods and the morphological methods are inconsistent, the image of the reticulocyte where the difference was detected (e.g., live image or stored image) is tagged. This tagging may further provide the operator with a warning or message to further analyze the obtained results. Alternatively, both the metric and morphological analyses may be rerun to confirm the results. Tagging may further include internal classification according to the measurement difference. For example, if the Heilmeyer class defined by the metric method differs by 1 from the class defined by the morphological method, e.g., metric-defined class = 2, morphological-defined class = 1, or vice versa, an internal value of 1 (= degree of difference) is tagged. If the Heilmeyer class defined by the metric method differs from the class defined by the morphological method by more than 1, for example by 2, e.g., the metrically defined class = 3, the morphologically defined class = 1, or vice versa, an internal value of 2 (= degree of difference) is assigned to the tag. Further analysis and / or control calculations, which may include checks on equipment, optical devices, image repositories, etc., may be performed according to the degree of difference.

[0078] To branch classifications where the degree of difference is 1, the present invention assumes, in embodiments of a particular group, the selection of results obtained by metric. To branch classifications where the degree of difference is 2, the present invention assumes, in embodiments of a particular group, the selection of classifications between classes defined by metric and morphology. To branch classifications, the present invention further assumes the tagging of samples containing different results, and separate analyses of samples of the same subject by different analytical or preparation methods, for example, based on blood smears, or further, for example, parallel samples.

[0079] In a further embodiment, the present invention relates to a computer implementation method for maturity classification of reticulocytes from one or more images obtained from a whole blood sample as defined above. The method includes the steps of determining, within the image, the parameters of each reticulocyte: (i) the ratio of reticular area (Ar) to total cell area (Ac) (Λ); and (ii) the ratio of reticular periphery (Ur) to reticular area (Ar) (Γ); and the steps of maturity classification of the reticulocytes into four main classes according to the values ​​of Λ and Γ. In some embodiments, the method further includes the step of enumerating the reticulocytes, preferably as defined above herein. The method includes the image acquisition and image processing steps as defined above herein. In certain embodiments, morphological comparison of images with an image repository of reticulocytes as defined herein can also be implemented and performed. The method may be implemented on any suitable storage or computer platform, for example, cloud-based, internet-based, or intranet-based, or it may reside on a local computer or mobile phone, etc.

[0080] In a further embodiment, the present invention relates to a data processing apparatus including means for performing the computer implementation method defined above. The apparatus includes means for performing any one or more steps of the computer implementation method of the present invention as referred to above herein. Thus, any of the computer implementation methods described herein can be performed in whole or in part on a computer system including one or more processors that can be configured to perform the steps. Accordingly, some of these embodiments relate to a computer system configured to perform any of the steps of the computer implementation methods described herein, together with different components that potentially perform each step or each group of steps. The corresponding steps of the methods can further be performed simultaneously or in different orders. Furthermore, some of these steps may be used together with some of the other steps of other methods. Also, all or some of the steps may be optional. Furthermore, any of the steps of any of the methods may be performed by modules, circuits or other means for performing these steps.

[0081] When a program is executed by a computer, a computer program is also conceivable that includes instructions to cause the computer to perform any of the computer implementation methods of the present invention as defined herein, or any one or more computerizable steps of the methods of the present invention referred to herein.

[0082] The provision of computer-readable storage media containing the computer program products defined above is also envisioned. The computer-readable storage media may be linked to a server element, reside in a cloud structure, or be linked to one or more database structures or customer databases via the internet or intranet.

[0083] Any software component or computer program or function described herein can be implemented as software code executed by a processor using any suitable computer language, such as Java, Python, Javascript, VB.Net, C++, C#, C, Swift, Rust, Objective-C, Ruby, PHP, or Perl, for example, using prior art or object-oriented techniques. The software code can be stored as a series of instructions or commands in a computer-readable medium for storage and / or transmission, suitable of which include random access memory (RAM), read-only memory (ROM), magnetic media such as hard drives, optical media such as compact discs (CDs) or DVDs (digital multipurpose discs), and flash memory. The computer-readable medium may be any combination of such storage or transmission devices. Such programs can also be coded and transmitted using carrier signals adapted for transmission over wired, optical, and / or wireless networks following various protocols, including the Internet. As such, a computer-readable medium according to the present invention can be made using data signals coded in such a program. Computer-readable media encoded in program code may be packaged with compatible devices or provided separately from other devices (e.g., via download from the Internet). Any such computer-readable media may reside in or within a single computer program product (e.g., a hard drive, CD, or an entire computer system), or may reside in or within different computer program products within a system or network. A computer system may include a monitor, printer, or other suitable display that provides the user with any of the results referred to herein.

[0084] In further embodiments, the present invention relates to an in vitro method for monitoring and determining the health status of a subject and / or the subject's response to treatment. The method includes performing a method for maturity classification of reticulocytes from a whole blood sample of a subject as defined herein. The method includes the step of enumerating the reticulocytes for the entire sample and / or by class as defined herein. After classifying and optionally enumerating the reticulocytes in the sample, the obtained results may be compared to one or more reference values ​​or numbers. For example, the obtained number of reticulocytes per sample volume or per class per sample volume may be compared to the number obtained from a normal or healthy subject. Furthermore, these may be compared to the number obtained from a subject diagnosed with a specific disease, such as a disease affecting erythrogenesis, the blood cell cycle, or the overall number of blood cells, such as anemia or bone marrow disorders. In further embodiments, the numbers may be compared to reference numbers from databases, texts, literature sources, hospital documents, etc.

[0085] As used herein, the term “health condition” refers to the presence or absence of the disease in question, for example, compared to a healthy individual. In certain embodiments, the term may further relate to developmental trends in health, such as deterioration or improvement of a medical or disease state.

[0086] In certain embodiments, a method for monitoring or determining the health status of a subject may result in a total number of reticulocytes in a sample that is lower than the number of reticulocytes in a reference sample of a healthy subject, or a Class 1 reticulocyte count that is lower than the number of reticulocytes in a reference sample of a healthy subject of the same Class 1, but other classes show similar figures in the subject and reference samples; or a Class 2 reticulocyte count that is lower than the number of reticulocytes in a reference sample of a healthy subject of the same Class 2, but other classes show similar figures in the subject and reference samples; or a Class 3 reticulocyte count in a reference sample of a healthy subject of the same Class 3. The results provide a class 3 reticulocyte count that is lower than the reticulocyte count in a reference sample of a healthy subject of the same class 4, while other classes show similar figures in the test and reference samples; or the results provide a class 4 reticulocyte count that is lower than the reticulocyte count in a reference sample of a healthy subject of the same class 4, while other classes show similar figures in the test and reference samples; or the results provide a class 1 and 2, class 2 and 3, class 3 and 4, class 1, 2 and 3 or class 2, 3 and 4 reticulocyte count that is lower than the reticulocyte count in a reference sample of a healthy subject of the corresponding class, while other classes show similar figures in the test and reference samples. These results can be interpreted as indicating, for example, bone marrow failure due to drugs, tumors, radiation therapy or infection; cirrhosis; anemia that may result from low iron levels or low vitamin B12 or folic acid levels; or chronic kidney disease. These diseases are understood as “hyporeticulosis” within the context of this invention.

[0087] In certain embodiments, a method for monitoring or determining the health status of a subject may result in a total number of reticulocytes in a sample that is higher than the number of reticulocytes in a reference sample of a healthy subject, or may result in a Class 1 reticulocyte count that is higher than the number of reticulocytes in a reference sample of a healthy subject of the same Class 1, but other classes show similar figures in the subject and reference samples; or may result in a Class 2 reticulocyte count that is higher than the number of reticulocytes in a reference sample of a healthy subject of the same Class 2, but other classes show similar figures in the subject and reference samples; or may result in a Class 3 reticulocyte count in a reference sample of a healthy subject of the same Class 3 The results provide a reticulocyte count of Class 3 that exceeds the reticulocyte count, while other classes show similar figures in the test and reference samples; or the results provide a reticulocyte count of Class 4 that exceeds the reticulocyte count in a reference sample of a healthy subject of the same Class 4, while other classes show similar figures in the test and reference samples; or the results provide a reticulocyte count of Class 1 and 2, Class 2 and 3, Class 3 and 4, Class 1, 2 and 3 or Class 2, 3 and 4 that exceeds the reticulocyte count in a reference sample of a healthy subject of the corresponding class, while other classifications show similar figures in the test and reference samples. These results can be interpreted as indicating anemia (i.e., hemolytic anemia) which may be caused by red blood cells being destroyed more quickly than usual; bleeding; fetal or neonatal blood disorders (fetal erythroblastosis); or renal disease due to increased erythropoietin production. These disorders are understood as “hyperreticulosis” within the context of this invention.

[0088] As used herein, the term “response to treatment” refers to a positive or negative response of a subject to treatment for a disease that may affect erythrocyte production or may be related to any of the diseases mentioned above or any other disease diagnosed with the same impairment.

[0089] A method envisioned for monitoring and determining the health status of a subject and / or its response to a treatment involves, in certain embodiments, performing the method of the present invention with one or more samples taken from the subject. For example, a sample may be taken at an initial point in time, and then further samples may be taken at a certain period of time. The period may be any period deemed suitable to those skilled in the art. The period may be determined by the disease or treatment of the subject, its hospitalization status, the subject's health status, or any other factors relating to the diagnosis. In certain embodiments, further samples are taken from the subject after a period of 2, 3, 4, 5, 6, 7 days or more. Ten days, 14 days, 3 weeks, 1 month, 2 months, 3 months, 4 months, 5 months, 6 months or more are also envisioned.

[0090] Comparing enumeration and classification procedures for samples taken from subjects at different time points may result in similar numbers and / or class distributions of reticulocytes, an overall increase in reticulocyte counts or within one or more classes, or a overall decrease in reticulocyte counts or within one or more classes. The corresponding results can then be used to draw diagnostic conclusions.

[0091] For example, when comparing initial and subsequent samples, an overall increase in class 2, 3, or 4 reticulocyte counts indicates a positive response to improvement in the subject's health and / or treatment in cases of hyporetic redness such as, for example, bone marrow failure due to drugs, tumors, radiation therapy, or infection; cirrhosis; anemia that may result from low iron levels or low vitamin B12 or folate levels; or chronic kidney disease.

[0092] Similarly, when comparing the initial and subsequent samples, an overall increase in the reticulocyte count, or in classes 2, 3, or 4, indicates deterioration of the subject's health and / or a negative response to treatment in cases of hyperreticulocytes such as hemolytic anemia; bleeding; fetal or neonatal hematological disorders (fetal erythroblastosis); or renal disease due to increased erythropoietin production.

[0093] In a further embodiment, when comparing the initial and subsequently collected samples, a decrease in the overall or class 2, 3, or 4 reticulocyte count indicates a positive response to improvement in the health and / or treatment of a subject in cases of hyperreticulosis such as hemolytic anemia; bleeding; fetal or neonatal hematological disorders (fetal erythroblastosis); or renal disease due to increased erythropoietin production.

[0094] In another embodiment, when comparing initially and subsequently collected samples, a general decrease in class 2, 3, or 4 reticulocyte counts indicates deterioration of the subject's health and / or a negative response to treatment in cases of hyporetic red blood cell disease, such as bone marrow failure due to drugs, tumors, radiation therapy, or infection; cirrhosis; anemia that may result from low iron levels or low vitamin B12 or folate levels; or chronic kidney disease. Further details can be obtained from suitable literature or internet sources, such as https: / / www.statpearls.com / kb / viewarticle / 28438 (last visited September 28, 2020).

[0095] In further embodiments, the present invention relates to an apparatus comprising means for carrying out the methods according to the present invention. The apparatus may therefore include a module capable of performing the staining operations defined above herein. The staining module may include elements for cell preparation, cell sorting, cell purification, chemical processing, staining, washing, etc., and optionally storage functions. The module may consist, for example, of a robotic body or function. The apparatus may further include a module capable of illuminating the stained sample. This module may take the form of a microscope or microscope system and may include a focusing optical element that can be designed as a lens, and / or a diaphragm. In certain embodiments, the microscope system may further include a system including a flow cytometer or flow cytometry function, or may be built essentially on these, particularly when using fluorescent dyes. The microscope module may further be coupled to an evaluation module capable of determining the parameters Λ and Γ described herein within reticulocytes. This evaluation module may, for example, be implemented or linked to the imaging module defined above herein, i.e., a unit capable of performing image processing procedures on acquired images of reticulocytes or other cellular components present in the sample described herein, preferably stained reticulocytes present in the sample described herein. The existence of a maturity classification module capable of assigning measured values ​​of the reticulocyte parameters Λ and Γ to Heilmeyer classes 1 to 4 is also envisioned. This module may, for example, be a computer-based module or an AI module or a neural network, computer system, computer network or interface, and may optionally be linked to a database, image repository or laboratory or hospital system. [Examples]

[0096] Detection of reticulocytes Materials and methods for detecting reticulocytes Samples. Peripheral blood was collected from healthy donors using a procedure approved by informed consent and application 316_14B of the Ethikkommission der Universitat Erlangen. Blood from each sample was collected in a 4.7 ml EDTA-coated tube. All samples were processed within 6 hours of collection.

[0097] Immature red blood cells were identified by using a hyperbiotic dye that precipitates cytoplasmic RNA into a reticular network. Reticulocytes were treated with basic neomethylene blue (NMB, c=0.5%, Sigma-Aldrich) dye. The dye solution was added to whole blood for 1 minute. Blood smears were prepared, and this filamentous network was visualized under a light microscope.

[0098] Optical configuration: Reticulocytes were identified by recording bright-field images in transmission mode using a camera (Apex 3-CMOS prism-based RGB camera, JAI) integrated into a DMi8 Leica inverted microscope and a 40x objective lens (HP PL APO 40× / 0.95 CORR PH2, Leica).

[0099] Statistical analysis. Images were analyzed using ImageJ. Segmentation of the stained reticular formation was achieved by converting the color images to RGB-grayscale images. An automated thresholding algorithm was applied to the green channel to determine the area and periphery of the reticular formation.

[0100] The following drawings (Figures 1, 2, and 3) are provided for illustrative purposes only. It should be understood that the drawings are not intended to be restrictive. Those skilled in the art can clearly envision further modifications of the principles described herein.

Claims

1. A method for classifying reticulocytes from whole blood samples by maturity. (a) A step of staining the sample with an agglutination staining reagent selected from a hyperbiotic agglutination staining reagent and a fluorescent agglutination staining reagent, which binds to nucleic acids containing intracellular structures and causes them to agglutinate; (b) A step of illuminating the stained sample with a light beam to detect reticulocytes; (c) A step of determining the parameters of each reticulocyte: (i) the ratio of the reticular area (Ar) to the total cell area (Ac) (Λ); and (ii) the ratio of the peripheral length of the reticular area (Ur) to the reticular area (Ar) (Γ); and (d) The process of classifying reticulocytes into one of four major maturity classes according to the values ​​determined for Λ and Γ. The method, including the method described above.

2. The method according to claim 1, further comprising the step of listing reticulocytes as a final step.

3. The method according to claim 1 or 2, wherein the staining is performed using a superbiotic agglutination staining reagent selected from NMB (new methylene blue), brilliant cresyl blue, crystal violet, methyl violet, and Nile blue.

4. The method according to claim 1 or 2, wherein the staining is performed with a fluorescent agglutinating dye selected from acridine orange, auramine O, ethidine bromide, pyronin Y, thioflavin-T, and thiazole orange.

5. Step (c) is performed in an apparatus including an imaging module, according to any one of claims 1 to 4.

6. The method according to any one of claims 1 to 5, further comprising, as step (e), a step of morphologically comparing each stained reticulocyte with an image repository of reticulocytes independently classified by an expert.

7. The method according to claim 6, wherein the morphological comparison comprises applying images of stained reticulocytes to a machine learning-based method trained on images from the image repository of reticulocytes.

8. The method according to any one of claims 1 to 7, wherein the maturity classification comprises assigning reticulocytes to one of four major maturity classes according to the ratio of Λ / Γ of a value determined for Λ / Γ.

9. The method according to claim 8, wherein a Λ / Γ value >2.5 indicates maturity class 1, a Λ / Γ value between 1 and 2.5 indicates maturity class 2, a Λ / Γ value between 0.35 and 1 indicates maturity class 3, and a Λ / Γ value <0.35 indicates maturity class 4.

10. A computer-implemented method for classifying reticulocytes by maturity from one or more images obtained from a whole blood sample, comprising the steps of: determining the parameters of each reticulocyte in the image: (i) the ratio of the reticular area (Ar) to the total cell area (Ac) (Λ); and (ii) the ratio of the peripheral length of the reticular structure (Ur) to the reticular area (Ar) (Γ); and classifying the reticulocytes by maturity into four main classes according to the values ​​of Λ and Γ.

11. An in vitro method for measuring an increase or decrease in the number of reticulocytes, comprising performing the method according to any one of claims 1 to 10 with one or more whole blood samples obtained from a subject.

12. The in vitro method according to claim 11, wherein an initial sample is compared with a second or subsequent sample taken from the subject after a period of 2, 3, 4, 5, 6, 7 days or more, and the increase or decrease in the number of reticulocytes in each maturity classification is measured.

13. An apparatus comprising means for performing the method according to any one of claims 1 to 9, 11, or 12.

14. A data processing device comprising means for performing the method described in claim 10.

15. A computer program that, when executed by a computer, includes instructions causing the computer to perform the method according to any one of claims 6, 7, or 10.

16. The method according to claim 3 or 4, wherein the staining further comprises the step of chemically crosslinking nucleic acids with nitrogen mustard, cis-diaminedichloroplatinum(II) or a derivative, or chloroethylnitrosourea (CENU).

Images