A model and system for assisting in diagnosing chronic myelocytic leukemia blast crisis

Abstract

The application relates to a model and system for assisting in diagnosing chronic myelocytic leukemia acute lymphoid leukemia. It belongs to the technical field of biological tests. The application combines FISH technology and DAPI staining, designs a double-reading standard of "nuclear morphology + gene signal", can quickly, accurately and standardize the distribution characteristics of BCR::ABL1 fusion genes in different cell types, solves the problem that the existing FISH technology cannot directly reflect the cell differentiation stage at the gene detection level, causes the lag of chronic myelocytic leukemia acute lymphoid leukemia diagnosis or needs additional cumbersome flow cytometry, and can provide similar immunotyping identification information without increasing the cost of reagents, improve the diagnosis efficiency and reduce the screening cost of patients.

Description

Technical Field

[0001] This application belongs to the field of biological testing technology and relates to a model and system for assisting in the diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia. Background Technology

[0002] Chronic myeloid leukemia (CML) is a malignant clonal disease originating from hematopoietic stem cells. Its characteristic genetic marker is the Philadelphia chromosome (Ph chromosome), specifically the BCR::ABL1 fusion gene formed by the t(9;22)(q34;q11) translocation. CML patients may experience blast crisis during disease progression, with lymphoid blast crisis being one of the most common types. CML lymphoid blast crisis (CML-BP-ALL) is the terminal stage of CML, characterized by a proportion of ≥20% primitive B lymphocytes in the bone marrow and / or peripheral blood (the 5th-WHO criteria consider a proportion of less than 10% primitive lymphocytes as diagnostic criteria for CML lymphoid blast crisis).

[0003] Acute B-lymphoblastic leukemia (B-ALL) with BCR::ABL1, also known as Philadelphia chromosome-positive B-ALL (Ph+B-ALL), is a specific molecular subtype of B-ALL. Its incidence in childhood B-ALL is approximately 2%-5%, while in adult B-ALL it accounts for about 25%-30%. The most common fusion protein formed by the Philadelphia chromosome is p190 (approximately 70%), followed by p210 (approximately 30%), with a very small number of cases showing double positivity.

[0004] Compared to Ph+B-ALL, patients with acute lymphoblastic leukemia (CML) have a higher relapse rate, are more prone to tyrosine kinase inhibitor (TKI) resistance, and have a worse overall prognosis. Therefore, accurate differentiation between the two is crucial for prognostic assessment and treatment strategies, but it is often challenging in clinical laboratory work.

[0005] Currently, the detection of the BCR::ABL1 fusion gene mainly relies on methods such as fluorescence in situ hybridization (FISH), polymerase chain reaction (PCR), and chromosome karyotype analysis. For example, CN103409505A discloses a FISH probe for detecting the BCR / ABL fusion gene without repetitive sequences. Using the BCR and ABL genes as templates, the non-repetitive sequences in the BCR and ABL genes are subjected to polymerase chain reaction, and the amplification products are DNA fragments of varying sizes ranging from 350 to 800 base pairs. The repetitive sequences in the BCR and ABL genes are eliminated, thus forming a product without repetitive sequences. After fluorescent labeling, the product is used to obtain a fluorescence in situ hybridization probe for detecting the BCR / ABL fusion gene without repetitive sequences. CN117230195A discloses a composition for detecting the BCR-ABL1 P210 fusion gene. The composition includes a pair of BCR-ABL1 P210 fusion gene detection primers and a BCR-ABL1 P210 fusion gene detection probe, as well as a pair of internal reference gene ABL1 detection primers and an internal reference gene ABL1 detection probe. Both ends of the BCR-ABL1 P210 fusion gene detection probe and the ABL1 detection probe are labeled with a fluorescent group and a quenching group, respectively.

[0006] However, existing technologies still have shortcomings in distinguishing BCR::ABL1 fusion genes and their status in different cell types, as well as in making comprehensive judgments based on cell morphology characteristics. They cannot achieve accurate detection of BCR::ABL1 fusion genes in specific cell populations. Fusion gene FISH tests mainly focus on the presence of fusion signals (positive / negative) in the cell nucleus, but the positive rate of fusion genes alone cannot distinguish whether a patient is in the "CML blast crisis (especially blast crisis)" or primary Ph+B-ALL. Diagnosis usually requires combined flow cytometry for immunophenotyping, which increases testing costs and sample consumption. Summary of the Invention

[0007] In view of the shortcomings of existing technologies and practical needs, this application provides a model and system for assisting in the diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia, with the aim of achieving rapid and low-cost auxiliary diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia.

[0008] To achieve the above objectives, this application adopts the following technical solution: In a first aspect, this application provides a model for the auxiliary diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia, the method for constructing the model comprising: FISH and DAPI staining images of samples from patients with acute lymphoblastic leukemia of chronic myeloid leukemia were acquired using fluorescence microscopy. Mononuclear cells, mature granulocytes, and FISH signals were identified in the images. The criteria for identifying mononuclear cells were round nuclei, and the criteria for identifying mature granulocytes were lobed or rod-shaped nuclei. The FISH signals included BCR gene signals, ABL1 gene signals, and BCR::ABL1 fusion gene signals. The dataset was divided into a test set and a training set to train the machine learning model, resulting in a model capable of classifying mononuclear cells, mature granulocytes, and FISH signals from FISH and DAPI stained images.

[0009] This application designs and trains a model to assist in the diagnosis of acute lymphoblastic leukemia (ALL). Combining FISH and DAPI staining images, a method is designed to rapidly identify mononuclear cells and mature granulocytes using the morphological features of DAPI staining, obtaining relevant classification information. This information is used as a dataset to train a machine learning model, resulting in a model capable of classifying mononuclear cells, mature granulocytes, and FISH signals in FISH and DAPI staining images. This classification information can be used as intermediate data to assist in calculating and determining the risk of ALL in chronic myeloid leukemia.

[0010] Optionally, the machine learning model may include any one of random forest, convolutional neural network, or K-nearest neighbor.

[0011] Secondly, this application provides a cell and FISH signal classification method, wherein the method uses the model for assisting in the diagnosis of acute lymphoblastic leukemia of chronic myeloid leukemia described in the first aspect to identify FISH and DAPI staining images, and classifies FISH and DAPI staining images into mononuclear cells, mature granulocytes and FISH signals.

[0012] The cell and FISH signal classification method of this application can be used to process and obtain intermediate data information for subsequent calculation and judgment.

[0013] Thirdly, this application provides a system for assisting in the diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia, the system comprising an image acquisition module, an image recognition module, and a calculation and determination module; The image acquisition module is used to perform the following: acquiring FISH and DAPI staining images of samples from patients with chronic myeloid leukemia; The image recognition module is used to perform the following: FISH recognition of mononuclear cells, mature granulocytes and FISH signal classification in FISH and DAPI staining images. The criteria for determining mononuclear cells are that the nucleus is round or oval, and the criteria for determining mature granulocytes are that the nucleus is lobed or rod-shaped. The FISH signal includes BCR gene signal, ABL1 gene signal and BCR::ABL1 fusion gene signal. The calculation and determination module is used to perform the following: calculating the proportion of BCR::ABL1 fusion gene signal positive cells in mononuclear cells and the proportion of BCR::ABL1 fusion gene signal positive cells in mature granulocytes; if the proportion of BCR::ABL1 fusion gene signal positive cells in mononuclear cells is ≥2% and the proportion of BCR::ABL1 fusion gene signal positive cells in mature granulocytes is ≥2%, then a high risk of acute lymphoblastic leukemia (ALL) is output (or a positive result for ALL); otherwise, a low risk of ALL is output (or a negative result for ALL).

[0014] Optionally, if only a single nucleated cell contains BCR::ABL1 fusion gene signal positive cells, and the proportion of BCR::ABL1 fusion gene signal positive cells among them is ≥2%, a high risk of Ph+ acute lymphoblastic leukemia can be output (or a positive result for Ph+ acute lymphoblastic leukemia can be indicated).

[0015] Optionally, the image recognition unit is configured to recognize FISH and DAPI stained images using the model for assisting in the diagnosis of acute lymphoblastic leukemia as described in the first aspect.

[0016] This application combines FISH and DAPI staining. By utilizing the differences in nuclear morphology (such as the difference between mononuclear cell morphology and mature segmented / band nuclei) displayed by DAPI staining, mononuclear cells and mature granulocytes can be quickly and intuitively distinguished. Cell classification can be performed directly at the gene detection level. By analyzing the positivity of BCR::ABL1 in cells with specific morphologies (such as primitive / immature mononuclear cells), the characteristic cell populations of the blast crisis can be captured more accurately, greatly improving the specificity and sensitivity of the diagnosis of "chronic myeloid leukemia blast crisis".

[0017] This application eliminates the need for additional fluorescence channels or special reagents. Utilizing only the DAPI channel, routinely used for counterstaining cell nuclei in FISH experiments, it provides information on cell differentiation stages through morphological interpretation, achieving "one detection, double benefit" (genetic information + morphological information). This simplifies the process and reduces clinical testing costs. Furthermore, it establishes an interpretation system based on the "morphology-signal" correspondence, reducing errors from relying solely on subjective experience to judge blast crisis phases, thus making FISH test results more clinically relevant.

[0018] Optionally, the sample may include a bone marrow sample or a peripheral blood sample.

[0019] Optionally, the method for obtaining the FISH and DAPI staining images includes: Samples were taken from patients with chronic myeloid leukemia, leukocyte suspensions were prepared, and FISH hybridization and DAPI staining were performed. FISH and DAPI staining images were obtained using a fluorescence microscope with graded filters.

[0020] Optionally, the FISH hybridization includes FISH hybridization using a two-color dual-fusion probe.

[0021] Optionally, the DAPI staining conditions include: a final DAPI concentration of 0.5-1.0 μg / mL, and incubation at 20-30°C in the dark for 5-10 min.

[0022] Fourthly, this application provides a computer device, which includes a memory and a processor. The memory stores a computer program, and when the computer program is executed by the processor, it performs the functions of the system for assisting in the diagnosis of acute lymphoblastic leukemia as described in the third aspect.

[0023] Fifthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the functions of the system for assisting in the diagnosis of acute lymphoblastic leukemia as described in the third aspect.

[0024] Compared with the prior art, this application has at least the following beneficial effects: This application utilizes FISH technology combined with DAPI staining and further designs an interpretation standard based on a dual dimension of "nuclear morphology + gene signal" to achieve rapid, accurate, and standardized identification of the distribution characteristics of the BCR::ABL1 fusion gene in different cell types. This addresses the problem that existing FISH technologies cannot directly reflect cell differentiation stages at the gene detection level, leading to delayed diagnosis of acute lymphoblastic leukemia (ALL) or reliance on additional cumbersome flow cytometry. Furthermore, traditional diagnosis of ALL usually requires bone marrow aspiration for flow cytometry immunophenotyping, which is time-consuming and costly. This application, without increasing reagent costs (using only the conventional dye DAPI), provides similar immunophenotyping-like identification information through morphological analysis, thereby improving diagnostic efficiency and reducing patient screening costs. Attached Figure Description

[0025] Figure 1 Image showing negative results for the BCR::ABL1 fusion gene in mononuclear cells; Figure 2Image showing positive results for the BCR::ABL1 fusion gene in mononuclear cells; Figure 3 Image showing the negative results of the BCR::ABL1 fusion gene in mature granulocytes (segmented / band nuclei); Figure 4 Image showing the positive results of the BCR::ABL1 fusion gene in mature granulocytes (segmented / band nuclei). Detailed Implementation

[0026] The technical solution of this application will be further described below with reference to the accompanying drawings and specific embodiments. However, the examples below are merely simplified examples of this application and do not represent or limit the scope of protection of this application. The scope of protection of this application shall be determined by the claims.

[0027] Where specific techniques or conditions are not specified in the examples, they shall be performed in accordance with the techniques or conditions described in the literature in this field, or in accordance with the product instructions. Reagents or instruments whose manufacturers are not specified are all conventional products that can be purchased from legitimate channels.

[0028] This application designs an analytical scheme based on FISH technology combined with DAPI staining. By combining the nuclear morphology displayed by DAPI staining (distinguishing between mononuclear cells and mature granulocytes (segmented / band cells)) with the BCR::ABL1 fusion gene detection results, a detection system capable of assisting in the differential diagnosis of whether chronic myeloid leukemia (CML) has undergone blast crisis is established. This aims to address the problem that existing FISH technology cannot directly reflect the cell differentiation stage at the gene detection level, thus leading to delayed diagnosis of CML blast crisis or reliance on additional cumbersome flow cytometry.

[0029] Example 1 This embodiment constructs a model for the auxiliary diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia. The method for constructing the model includes: FISH and DAPI staining images of samples from patients with acute lymphoblastic leukemia (ALL) were acquired using fluorescence microscopy. Mononuclear cells, mature granulocytes, and FISH signals were identified and classified in the images. Mononuclear cells were defined as having round or oval nuclei, and mature granulocytes as having lobulated or rod-shaped nuclei. FISH signals included BCR gene signals, ABL1 gene signals, and BCR::ABL1 fusion gene signals. The samples were divided into test and training sets to train a machine learning model, resulting in a model capable of classifying mononuclear cells, mature granulocytes, and FISH signals from FISH and DAPI staining images.

[0030] Specifically, the following steps are included: Step 1: Sample Preparation 1. Collect bone marrow or peripheral blood samples from the patient; 2. Prepare a leukocyte suspension; The specific experimental procedure includes: (1) Take 2 ml of bone marrow or peripheral blood sample from the patient within 7 days and put it into a 10 ml centrifuge tube; (2) Centrifuge at room temperature, 2200 rpm, for 7 minutes; (3) After centrifugation, discard the supernatant of the sample, add 0.075 M potassium chloride to 10 mL, and mix by pipetting; (4) Water bath at 37℃ for 40 min; (5) Take out the sample after water bath, add Carnoy's fixative to 10 ml, mix by pipetting, and centrifuge at room temperature at 2200 rpm for 7 minutes; (6) Remove the centrifuged sample, discard the supernatant, add Carnoy's fixative to 10 mL, mix well by pipetting, and centrifuge at room temperature at 2200 rpm for 7 minutes. Repeat this operation 3 times; (7) After centrifugation, the supernatant is completely discarded, and the cell suspension is prepared and then dropped to make cell droplets.

[0031] Step 2: FISH hybridization 1. FISH hybridization was performed using a specific BCR::ABL1 fusion gene dual-color dual-fusion probe (Wuhan Kanglu Biotechnology Co., Ltd., product code: FP-003, product name: BCR / ABL fusion gene detection kit (fluorescence in situ hybridization)); The specific experimental procedure includes: (1) The prepared cell droplets were placed in 2×SSC, 75% ethanol, 85% ethanol and anhydrous ethanol for 3 minutes each, and then dried at room temperature. (2) Remove the probe from the frozen condition and allow it to return to room temperature; (3) Place the probe on the sample area, cover with a coverslip and seal the edges with rubber; (4) Place the slides on the hybridization apparatus, start the program to denature at 85℃ for 5 minutes, and hybridize at 42℃ for 2-16 hours; 2. Fluorescent labeling of probe gene sites: BCR gene is labeled with green fluorescence, and ABL1 gene is labeled with red fluorescence; Step 3: DAPI staining 1. Wash after hybridization to remove unbound probes; 2. Stain the cell nuclei with DAPI dye; 3. The final DAPI concentration is 0.5-1.0 μg / mL, and the mixture is incubated at room temperature in the dark for 5-10 minutes. The specific experimental procedure includes: (1) Remove the coverslip and wash with 2×SSC / 0.1% NP-40 washing solution at 73°C for 2 minutes, then wash with 2×SSC at room temperature for 2 minutes to remove non-specific binding; (2) After drying in the dark, add DAPI-containing anti-quenching sealing agent and cover with a coverslip; (3) Observe the results under a fluorescence microscope and collect images for analysis.

[0032] Step 4: Microscopic observation and analysis 1. Observe using a fluorescence microscope under appropriate filters ([C-3010-001-MS]DAPI filter, [C-3010-002-MS]GREEN filter, [C-3010-003-MS]ORANGE filter); 2. Simultaneously observe the nuclear morphology and FISH signal of cells stained with DAPI; 3. Differentiate cell types based on nuclear morphology: Mononuclear cells: The nucleus is round; Mature granulocytes: The nucleus is lobed or rod-shaped; 4. Record the BCR::ABL1 fusion gene and its status: Positive: A two-color, two-fusion signal appears (red and green merge into yellow); Negative: Red and green signals separated.

[0033] The specific training process includes: FISH and DAPI staining images of samples from patients with acute lymphoblastic leukemia (ALL) were collected using the method described above. Mononuclear cells, mature granulocytes, and FISH signals were identified and classified in the images. The FISH signals included BCR gene signals, ABL1 gene signals, and BCR::ABL1 fusion gene signals. The obtained data were divided into test and training sets to train the machine learning model. Five-fold cross-validation was used to evaluate the model's reliability. The machine learning model could use any of the following: random forest, convolutional neural network (such as RetinaNet), or K-nearest neighbor.

[0034] Example 2 This embodiment constructs a system for assisting in the diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia. The system includes an image acquisition module, an image recognition module, and a calculation and judgment module. The image acquisition module is used to perform the following: acquiring FISH and DAPI staining images of samples from patients with acute lymphoblastic leukemia of chronic myeloid leukemia; The image recognition module is used to perform the following: identifying mononuclear cells, mature granulocytes, and classifying FISH signals in FISH and DAPI stained images. The criteria for identifying mononuclear cells are that the nucleus is round, and the criteria for identifying mature granulocytes are that the nucleus is lobed or rod-shaped. The FISH signals include BCR gene signals, ABL1 gene signals, and BCR::ABL1 fusion gene signals. The calculation and determination module is used to perform the following: calculating the proportion of BCR::ABL1 fusion gene signal positive cells in mononuclear cells and the proportion of BCR::ABL1 fusion gene signal positive cells in mature granulocytes; if the proportion of BCR::ABL1 fusion gene signal positive cells in mononuclear cells is ≥2% and the proportion of BCR::ABL1 fusion gene signal positive cells in mature granulocytes is ≥2%, then a high risk of acute lymphoblastic leukemia (ALL) is output (or a positive result for ALL); otherwise, a low risk of ALL is output (or a negative result for ALL).

[0035] Specifically, the method for obtaining FISH and DAPI staining images of samples from patients with acute lymphoblastic leukemia of chronic myeloid leukemia is as described in Example 1. Image recognition can be performed by manual visual recognition or by using the model constructed in Example 1.

[0036] Example 3 This embodiment verifies the system for assisting in the diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia constructed in Example 2.

[0037] Peripheral blood samples (sample numbers 1-21) were collected from 21 patients with acute lymphoblastic leukemia (ALL) who had undergone clinical evaluation, bone marrow samples (sample numbers 22-32) were collected from 11 patients with Ph+ acute lymphoblastic leukemia, and 3 samples (sample numbers 33-35) were negative for the BCR::ABL1 fusion gene. FISH hybridization and DAPI staining images were obtained according to the method in Example 1. Representative results are shown below. Figures 1-4 As shown, the system for assisting in the diagnosis of acute lymphoblastic leukemia of chronic myeloid leukemia constructed in Example 2 was used for diagnosis.

[0038] The specific results are as follows: Specimens numbered 1-21 showed BCR::ABL1 fusion gene signal-positive cells in both mononuclear cells and mature granulocytes, with the proportion of BCR::ABL1 fusion gene signal-positive cells in both mononuclear cells and mature granulocytes being ≥2%, meeting the FISH diagnostic criteria for acute lymphoblastic leukemia (ALL). Specimens numbered 22-32 showed BCR::ABL1 fusion gene signal-positive cells only in mononuclear cells, with the proportion of BCR::ABL1 fusion gene signal-positive cells being ≥2%, meeting the FISH diagnostic criteria for Ph+ acute lymphoblastic leukemia. Specimens numbered 33-35 detected both mononuclear cells and mature granulocytes, but the proportion of BCR::ABL1 fusion gene signal-positive cells was 0%, meeting the FISH diagnostic criteria for a negative BCR::ABL1 fusion gene.

[0039] The prediction results of this application system were compared with clinical results, as shown in Table 1. ABL::BCR fusion gene positivity includes the following situations: Positive ①: ABL::BCR fusion gene positivity, with the proportion of fusion gene-positive cells in mononuclear cells and mature granulocytes ≥2%, consistent with acute lymphoblastic leukemia of chronic myeloid leukemia; Positive ②: Only mononuclear cells are positive for ABL::BCR fusion gene, and the proportion of fusion gene-positive cells is ≥2%, consistent with patients with Ph+ acute lymphoblastic leukemia; The detection accuracy, sensitivity, and specificity of this application system were calculated.

[0040] Table 1 In this test, the specificity of the test is defined as the concordance rate of negative results. Among the 35 samples completed, 3 were negative. The test results of the diagnostic system of this application are consistent with the known clinical results, all of which are negative. The concordance rate of FISH test = 3 / 3 × 100% = 100%.

[0041] In this test, the detection sensitivity is defined as the concordance rate of positive results. Among the 35 completed samples, 11 cases were positive for Ph+ acute lymphoblastic leukemia, and the detection results of the diagnostic system of this application were consistent with the known clinical results, all showing positive results; 21 cases were positive for acute lymphoblastic leukemia of chronic myeloid leukemia, and the detection results of the diagnostic system of this application were consistent with the known clinical results, all showing positive results. Detection sensitivity = 32 / 32 × 100% = 100%.

[0042] The detection results of the system in this application are consistent with the clinical results, and the specificity and sensitivity are both 100%, indicating that the system in this application can be effectively used to assist in the diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia.

[0043] Example 4 This embodiment provides an electronic device, a computer device, which includes a memory and a processor. The memory stores a computer program, and when the computer program is executed by the processor, it implements the function of the system for assisting in the diagnosis of acute lymphoblastic leukemia of chronic myeloid leukemia described in Embodiment 2.

[0044] Those skilled in the art will understand that the device of this application can be obtained using various forms of hardware, software, firmware, dedicated processors, or combinations thereof.

[0045] Example 5 This embodiment provides a computer-readable storage medium storing a computer program. The computer program has program code, which, when run in a corresponding processor, controller, computing device, or terminal, implements the function of the system for assisting in the diagnosis of acute lymphoblastic leukemia of chronic myeloid leukemia as described in Embodiment 2.

[0046] In summary, this application, based on FISH technology combined with DAPI staining, and further designed an interpretation standard based on the dual dimensions of "nuclear morphology + gene signal," enables rapid, accurate, and standardized identification of the distribution characteristics of the BCR::ABL1 fusion gene in different cell types. This addresses the problem that existing FISH technologies cannot directly reflect cell differentiation stages at the gene detection level, leading to delayed diagnosis of acute lymphoblastic leukemia (ALL) or reliance on additional cumbersome flow cytometry. Furthermore, traditional diagnosis of ALL typically requires bone marrow aspiration for flow cytometry immunophenotyping, which is time-consuming and costly. This application, without increasing reagent costs (using only the conventional dye DAPI), provides similar immunophenotyping-like identification information through morphological analysis, thereby improving diagnostic efficiency and reducing patient screening costs.

[0047] The applicant declares that the above description is only a specific implementation of this application, but the protection scope of this application is not limited thereto. Those skilled in the art should understand that any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application fall within the protection and disclosure scope of this application.

Claims

1. A model for the auxiliary diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia, characterized in that, The method for constructing the model includes: FISH and DAPI staining images of samples from patients with acute lymphoblastic leukemia of chronic myeloid leukemia were acquired using fluorescence microscopy. Mononuclear cells, mature granulocytes, and FISH signals were identified in the images. The criteria for identifying mononuclear cells were round nuclei, and the criteria for identifying mature granulocytes were lobed or rod-shaped nuclei. The FISH signals included BCR gene signals, ABL1 gene signals, and BCR::ABL1 fusion gene signals. The dataset was divided into a test set and a training set to train the machine learning model, resulting in a model capable of classifying mononuclear cells, mature granulocytes, and FISH signals from FISH and DAPI stained images.

2. The model for auxiliary diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia according to claim 1, characterized in that, The machine learning model includes any one of random forest, convolutional neural network, or K-nearest neighbor.

3. A method for classifying cells and FISH signals, characterized in that, The method utilizes the model for assisting the diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia as described in claim 1 or 2 to identify FISH and DAPI staining images, and classifies FISH and DAPI staining images into mononuclear cells, mature granulocytes, and FISH signals.

4. A system for assisting in the diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia, characterized in that, The system includes an image acquisition module, an image recognition module, and a calculation and determination module; The image acquisition module is used to perform the following: acquiring FISH and DAPI staining images of samples from patients with acute lymphoblastic leukemia of chronic myeloid leukemia; The image recognition module is used to perform the following: identifying mononuclear cells, mature granulocytes, and classifying FISH signals in FISH and DAPI stained images. The criteria for identifying mononuclear cells are that the nucleus is round, and the criteria for identifying mature granulocytes are that the nucleus is lobed or rod-shaped. The FISH signals include BCR gene signals, ABL1 gene signals, and BCR::ABL1 fusion gene signals. The calculation and determination module is used to perform the following: calculate the proportion of BCR::ABL1 fusion gene signal positive cells in mononuclear cells and the proportion of BCR::ABL1 fusion gene signal positive cells in mature granulocytes; if the proportion of BCR::ABL1 fusion gene signal positive cells in mononuclear cells is ≥2% and the proportion of BCR::ABL1 fusion gene signal positive cells in mature granulocytes is ≥2%, then output a high risk of chronic myeloid leukemia acute lymphoblastic leukemia (CML); otherwise, output a low risk of CML.

5. The system for assisting in the diagnosis of acute lymphoblastic leukemia in chronic myeloid leukemia according to claim 4, characterized in that, The image recognition unit is configured to recognize FISH results and DAPI staining images using the model for assisting in the diagnosis of acute lymphoblastic leukemia as described in claim 1 or 2.

6. The system for assisting in the diagnosis of acute lymphoblastic leukemia according to claim 4, characterized in that, The samples include bone marrow samples or peripheral blood samples.

7. A computer device, characterized in that, The computer device includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, performs the function of the system for assisting in the diagnosis of acute lymphoblastic leukemia as described in any one of claims 4-6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it performs the functions of the system for assisting in the diagnosis of acute lymphoblastic leukemia as described in any one of claims 4-6.

Images