Method for screening early stage of coronary heart disease N1 protein by using kit

Abstract

The invention provides a method for screening an early stage of coronary heart disease N1 protein by using a kit, and belongs to the technical field of biomedical detection, and the method comprises the following steps: sorting mononuclear cells; performing incubation; collecting a single cell fluorescence image; performing algorithm analysis; a single cell imaging and AI quantitative strategy is adopted, target mononuclear cells are enriched through high-purity sorting, target protein is labeled with a high-affinity antibody, single cell fluorescence distribution is captured through high-resolution microscopic imaging, integral optical density is calculated through an AI algorithm, and subcellular level spatial quantification of protein expression is achieved. Biomarker detection is improved from a population mean value to a single cell resolution ratio, and the analysis capability on trace expression is remarkably improved.

Description

Technical Field

[0001] The invention belongs to the technical field of biomedical detection, and in particular to a method for screening early-stage coronary heart disease N1 protein using a kit. Background Art

[0002] Coronary heart disease, whose full name is coronary atherosclerotic heart disease, refers to a type of heart disease caused by atherosclerosis of the coronary arteries, which leads to thickening of the blood vessel walls, narrowing of the lumen, or even blockage, thereby causing myocardial ischemia, hypoxia, or necrosis. Its essence is that there is a problem with the "oil pipeline" of the heart, and blood and oxygen cannot reach the myocardial cells smoothly. The most direct harm of coronary heart disease is angina pectoris, which manifests as oppressive pain or discomfort in the chest, often triggered by physical activity or emotional excitement. Unstable atherosclerotic plaques may suddenly rupture, forming a blood clot that completely blocks the blood vessels, leading to acute myocardial infarction, causing large areas of myocardial cells to die irreversibly in a short period of time. This is the most common cause of sudden cardiac death.

[0003] The early symptoms of coronary artery disease (CAD) are often subtle or atypical, making them easily overlooked. The most typical early sign is angina pectoris, characterized by a sensation of pressure, tightness, burning, or heaviness in the center or left side of the chest. The pain may radiate to the left shoulder, inner left arm, neck, jaw, or back. This discomfort is often triggered by physical activity (such as brisk walking, climbing, or lifting heavy objects), emotional excitement, a large meal, or cold surroundings, and is relieved within minutes by rest or sublingual nitroglycerin. However, early symptoms can also be quite atypical: for example, they may present only with mild chest tightness, shortness of breath (especially during activity), unexplained fatigue, or dizziness; or with upper abdominal discomfort, heartburn, and nausea, which can be mistaken for indigestion; or with "misplaced" pain such as jaw pain, toothache, throat tightness, or left shoulder or arm pain. Some patients, particularly women or those with diabetes, may not experience any obvious chest pain in the early stages, instead experiencing only decreased activity tolerance or fatigue without a clear cause. Recognizing these early, mild, and atypical signs is crucial.

[0004] Early detection of coronary heart disease is extremely important. Coronary atherosclerosis is a long and gradual process. In the early stages when the degree of vascular stenosis is not yet serious, patients may only feel mild discomfort or even no symptoms during strenuous activities (called "latent coronary heart disease"), but pathological changes have already occurred and unstable plaques are at risk of rupture at any time. The core value of early detection lies in grasping the "time window": (1) Preventing catastrophic events: By screening to detect early lesions, strong intervention can be carried out before myocardial infarction or sudden death occurs, such as intensive drug treatment (statins, antiplatelet drugs, etc.) and strict lifestyle changes (smoking cessation, healthy diet, exercise) to stabilize plaques and prevent them from rupture. (2) Delaying disease progression: Early intervention can effectively slow down or even reverse the process of atherosclerosis, avoid rapid aggravation of vascular stenosis, and delay or avoid the worsening of angina pectoris and the occurrence of heart failure. (3) Improve treatment effect and prognosis: Compared with rescue treatment after myocardial infarction (such as stents, bypass surgery), intervention before myocardial necrosis has better treatment effect, higher quality of life for patients, longer life expectancy, and lower medical costs. (4) Identify high-risk groups: For people with high-risk factors such as hypertension, hyperlipidemia, diabetes, smoking, obesity, and a family history of premature coronary heart disease, early detection can accurately identify those who need key intervention and achieve individualized prevention. Therefore, paying attention to the early symptoms of coronary heart disease, especially atypical manifestations, and actively conducting targeted examinations, especially for high-risk groups, is a key line of defense to effectively fight this "silent killer" and protect heart health.

[0005] Research has shown that NFAM1 (also known as N1 protein), an immune receptor specifically expressed on monocytes, plays a key role in the development and progression of coronary artery disease. When the vascular endothelium is damaged, oxidized low-density lipoprotein activates the NF-κB pathway in monocytes, inducing upregulation of NFAM1 expression. This upregulation amplifies the inflammatory response by modulating the TLR4 / MyD88 signaling cascade, directly promoting the formation of atherosclerotic plaques. Traditional NFAM1 flow cytometry detection relies on the analysis of the mean fluorescence of a population of cells, which lacks sensitivity for low-abundance, scattered expression and results in high background noise due to low sorting purity. Furthermore, conventional antibodies have limited affinity and are prone to cross-reaction with degradation fragments, resulting in a false-positive rate exceeding 12%.

[0006] In view of the above problems, the present invention proposes a method for early screening of N1 protein in coronary heart disease using a kit. By adopting single-cell imaging and AI quantification strategies, the target monocytes are first enriched by high-purity sorting, and then the target protein is labeled with a high-affinity antibody. Finally, the single-cell fluorescence distribution is captured by high-resolution microscopic imaging, and the integrated optical density is calculated by the AI ​​algorithm to achieve subcellular spatial quantification of protein expression, thereby improving the biomarker detection from "population mean" to "single-cell resolution", significantly improving the ability to analyze trace expressions. Summary of the Invention

[0007] The purpose of the present invention is to address the above-mentioned problems in the existing technology and to propose a method for screening the early stage of coronary heart disease N1 protein using a kit.

[0008] The purpose of the present invention can be achieved through the following technical solutions:

[0009] A method for screening early-stage coronary artery disease N1 protein using a kit, comprising:

[0010] S1. Isolate monocytes from peripheral blood samples using an immunomagnetic bead-microfluidic chip system. The immunomagnetic beads are surface-coupled with anti-CD14 antibodies. The microfluidic chip has a spiral channel structure with a width of 50 ± 5 μm. The sorting conditions are: magnetic field intensity of 0.5 T, flow rate of 1.5 mL / min, and sorting purity ≥ 98%;

[0011] S2. incubating the sorted monocytes with an anti-N1 protein fluorescently labeled antibody, wherein the antibody is a monoclonal antibody with clone number N1-3F8 and is labeled with an Alexa Fluor 647 fluorescent group;

[0012] S3. Single-cell fluorescence images were collected using a fully automated cell morphology analyzer equipped with a 647 nm laser excitation source and a 0.2 μm resolution microscope.

[0013] S4. Analyze fluorescence images using an artificial intelligence image recognition algorithm to calculate the integrated optical density (IOD) of single-cell N1 protein expression. The algorithm is based on a convolutional neural network (CNN) model, and the training set contains single-cell images of 5,000 patients with coronary artery disease.

[0014] S5. When the IOD value is ≥25.0, the subject is judged to be at high risk of coronary heart disease.

[0015] A magnetic bead-microfluidic combined system (50±5μm spiral channel + 0.5T magnetic field) was used to achieve monocyte purity sorting of ≥98%, an increase of more than 18% compared to the traditional Ficoll method (≤80%), and the time was shortened from 60 minutes to 10 minutes; AI-driven single-cell IOD quantification can detect N1 protein ≥10pg / mL, a 100-fold increase compared to flow cytometry (sensitivity 1ng / mL), achieving a breakthrough in detection sensitivity; the threshold of IOD ≥25.0 was verified by ROC (AUC=0.96), and the detection rate for early patients with coronary artery stenosis <50% reached 92% (traditional markers <50%).

[0016] Preferably, the immunomagnetic bead-microfluidic chip combined system comprises:

[0017] The anti-CD14-PE labeled magnetic microbeads with a particle size of 1.0±0.2μm were used to optimize the magnetic response speed and achieve a sorting throughput of 10 6 cells / minute, which is significantly improved compared with traditional magnetic beads;

[0018] The multilayer spiral microfluidic chip with the inner wall of the channel coated with polyethylene glycol-silane copolymer and the anti-adhesion coating design (polyethylene glycol-silane copolymer) increased the cell recovery rate from 85% to ≥98%, which can avoid false negatives caused by cell retention.

[0019] Preferably, the artificial intelligence image recognition algorithm sequentially performs image preprocessing, fluorescence signal positioning, and risk grading operations. The algorithm flow adopts a three-step closed-loop analysis, which avoids subjective errors in manual interpretation through preprocessing → positioning → grading.

[0020] Preferably, the image preprocessing uses Gaussian filtering for noise reduction and U-Net network for segmentation of individual cells. The U-Net cell segmentation accuracy reaches 99.2%, overcoming signal interference caused by cell overlap.

[0021] Preferably, the fluorescence signal localization is performed by using a ResNet-34 model to identify the subcellular distribution of the N1 protein.

[0022] Preferably, the risk grading includes: outputting a high-risk warning when IOD ≥ 25.0, and generating a heat map visualization report, visually displaying the risk area through the heat map visualization, and assisting doctors in locating the pathological mechanism.

[0023] Preferably, the light chain CDR3 region of the anti-N1 protein fluorescently labeled antibody comprises the amino acid sequence: Ser-Ala-Ser-Ser-Ser-Val-Ser (SEQ ID NO: 2).

[0024] The light chain CDR3 sequence (SEQ ID NO: 2) increases the antibody affinity (KD) to 10 -11 M, compared with conventional antibodies (KD~10 -9 M) increased by 100 times, significantly reducing nonspecific binding. The epitope binding region targets the functional domain of the N1 protein, avoiding interference from degradation fragments, and the false positive rate is <5%.

[0025] Preferably, the method is used for screening early coronary heart disease with a coronary artery stenosis degree of less than 50%, with a positive coincidence rate of ≥92%.

[0026] It is specially designed for early lesions with coronary artery stenosis <50%, making up for the blind spot of traditional imaging (CTA / angiography) for microvascular lesions. The positive compliance rate is ≥92% (95% CI: 89-95%), which is 40% higher than the NFAM1 flow cytometry method (65.6%). It can provide early warning of coronary heart disease 2-5 years in advance.

[0027] Preferably, the method is implemented by a human peripheral blood mononuclear cell automated sorting and analysis system (HPB-MASAS), which includes a sorting module, an imaging module and an analysis module;

[0028] The sorting module integrates a magnetic field generator and a microfluidic chip, reducing its size to 1 / 5 of traditional equipment and making it suitable for primary care clinics.

[0029] The imaging module is equipped with a 647nm laser and an EMCCD camera. The EMCCD camera has a signal-to-noise ratio of 25:1 under low light conditions, ensuring accurate capture of weak fluorescence signals (such as those in early-stage patients).

[0030] The analysis module runs an artificial intelligence image recognition algorithm through an embedded processor to shorten the result output time.

[0031] Compared with the existing technology, the method of using the kit to screen for early-stage coronary heart disease N1 protein has the following beneficial effects:

[0032] 1. The present invention provides a method for screening for early-stage N1 protein in coronary heart disease using a kit. By combining magnetic beads and spiral microfluidic chips, the purity of monocytes sorted is ≥98% and the recovery rate is ≥98%, which is significantly improved compared with traditional methods. It also shortens the sorting time and greatly reduces the risk of false negatives caused by cell loss.

[0033] 2. The present invention provides a method for screening N1 protein in the early stages of coronary heart disease using a kit. It uses high-affinity antibodies combined with AI-driven single-cell IOD quantitative analysis, which has greatly improved sensitivity compared to flow cytometry and can accurately capture early trace biomarkers.

[0034] 3. The present invention provides a method for screening N1 protein in the early stage of coronary heart disease using a kit. Based on the ROC verification threshold, it has a high detection rate for early coronary heart disease with coronary artery stenosis <50%, which is significantly better than traditional markers and fills the imaging blind spot of microvascular lesions.

[0035] 4. The present invention provides a method for screening early-stage N1 protein of coronary heart disease using a kit, which achieves high-precision cell segmentation through an AI three-step closed-loop algorithm, overcomes the interference of cell overlap, and at the same time reduces the false positive rate by the antibody light chain CDR3 sequence, avoiding misjudgment of degradation fragments.

[0036] 5. The present invention provides a method for screening N1 protein in the early stage of coronary heart disease using a kit, which adopts the integrated equipment HPB-MASAS system, equipped with a low-light and high signal-to-noise ratio EMCCD camera, and embedded AI analysis to reduce the time consumption of the entire process and adapt to scenarios with limited resources.

[0037] 6. The present invention provides a method for screening N1 protein in the early stage of coronary heart disease using a kit. Through high-risk early warning and heat map visualization report, the pathological mechanism can be intuitively located to assist doctors in formulating intervention strategies.

[0038] In summary, the present invention provides a method for early screening of N1 protein in coronary heart disease using a kit. It adopts single-cell imaging and AI quantification strategies. The target monocytes are first enriched by high-purity sorting, and then the target protein is labeled with a high-affinity antibody. Finally, the single-cell fluorescence distribution is captured by high-resolution microscopic imaging, and the integrated optical density is calculated by the AI ​​algorithm to achieve subcellular spatial quantification of protein expression, thereby improving biomarker detection from "population mean" to "single-cell resolution", significantly improving the ability to analyze trace expressions. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 This is a diagram of the auxiliary tool for the sorting system in the first specific embodiment.

[0040] Figure 2 This is a diagram of the artificial intelligence image recognition algorithm software independently developed in specific embodiment 1. DETAILED DESCRIPTION

[0041] The following are specific embodiments of the present invention and the accompanying drawings to further describe the technical solutions of the present invention, but the present invention is not limited to these embodiments. Specific embodiment one:

[0043] A method for screening early-stage coronary artery disease N1 protein using a kit, comprising:

[0044] S1. Isolate monocytes from peripheral blood samples using an immunomagnetic bead-microfluidic chip system. The immunomagnetic beads are surface-coupled with anti-CD14 antibodies. The microfluidic chip has a spiral channel structure with a width of 50 ± 5 μm. The sorting conditions are: magnetic field intensity of 0.5 T, flow rate of 1.5 mL / min, and sorting purity ≥ 98%;

[0045] The immunomagnetic bead-microfluidic chip combined system includes:

[0046] The anti-CD14-PE labeled magnetic microbeads with a particle size of 1.0±0.2μm were used to optimize the magnetic response speed and achieve a sorting throughput of 10 6 cells / minute, which is significantly improved compared with traditional magnetic beads;

[0047] The multilayer spiral microfluidic chip with the inner wall of the channel coated with polyethylene glycol-silane copolymer and the anti-adhesion coating design (polyethylene glycol-silane copolymer) increased the cell recovery rate from 85% to ≥98%, which can avoid false negatives caused by cell retention.

[0048] S2. incubating the sorted monocytes with an anti-N1 protein fluorescently labeled antibody, wherein the antibody is a monoclonal antibody with clone number N1-3F8 and is labeled with an Alexa Fluor 647 fluorescent group;

[0049] The light chain CDR3 region of the anti-N1 protein fluorescently labeled antibody comprises the amino acid sequence: Ser-Ala-Ser-Ser-Ser-Val-Ser (SEQ ID NO: 2).

[0050] S3. Single-cell fluorescence images were collected using a fully automated cell morphology analyzer equipped with a 647 nm laser excitation source and a 0.2 μm resolution microscope.

[0051] S4. Analyze fluorescence images using an artificial intelligence image recognition algorithm to calculate the integrated optical density (IOD) of single-cell N1 protein expression. The algorithm is based on a convolutional neural network (CNN) model, and the training set contains single-cell images of 5,000 patients with coronary artery disease.

[0052] The AI-powered image recognition algorithm sequentially performs image preprocessing, fluorescence signal localization, and risk grading. The algorithm utilizes a three-step closed-loop analysis process, from preprocessing to localization to grading, to avoid subjective errors in manual interpretation. Image preprocessing utilizes Gaussian filtering for noise reduction and a U-Net network for individual cell segmentation. The U-Net cell segmentation accuracy reaches 99.2%, overcoming signal interference caused by cell overlap. Fluorescence signal localization utilizes a ResNet-34 model to identify the subcellular distribution of the N1 protein. Risk grading includes a high-risk warning when the IOD ≥ 25.0, and generates a heatmap visualization report. This heatmap visualization intuitively displays risk areas, assisting physicians in locating the pathological mechanism.

[0053] S5. When the IOD value is ≥25.0, the subject is judged to be at high risk of coronary heart disease.

[0054] A magnetic bead-microfluidic combined system (50±5μm spiral channel + 0.5T magnetic field) was used to achieve monocyte purity sorting of ≥98%, an increase of more than 18% compared to the traditional Ficoll method (≤80%), and the time was shortened from 60 minutes to 10 minutes; AI-driven single-cell IOD quantification can detect N1 protein ≥10pg / mL, a 100-fold increase compared to flow cytometry (sensitivity 1ng / mL), achieving a breakthrough in detection sensitivity; the threshold of IOD ≥25.0 was verified by ROC (AUC=0.96), and the detection rate for early patients with coronary artery stenosis <50% reached 92% (traditional markers <50%).

[0055] The method is implemented using a human peripheral blood mononuclear cell automated sorting and analysis system (HPB-MASAS). The system includes a sorting module, an imaging module, and an analysis module. The sorting module integrates a magnetic field generator and a microfluidic chip, reducing its size to 1 / 5 of that of traditional equipment and making it suitable for primary care clinics. The imaging module is equipped with a 647nm laser and an EMCCD camera. The EMCCD camera has a signal-to-noise ratio of 25:1 under low light conditions, ensuring accurate capture of weak fluorescence signals (such as those of early-stage patients). The analysis module runs an artificial intelligence image recognition algorithm through an embedded processor, shortening the time it takes to output results. Specific embodiment two:

[0057] The present invention provides an early screening kit for coronary heart disease based on the combination of immunomagnetic beads and microfluidic chips and artificial intelligence analysis. The core lies in breaking through the bottleneck of traditional detection through the synergy of triple technologies. First, an anti-CD14 antibody-coupled magnetic microbeads with a particle size of 1.0±0.2μm is used in conjunction with a 50±5μm spiral microfluidic chip system. The inner wall of the channel is coated with a polyethylene glycol-silane copolymer anti-adhesion coating. The purity and recovery rate of monocyte sorting are both ≥98% under a 0.5T magnetic field and a flow rate of 1.5mL / min, which is more than 18% higher than the traditional Ficoll method. The sorting time is shortened from 60 minutes to 10 minutes, significantly reducing the risk of false negatives caused by cell loss. Secondly, a high-affinity fluorescent-labeled antibody (clone number N1-3F8) is used to target the N1 protein. The specific sequence of the light chain CDR3 region, Ser-Ala-Ser-Ser-Ser-Val-Ser (SEQ ID NO: 2), makes the antibody affinity reach KD=10 -11 M, 100 times more potent than conventional antibodies, combined with Alexa Fluor 647 fluorescent labeling, effectively avoids interference from degradation fragments.

[0058] Finally, single-cell images were collected using a fully automated cell morphology analyzer. An AI-driven three-step closed-loop algorithm was employed: a U-Net network achieved 99.2% cell segmentation accuracy to overcome overlapping interference, a ResNet-34 model localized the subcellular distribution of the N1 protein, and a CNN model calculated the integrated optical density (IOD). An IOD ≥ 25.0 was used to identify individuals at high risk of CHD, a threshold validated by a receiver operating characteristic (ROC) curve (AUC = 0.96). Systematic testing demonstrated that this method achieved a sensitivity of 10 pg / mL, a 100-fold improvement over flow cytometry. The detection rate for early-stage patients with coronary artery stenosis <50% reached 92.1% (compared to <50% for traditional methods), with a false-positive rate of <5%. Furthermore, the integrated HPB-MASAS system compressed the entire process to 40 minutes, reducing the device size by 80% and making it suitable for grassroots use. In clinical trials, 58 asymptomatic patients with early-stage CHD were successfully identified, detecting the risk of lesion progression an average of 2.3 years earlier. This method fundamentally addresses the three key challenges of early CHD screening: insufficient sensitivity, microvascular blind spots, and complex procedures. Specific embodiment three:

[0060] In a comparative experiment for sorting efficiency, peripheral venous blood samples (5 mL EDTA anticoagulated per sample) were collected from 100 subjects. Each sample was divided equally into two aliquots: one was processed using conventional Ficoll density gradient centrifugation, and the other was sorted using the immunomagnetic bead-microfluidic system of the kit of the present invention. The conventional group followed the standard procedure: blood was diluted 1:1 with PBS and slowly layered onto Ficoll separation buffer. The blood was centrifuged at 400 × g for 30 minutes, the buffy coat cells were aspirated, and the cells were washed twice with PBS. In the inventive group, whole blood was mixed with anti-CD14-PE-labeled magnetic microbeads (particle size 1.0 ± 0.2 μm) at a volume ratio of 1:50, incubated at room temperature for 15 minutes, and then injected into a microfluidic chip (channel width 50 ± 5 μm, inner wall coated with polyethylene glycol-silane copolymer). Mononuclear cells were sorted at a magnetic field strength of 0.5 T and a flow rate of 1.5 mL / min. After sorting, the purity (percentage of CD14+ cells) and recovery rate (number of recovered cells / initial number of monocytes × 100%) of the two groups of cells were detected by flow cytometry (BD FACSAria III). The recovery rate was verified by counting the number of active cells using a hemocytometer and trypan blue staining.

[0061] Table 1 Comparison of sorting efficiency

[0062]

[0063] Experimental data showed that the average purity of the traditional Ficoll method was only 78.2±3.1%, the recovery rate was 84.7±2.9%, and the sorting time was 58.4±5.2 minutes; while the sorting purity of the group of the present invention reached 98.5±0.8% (p<0.001), the recovery rate was increased to 98.3±1.2% (p<0.001), and the time was shortened to 9.8±1.5 minutes. The results showed that the microfluidic spiral channel design combined with the anti-adhesion coating significantly reduced cell retention (retention rate of the traditional method was >15%, and this method was <2%), and the 1.0μm magnetic beads optimized the magnetic field response speed, making the throughput reach 1.2×10 6 cells / minute, which is 3 times higher than traditional magnetic beads (particle size 4.5μm), fundamentally solving the false negative problem caused by cell loss. Specific embodiment four:

[0065] In the sensitivity and specificity validation experiment, a gradient concentration standard containing recombinant human N1 protein (0.1pg / mL, 1pg / mL, 10pg / mL, 100pg / mL, 1ng / mL, 10ng / mL, 100ng / mL) was first prepared, and 10% healthy human serum was added to PBS to simulate the real sample environment. Each concentration standard was divided into two equal parts: one was detected by flow cytometry (BD FACS Celesta), and the other was detected by commercial anti-N1 antibody (clone 4D2, KD ~ 10 -9 M) were incubated according to standard procedures and then loaded onto the instrument; the other was processed using this kit system: first, a high-affinity anti-N1 antibody (clone N1-3F8, CDR3 sequence Ser-Ala-Ser-Ser-Ser-Val-Ser, labeled with Alexa Fluor 647) was added and incubated at 37°C in the dark for 30 minutes. Single-cell fluorescence images were then acquired using the HPB-MASAS system imaging module (647nm laser / EMCCD camera), and the IOD value was calculated using an AI algorithm. To verify specificity, three additional control groups were set up: healthy human serum samples (n=50), proteinase K-treated N1 degradation fragment samples (n=30), and an isotype control antibody group. Sensitivity was evaluated by limit of detection (LoD) and functional sensitivity (20% CV concentration), and specificity was assessed by false positive rate (the proportion of non-target samples with an IOD ≥ 25.0).

[0066] Table 2 Verification of detection sensitivity and specificity

[0067]

[0068] Experimental data showed that flow cytometry had a significant signal at ≥1 ng / mL (S / N>3) and a functional sensitivity of 1.2 ng / mL (CV=19.8%), while the kit achieved an S / N=4.1 at 10 pg / mL and a functional sensitivity of 8.7 pg / mL (CV=18.3%), a 115-fold increase compared to flow cytometry. In the false positive rate test, the flow cytometry group had a false positive rate of 12.3% (6 / 50 healthy samples were false positive), while the kit had a false positive rate of only 4.7% (2 / 50 healthy samples were false positive, and all degradation fragment groups were negative), attributed to the high-affinity antibody (surface plasmon resonance verified KD=1.3×10 -11 M) and AI algorithm to filter the debris signal (U-Net segmentation excludes 95% extracellular debris). Statistical analysis confirmed that the IOD value of this method was strongly linearly correlated with the concentration in the low concentration range of 1-100 pg / mL (R 2 =0.992, p<0.001), while flow cytometry had no response in the same interval (R 2 =0.21, p=0.38). Specific embodiment five:

[0070] In a trial evaluating the efficacy of early coronary artery disease screening, 300 patients diagnosed by coronary angiography (excluding those with concomitant tumors or infectious diseases) were consecutively enrolled and divided into three groups according to the degree of stenosis: Group A (stenosis <50%, n=120), Group B (stenosis 50-70%, n=100), and Group C (stenosis >70%, n=80). Fasting peripheral venous blood (5 mL EDTA anticoagulated) was collected from all subjects in the morning. After double-blind identification, the samples were subjected to conventional NFAM1 flow cytometry (commercial antibody) and this kit. The kit's processing workflow includes: first, monocytes are isolated using an immunomagnetic bead-microfluidics system (magnetic field 0.5T, flow rate 1.5mL / min). The resulting cell suspension is then incubated with a high-affinity anti-N1 fluorescent antibody (clone N1-3F8, labeled with Alexa Fluor 647) at 37°C in the dark for 30 minutes. Single-cell fluorescence images are then acquired using the HPB-MASAS system imaging module (647nm laser / 0.2μm resolution lens / EMCCD camera). Finally, an AI algorithm performs a three-step analysis: U-Net segmentation of single cells, ResNet-34 localization of protein distribution, and CNN model calculation of IOD values. Using coronary angiography results as the gold standard, a receiver operating characteristic (ROC) curve was plotted to determine the optimal IOD threshold, and the detection rate for each group was calculated.

[0071] Table 3 Detection sensitivity and specificity verification

[0072]

[0073] Data showed that the area under the receiver operating characteristic (ROC) curve (AUC) was 0.96 (95% CI: 0.93–0.99), and the Youden index defined an IOD ≥ 25.0 as a high-risk threshold (sensitivity 92.1%, specificity 95.3%). In group A (early lesions), the proposed method detected 111 positive cases (92.1%), significantly higher than the NFAM1 flow cytometry method (65.6%, p < 0.001). The detection rates in groups B and C were 96.3% and 98.8%, respectively. The mean IOD increased with the degree of stenosis (group A: 26.8±3.2, group B: 41.5±5.7, group C: 68.9±8.4, p < 0.001). Subgroup analysis revealed that in patients with diabetes and microvascular disease (n = 45), the proposed method had a detection rate of 94.2%, compared with only 28.9% using coronary CTA (p < 0.001). The false-positive rate was only 4.3% (13 / 300), significantly lower than the 12.7% of flow cytometry (p = 0.002). Typical case follow-up revealed that among 58 asymptomatic patients with IOD ≥ 25.0 but coronary artery stenosis < 50%, 51 (87.9%) experienced stenosis progression or cardiovascular events during an average follow-up of 2.3 years, including one patient who suffered a myocardial infarction six months later (initial stenosis 40%, progressed to 85%). This method demonstrates that this approach can provide early warning of risk an average of 2.5 years earlier than conventional imaging (HR = 8.7, 95% CI: 4.1–18.3). AI-generated heat maps further demonstrated that N1 protein in early-stage patients primarily accumulates on the mitochondrial membrane of monocytes (accounting for 79.3%), providing a spatial basis for targeting interventions.

[0074] The specific embodiments described herein are merely illustrative of the spirit of the present invention. Persons skilled in the art may make various modifications, additions, or substitutions to the described specific embodiments without departing from the spirit of the present invention or exceeding the scope of the appended claims.

Claims

1. A method for screening early-stage coronary heart disease N1 protein using a kit, characterized in that: include: S1. Isolate monocytes from peripheral blood samples using an immunomagnetic bead-microfluidic chip system. The immunomagnetic beads are surface-coupled with anti-CD14 antibodies. The microfluidic chip has a spiral channel structure with a width of 50 ± 5 μm. The sorting conditions are: magnetic field intensity of 0.5 T, flow rate of 1.5 mL / min, and sorting purity ≥ 98%; S2. incubating the sorted monocytes with an anti-N1 protein fluorescently labeled antibody, wherein the antibody is a monoclonal antibody with clone number N1-3F8 and is labeled with an Alexa Fluor 647 fluorescent group; S3. Single-cell fluorescence images were collected using a fully automated cell morphology analyzer equipped with a 647 nm laser excitation source and a 0.2 μm resolution microscope. S4. Analyze fluorescence images using an artificial intelligence image recognition algorithm to calculate the integrated optical density (IOD) of single-cell N1 protein expression. The algorithm is based on a convolutional neural network (CNN) model, and the training set contains single-cell images of 5,000 patients with coronary artery disease. S5. When the IOD value is ≥25.0, the subject is judged to be at high risk of coronary heart disease.

2. A method for screening early-stage coronary heart disease N1 protein using a kit according to claim 1, characterized in that: The immunomagnetic bead-microfluidic chip combined system includes: anti-CD14-PE labeled magnetic microbeads with a particle size of 1.0 ± 0.2 μm; A multilayer spiral microfluidic chip with the inner wall of the channel coated with polyethylene glycol-silane copolymer.

3. The method for screening early-stage coronary heart disease N1 protein using a kit according to claim 1, characterized in that: The artificial intelligence image recognition algorithm sequentially performs image preprocessing, fluorescence signal localization, and risk grading operations.

4. A method for screening early-stage coronary heart disease N1 protein using a kit according to claim 3, characterized in that: The image preprocessing uses Gaussian filtering for noise reduction and U-Net network for segmentation of single cells.

5. The method for screening early-stage coronary heart disease N1 protein using a kit according to claim 3, characterized in that: The fluorescence signal localization is performed using the ResNet-34 model to identify the subcellular distribution of the N1 protein.

6. The method for screening early-stage coronary heart disease N1 protein using a kit according to claim 3, characterized in that: The risk grading includes: outputting a high-risk warning when IOD ≥ 25.0, and generating a heat map visualization report.

7. The method for screening early-stage coronary heart disease N1 protein using a kit according to claim 1, characterized in that: The light chain CDR3 region of the anti-N1 protein fluorescently labeled antibody comprises the amino acid sequence: Ser-Ala-Ser-Ser-Ser-Val-Ser (SEQ ID NO: 2).

8. The method for screening early-stage coronary heart disease N1 protein using a kit according to claim 1, characterized in that: The method is used for screening early coronary heart disease with a coronary artery stenosis degree of less than 50%, and the positive coincidence rate is ≥92%.

9. The method for screening early-stage coronary heart disease N1 protein using a kit according to claim 1, characterized in that: The method is implemented by a human peripheral blood mononuclear cell automated sorting and analysis system (HPB-MASAS), which includes a sorting module, an imaging module, and an analysis module; Wherein, the sorting module integrates a magnetic field generator and a microfluidic chip; Wherein, the imaging module is equipped with a 647nm laser and an EMCCD camera; The analysis module runs an artificial intelligence image recognition algorithm through an embedded processor.

Images