Diabetic complication identification method based on exhaled gas biomarker combination

Abstract

The invention provides a diabetic complication identification method based on an exhaled gas biomarker combination. A marker is selected from one or more of isoprene, acetone, isopropanol, benzene, toluene, ethylbenzene, p-xylene, undecane and tetradecane. The invention provides a safe, convenient, efficient and non-invasive diagnosis method so as to monitor the health condition of the human body in time, thereby providing a new thought and method for early diagnosis, treatment and prevention of the human body.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of screening for diabetic complications; in particular, it relates to a method for identifying diabetic complications based on a combination of biomarkers in exhaled breath. BACKGROUND

[0002] Diabetic patients may suffer from various diabetic complications due to long-term maintenance of high blood glucose levels, which may cause damage to blood vessels, nerves and other tissues. The diabetic complication stage is an irreversible process in the late stage of the diabetic course, which seriously affects the health and quality of life of patients.

[0003] Diagnosis of diabetic complications usually requires a combination of clinical symptoms, physical examination, laboratory examination and imaging examination. However, in most cases, patients have already suffered from damage to blood vessels, nerves or other tissues when they are diagnosed with diabetic complications, and it is not possible to control them as early as possible, in a timely and effective manner. For example, the diagnosis of diabetic retinopathy is usually made by observing the retina with an ophthalmoscope or fundus camera to check for diabetic changes such as hemorrhage, exudation, edema, macular edema. Diabetic nephropathy is diagnosed with the aid of urine protein quantification, serum creatinine determination or kidney ultrasound examination. For diabetic peripheral neuropathy, nerve electrophysiological examination, including nerve conduction velocity determination and electromyography, is used to assess the degree of nerve function impairment, and somatosensory tests such as touch, temperature sensation or vibration sensation can also be performed. For diabetic foot disease, the skin, nails and soles of the feet are observed, and the foot pulse and foot temperature are measured to check for abnormal conditions such as ulcers and infections. Diabetic cardiovascular disease is usually assessed by electrocardiogram to determine whether there is myocardial ischemia, arrhythmia or other abnormal conditions. By analyzing biomarkers in the exhaled breath of diabetic complication patients, it may be possible to find features related to diabetic complications, thereby providing new ideas and methods for early diagnosis, treatment and prevention. SUMMARY

[0004] In view of the above-mentioned shortcomings of the prior art, the purpose of the present application is to provide a combination of exhaled breath for distinguishing between diabetic patients and diabetic complication patients, and to identify diabetic complications by analyzing the concentrations of the compounds in the biomarker combination.

[0005] The first aspect of the present application discloses a marker combination for identifying and / or predicting diabetic complications, the markers being selected from one or more of isoprene, acetone, isopropanol, benzene, toluene, ethylbenzene, p-xylene, undecane and tetradecane.

[0006] In some embodiments of the present application, the markers are selected from a combination of isoprene and acetone.

[0007] In certain embodiments of the application, the marker is selected from the group of isoprene and isopropanol.

[0008] In certain embodiments of the application, the marker is selected from the group of isoprene and tetradecane.

[0009] In certain embodiments of the application, the marker is selected from the group of acetone and isopropanol.

[0010] In certain embodiments of the application, the marker is selected from the group of acetone and tetradecane.

[0011] In certain embodiments of the application, the marker is selected from the group of isopropanol and tetradecane.

[0012] In certain embodiments of the application, the marker is selected from the group of isoprene, acetone and isopropanol.

[0013] In certain embodiments of the application, the marker is selected from the group of isoprene, acetone and tetradecane.

[0014] In certain embodiments of the application, the marker is selected from the group of isoprene, isopropanol and tetradecane.

[0015] In certain embodiments of the application, the marker is selected from the group of isopropanol, acetone and tetradecane.

[0016] In certain embodiments of the application, the marker is selected from the group of isoprene, acetone, isopropanol and tetradecane.

[0017] In certain embodiments of the application, the marker is selected from the group of isoprene, acetone, isopropanol, tetradecane and benzene.

[0018] In certain embodiments of the application, the marker is selected from the group of isoprene, acetone, isopropanol, tetradecane, benzene and ethylbenzene.

[0019] In certain embodiments of the application, the marker is selected from the group of isoprene, acetone, isopropanol, tetradecane, benzene, ethylbenzene and p-xylene.

[0020] In certain embodiments of the application, the marker is selected from the group of isoprene, acetone, isopropanol, tetradecane, benzene, ethylbenzene, p-xylene and undecane.

[0021] In certain embodiments of the application, the marker is selected from the group of isoprene, acetone, isopropanol, tetradecane, toluene, ethylbenzene, p-xylene, benzene and undecane.

[0022] Preferably, the marker is derived from exhaled breath. Preferably, it is alveolar exhaled breath.

[0023] The second aspect of the present application discloses a use of the above-mentioned marker detecting substance in the preparation of a product for identifying and / or predicting diabetic complications.

[0024] Preferably, the marker detecting substance refers to a substance for detecting the volume or mass concentration of the marker.

[0025] The third aspect of the present application discloses a product for identifying and / or predicting diabetic complications in vitro, which comprises the above-mentioned marker detecting substance.

[0026] Preferably, the product comprises at least one of a kit, a membrane strip, a chip and a detection system.

[0027] Preferably, the detection system is selected from a gas sensing system or a gas chromatograph.

[0028] The fourth aspect of the present application discloses a use of a product for identifying and / or predicting diabetic complications in vitro in the screening of drugs for treating and / or preventing diabetic complications.

[0029] Preferably, the product comprises the above-mentioned marker detecting substance.

[0030] The fifth aspect of the present application discloses a screening method of the marker, comprising the following steps:

[0031] 1) detecting the types and contents of compounds in a sample of a subject;

[0032] 2) analyzing the detection results to obtain the marker.

[0033] Preferably, the subject comprises a person with diabetic complications and a diabetic patient.

[0034] Preferably, the analysis comprises analyzing the types and contents of the above-mentioned compounds.

[0035] Preferably, the analysis comprises identifying the compounds in exhaled breath according to the retention time of NIST14 database and standard substance.

[0036] Preferably, the analysis comprises quantification using spectral peak area.

[0037] Preferably, the analysis comprises t-test or z-test analysis.

[0038] Preferably, the analysis comprises calculating P value.

[0039] Further preferably, the analysis comprises obtaining compounds with P value <0.05.

[0040] Preferably, the compounds with P value <0.05 are the marker.

[0041] Preferably, the sample is exhaled breath, more preferably alveolar exhaled breath.

[0042] In some embodiments of the present application, the method for detecting the compound is:

[0043] The exhaled breath of the subject is enriched using an adsorbent;

[0044] The adsorbent after enrichment of the exhaled breath is thermally desorbed using a thermal desorber, and then detected using gas chromatography-mass spectrometry, and the detection results are subjected to qualitative analysis and quantitative analysis, the type of the compound is obtained according to the qualitative analysis, and the content of the compound is obtained according to the quantitative analysis.

[0045] Preferably, the adsorbent is selected from one or more of Tenax, Carbograph, Chromosorb and PoraPak.

[0046] Further preferably, the adsorbent is Tenax TA.

[0047] Preferably, the thermal desorption is divided into first desorption and second desorption.

[0048] Preferably, the first desorption temperature is 250-350℃.

[0049] Preferably, the first desorption time is 5-15min.

[0050] Preferably, the second desorption temperature is 280-370℃.

[0051] Preferably, the second desorption time is 1-5min.

[0052] Preferably, the chromatographic column of the gas chromatograph is: DB624 UI.

[0053] Preferably, the carrier gas flow rate of the gas chromatograph is 0.5-2mL / min in constant flow mode, more preferably 1mL / min.

[0054] Preferably, the carrier gas pressure is 5-15psi.

[0055] Preferably, the initial temperature of the gas chromatograph is 30-50℃ for 2-8min.

[0056] Preferably, the temperature rising program of the gas chromatograph is a rate of 5-15℃ / min to 150-250℃ for 5-15min.

[0057] Preferably, the transmission line temperature of the mass spectrometer is 200-250℃, more preferably 240℃.

[0058] Preferably, the ionization mode of the mass spectrometer is electron impact (EI).

[0059] Preferably, the ionization energy of the mass spectrometer is 60-80 eV.

[0060] Preferably, the ion source temperature of the mass spectrometer is 200-250 °C.

[0061] Preferably, the quadrupole rod temperature of the mass spectrometer is 130-170 °C.

[0062] Preferably, the scan range of the mass spectrometer is m / z 35-390.

[0063] The sixth aspect of the present application discloses a method for constructing a diabetic complication identification model, comprising the following steps:

[0064] 1) Constructing a sample data set based on the detection amount of the marker in the sample of a subject, wherein the subject comprises a patient diagnosed with diabetic complications and a diabetic patient;

[0065] 2) Randomly dividing the sample data set into a test set and a training set, learning the training set using a machine learning method,

[0066] obtaining the diabetic complication prediction model.

[0067] Preferably, the sample is selected from exhaled breath, and further preferably is alveolar exhaled breath.

[0068] Preferably, the machine learning method is selected from at least one of a random forest algorithm, a support vector machine algorithm, a decision tree algorithm, a K-nearest neighbor algorithm, a logistic regression algorithm, and a neural network algorithm, and further preferably is a decision tree algorithm and / or a support vector machine algorithm.

[0069] Preferably, the division ratio of the test set and the training set is 1:(1-5). In some preferred embodiments, it is 1:4.

[0070] Preferably, the method further comprises testing the diabetic complication prediction model using the test set.

[0071] The seventh aspect of the present application discloses a method for screening diabetic complications, comprising the following steps:

[0072] S1, obtaining the detection amount data of the marker in the sample of a subject to be tested;

[0073] S2, processing the detection amount data using the diabetic complication prediction model obtained by the construction method to output a diabetic complication diagnosis result.

[0074] Preferably, the sample is exhaled breath, and further preferably is alveolar exhaled breath.

[0075] The eighth aspect of the present application discloses a device for diagnosing diabetic complications, which comprises:

[0076] a data module for obtaining the detection amount data of the markers in the sample of the subject;

[0077] an evaluation module for processing the detection amount data by using the diabetic complications prediction model obtained by the construction method to output the diagnosis result of diabetic complications.

[0078] The ninth aspect of the present application discloses a device comprising a processor and a memory, wherein the memory is used for storing a computer program, and the processor is used for executing the computer program stored in the memory to make the device execute the screening method or the construction method.

[0079] The tenth aspect of the present application discloses a computer readable storage medium, which stores a computer program, and the program is executed by a processor to realize the screening method or the construction method.

[0080] The purpose of the present application is to provide an exhaled gas combination for distinguishing diabetic patients and diabetic complication patients, and to identify diabetic complications by analyzing the concentration of each compound in the biomarker combination.

[0081] As described above, compared with the prior art, the present application has the following beneficial effects:

[0082] 1) The two exhaled gas biomarker combination models involved in the present application can be applied to large-scale preliminary screening of diabetic complications.

[0083] 2) The exhaled gas sample involved in the present application is non-invasive and simple and fast in the collection process.

[0084] 3) The biomarker combination involved in the present application contains a large number of compounds, which can exclude the interference of other diseases as much as possible. BRIEF DESCRIPTION OF DRAWINGS

[0085] Figure 1 The structure diagram of the CO2 controlled exhaled gas collection device used in the present application is shown. (A) The CO2 in the exhaled gas is <C50, and the exhaled gas is not collected; (B) The CO2 in the exhaled gas is ≥C50, and the exhaled gas is collected in the gas bag.

[0086] Figure 2 The diabetic complication identification model of Example 2 of the present application is shown. DETAILED DESCRIPTION

[0087] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. Test methods in the following examples that do not specify specific conditions are generally performed under conventional conditions or according to the conditions recommended by the respective manufacturers.

[0088] When numerical ranges are given in the embodiments, it should be understood that, unless otherwise stated in the present invention, both endpoints of each numerical range and any value between the two endpoints may be selected. Unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. In addition to the specific methods, apparatus, and materials used in the embodiments, based on the knowledge of the prior art possessed by one of ordinary skill in the art and the description of this invention, any prior art methods, apparatus, and materials similar to or equivalent to those described, apparatus, and materials in the embodiments of this invention may be used to implement the present invention.

[0089] The underlying principle of this invention is that diabetes complications, as a metabolic disorder, result from a lack of insulin or impaired insulin action, leading to elevated blood sugar levels and an imbalance of metabolic products in the body. This, in turn, causes changes in the concentration of volatile organic compounds (VOCs) in the body. These VOCs can cross the blood-air barrier in the lungs and enter exhaled air, thus altering their concentration in the exhaled air. By detecting changes in the concentration of specific compounds in exhaled air, the disease can be identified.

[0090] The nine compounds involved in this invention's method for identifying diabetic complications based on a combination of exhaled breath biomarkers include: isoprene, acetone, isopropanol, toluene, ethylbenzene, p-xylene, and tetradecane. Two machine learning models were constructed for each compound.

[0091] (1) A decision tree model was constructed based on isoprene + acetone + isopropanol + tetradecane;

[0092] (2) A support vector machine model was constructed based on isoprene + acetone + isopropanol + tetradecane.

[0093] Figure 1This is a schematic diagram of the exhaled air collection device used in this invention. The exhaled air analysis device used in this invention includes an exhaled air collection device, a pre-concentration device, an adsorption tube, a thermal desorption instrument, and gas chromatography-mass spectrometry (GC-MS). The specific detection method is as follows: The CO2-controlled alveolar exhaled air collection device collects the subject's alveolar exhaled air into a Teflon gas bag. The exhaled air in the gas bag is then transferred to an adsorption tube containing Tenax TA adsorption material for pre-concentration. The adsorption tube containing the concentrated exhaled air components is placed in a thermal desorption instrument for heating to desorb the exhaled air components. The desorbed exhaled air components are then delivered to a GC-MS instrument with a DB624 UI column for detection.

[0094] The following embodiments of this application employ a UNITY-xr TD thermal desorption system (Markes International, UK) and an 8890 / 5977B gas chromatography-mass spectrometry system (Agilent Technologies, USA).

[0095] The conditions for thermal desorption are:

[0096] The primary desorption temperature is 300℃, and the desorption time is 10 min.

[0097] The secondary desorption temperature was 320℃, and the desorption time was 3 minutes.

[0098] Transmission line temperature: 200℃;

[0099] No diversion.

[0100] The determination conditions for gas chromatography are as follows:

[0101] Column: DB624 UI;

[0102] Carrier gas: Helium purity ≥ 99.999%;

[0103] Carrier gas pressure: 10 psi;

[0104] Temperature ramp-up program: Initial temperature 40℃, hold for 5 min; then ramp up to 200℃ at a rate of 10℃ / min and hold for 10 min.

[0105] The determination conditions for mass spectrometry are as follows:

[0106] Transmission line temperature: 240℃;

[0107] Ionization method: Electron bombardment (EI);

[0108] Ionization voltage: 70 eV;

[0109] Ion source temperature: 230℃;

[0110] Quadrupole temperature: 150℃;

[0111] Detection mode: Qualitative analysis using total ion chromatogram (TIC) with full scan, and semi-quantitative analysis using peak area method;

[0112] Quality range: 35~390amu.

[0113] All materials, reagents and instruments used in this invention are well known in the art, but this does not limit the implementation of the invention. Other reagents and devices well known in the art can be applied to the implementation of the following embodiments of the invention.

[0114] Example 1

[0115] In this embodiment, the following is adopted: Figure 1 The CO2-controlled exhaled breath collection device shown collects alveolar exhaled breath from subjects and analyzes the collected alveolar exhaled breath to screen for compounds whose levels differ significantly between patients with diabetic complications and diabetic patients as biomarkers. These include the following:

[0116] Subject grouping: The subjects were divided into a diabetic comorbidity group and a diabetic group. The diabetic comorbidity group consisted of 38 patients with diabetic comorbidity, and the diabetic group consisted of 38 patients with diabetic comorbidity.

[0117] Both the diabetic complication group and the diabetic group rinsed their mouths with purified water before exhaling.

[0118] Method Design: Exhaled breath samples were collected from patients with diabetic complications and those with diabetes. After pretreatment, the exhaled breath was analyzed by gas chromatography-mass spectrometry to obtain the chromatographic peak areas of compounds. Univariate analysis was performed on the chromatographic peak areas of compounds in the exhaled breath samples from patients with diabetic complications and those with diabetes to screen out compounds that showed significant differences between the two groups.

[0119] 1.1 Exhaled breath sampling

[0120] Based on the Fowler model, the exhaled CO2 concentration was increased by 50% (C50) to distinguish between anatomical dead space exhaled air and alveolar exhaled air. When the exhaled air concentration was below C50, the collection device did not collect this portion of exhaled air; when the exhaled air concentration was above C50, the collection device collected this portion of exhaled air in a Teflon bag, with an exhaled air collection volume of 3L.

[0121] 1.2 Exhaled Breath Detection

[0122] Exhaled air was transferred to an adsorption tube containing Tenax TA adsorbent for pre-concentration, and then detected by gas chromatography-mass spectrometry via thermal desorption injection to obtain the detection results.

[0123] 1.3 Analysis of Test Results

[0124] Qualitative and quantitative analyses were performed on the detection results obtained in step 1.2 based on the retention times of the NIST14 database and standard reference materials.

[0125] 1.4 Screening of Volatile Organic Compounds as Markers

[0126] The chromatographic peak areas of nine compounds identified in patients with diabetic complications and those with diabetes were compared using a t-test. Compounds showing significant differences between the two groups (P<0.05) were identified as biomarkers. Four volatile organic compounds—isoprene, acetone, isopropanol, and tetradecane—were screened out. The names of the compounds in exhaled breath, their chromatographic peak areas, and the results of their t-test analysis are shown in Table 1.

[0127] Table 1. Comparison of chromatographic peak areas of exhaled compounds between the diabetes group and the diabetic complication group.

[0128]

[0129] The results in Table 1 show significant differences in the composition of exhaled breath between the diabetes group and the diabetes complication group. Specifically, the levels of isoprene, acetone, and isopropanol in the exhaled breath of the diabetes group were significantly higher than those of the diabetes complication group (P < 0.05), while the level of tetradecane in the exhaled breath of the diabetes group was significantly lower than that of the diabetes complication group (P < 0.05). In the above comparisons, the P values ​​for isoprene, acetone, and tetradecane were all less than 0.001, and the P value for isopropanol was less than 0.01. In contrast, there were no significant differences in benzene, toluene, ethylbenzene, p-xylene, and undecane between the diabetes group and the diabetes complication group. This invention uses a Gini coefficient-based method to evaluate the importance of features in a decision tree model. Four biomarkers—isoprene, acetone, isopropanol, and tetradecane—were used as features to construct the model.

[0130] Example 2

[0131] In this embodiment, a sample dataset is constructed based on the detection levels of biomarkers in the subjects of Example 1, and a decision tree model and a support vector machine model are used to construct a model for identifying diabetic complications. This includes the following:

[0132] (1) Decision tree model:

[0133] a) A decision tree model was used to learn the training set, resulting in a model for identifying diabetic complications. (See model diagram below.) Figure 2 .

[0134] The following command line is used: `tree = DecisionTreeClassifier(random_state = 0)` and `tree.fit(X_train, y_train)`, where `tree` represents the modeling formula, `X_train` and `y_train` represent the training set data and training set labels. The dataset is divided into a test set (20%) and a training set (80%).

[0135] b) The test set was substituted into the diabetes complication identification model for evaluation.

[0136] Use the following command line:

[0137] y_pred = tree.predict(X_test), y_true = y_test, conf_matrix = confusion_matrix(y_true, y_pred) where tree.predict represents the prediction model, X_test represents the test set data, y_pred represents the prediction result of the test set, and y_test represents the true result of the test set.

[0138] (2) Support Vector Machine Model:

[0139] a) A support vector machine model was used to learn the training set to obtain a model for recognizing diabetic complications.

[0140] The following command line is used: `svm_classifier = SVC(kernel='linear', C=1.0) svm_classifier.fit(X_train, y_train)`, where `svm_classifier` represents the modeling formula, `X_train` and `y_train` represent the training set data and training set labels. The dataset is divided into 20% for the test set and 80% for the training set.

[0141] b) The test set was substituted into the diabetes complication identification model for evaluation.

[0142] Use the following command line:

[0143] y_pred = svm_classifier.predict(X_test) cm = confusion_matrix(y_test, y_pred) where svm_classifier.predict represents the prediction model, confusion_matrix(y_test, y_pred) represents the confusion matrix formula, X_test represents the test set data, y_pred represents the prediction result of the test set, and y_test represents the actual result of the test set.

[0144] (3) The model for identifying diabetic complications was evaluated using data from the test set. The evaluation metrics included sensitivity, specificity, and accuracy. The results are shown in Table 2.

[0145] Sensitivity: The proportion of subjects with diabetic complications who were correctly identified out of all subjects with diabetic complications.

[0146] Specificity: The proportion of correctly identified diabetic patients out of all diabetic patients.

[0147] Accuracy: The percentage of subjects correctly identified as having diabetic complications or diabetes patients out of all subjects.

[0148] Table 2. Results of classification of diabetic patients and patients with diabetic complications using the decision tree model and support vector machine model.

[0149]

[0150] Table 2 shows that the decision tree model based on the combination of isoprene + acetone + isopropanol + tetradecane achieved an accuracy of 87.5%, a sensitivity of 100%, and a specificity of 80% in distinguishing between diabetic patients and diabetic complications. In contrast, the support vector machine model based on the same combination achieved an accuracy of 87.5%, a sensitivity of 80%, and a specificity of 100% in distinguishing between diabetic patients and diabetic complications.

[0151] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any form or substance. It should be noted that those skilled in the art can make various improvements and additions without departing from the method of the present invention, and these improvements and additions should also be considered within the scope of protection of the present invention. Any modifications, alterations, and equivalent changes made by those skilled in the art based on the above-disclosed technical content without departing from the spirit and scope of the present invention are equivalent embodiments of the present invention. Furthermore, any modifications, alterations, and evolutions made to the above embodiments based on the essential technology of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A combination of biomarkers for identifying and / or predicting diabetic complications, characterized in that, The marker is selected from one or more of isoprene, acetone, isopropanol, benzene, toluene, ethylbenzene, p-xylene, undecane, and tetradecane.

2. The marker as described in claim 1, characterized in that, Includes at least one of the following technical features: 1) The marker is selected from a combination of isoprene and acetone; 2) The marker is selected from a combination of isoprene and isopropanol; 3) The marker is selected from a combination of isoprene and tetradecane; 4) The marker is selected from a combination of acetone and isopropanol; 5) The marker is selected from a combination of acetone and tetradecane; 6) The marker is selected from a combination of isopropanol and tetradecane; 7) The marker is selected from a combination of isoprene, acetone, and isopropanol; 8) The marker is selected from a combination of isoprene, acetone, and tetradecane; 9) The marker is selected from a combination of isoprene, isopropanol, and tetradecane; 10) The marker is selected from a combination of isopropanol, acetone, and tetradecane; 11) The marker is selected from a combination of isoprene, acetone, isopropanol and tetradecane; 12) The marker is selected from a combination of isoprene, acetone, isopropanol, tetradecane, and benzene; 13) The marker is selected from a combination of isoprene, acetone, isopropanol, tetradecane, benzene, and ethylbenzene; 14) The marker is selected from a combination of isoprene, acetone, isopropanol, tetradecane, benzene, ethylbenzene and p-xylene; 15) The marker is selected from a combination of isoprene, acetone, isopropanol, tetradecane, benzene, ethylbenzene, p-xylene, and undecane; 16) The marker is selected from a combination of isoprene, acetone, isopropanol, tetradecane, toluene, ethylbenzene, p-xylene, benzene and undecane.

3. Use of a substance that detects the biomarker as described in claim 1 or 2 in the preparation of a product for identifying and / or predicting diabetic complications, preferably, the substance that detects the biomarker comprises a substance that detects the volume or mass concentration of the biomarker.

4. A product for in vitro identification and / or prediction of diabetic complications, characterized in that, The product includes the substance used for detecting markers as described in claim 3.

5. Use of the biomarker as described in claim 1 or 2 or the product as described in claim 4 in screening for medications to treat and / or prevent complications of diabetes.

6. A method for constructing a model for identifying diabetic complications, characterized in that, The construction method includes the following: 1) Construct a sample dataset based on the detection levels of the biomarkers as described in claim 1 in the subject samples, wherein the subjects include patients diagnosed with diabetic complications and diabetic patients; 2) The sample dataset is randomly divided into a test set and a training set. The training set is learned using machine learning methods to obtain the diabetes complication prediction model. Preferably, the sample is selected from exhaled breath; And / or, the machine learning method is selected from at least one of the following: random forest algorithm, support vector machine algorithm, decision tree algorithm, K-nearest neighbor algorithm, logistic regression algorithm, and neural network algorithm; And / or, the split ratio of the test set to the training set is 1:(1~5); And / or, the construction method further includes testing the diabetes complication prediction model using a test set.

7. A screening method for diabetic complications, characterized in that, Includes the following steps: S1. Obtain the detection data of the biomarker as described in claim 1 in the sample of the subject; S2. The detection data is processed using the diabetes complication prediction model obtained by the construction method described in claim 6 to output the diagnosis results of diabetes complications.

8. A device for diagnosing diabetic complications, characterized in that, The device includes: The data module is used to acquire the detection data of the biomarker as described in claim 1 in the sample of the subject; The evaluation module processes the detected data using the diabetes complication prediction model obtained by the construction method described in claim 6, in order to output the diagnosis results of diabetes complications.

9. A device comprising a processor and a memory, the memory being used to store a computer program, characterized in that, The processor is configured to execute a computer program stored in the memory to cause the device to perform the screening method as described in claim 7 or the construction method as described in claim 6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the screening method as described in claim 7 or the construction method as described in claim 6.

Images