Method and system for screening risk of obese children in early stage of diabetes mellitus based on mid-infrared spectrum

Abstract

The invention discloses a mid-infrared spectrum-based obese infant early diabetes risk screening method and system, and belongs to the technical field of combination of medical detection and infrared spectrum, the mid-infrared spectrum-based obese infant early diabetes risk screening system comprises a micro-fluidic chip used for performing serum separation and protein enrichment deposition on dropped tissue peripheral blood of a patient, serum samples rich in different proteins are obtained; the mid-infrared spectrometer is used for carrying out in-situ mid-infrared spectrum acquisition on the blood sample to obtain spectrum data; and the risk screening module is used for preprocessing the spectral data, inputting the preprocessed spectral data into a risk prediction model to carry out pre-diabetic risk prediction on the obese children, and carrying out pre-diabetic risk screening according to a risk prediction probability value. The technical scheme is clear in structure, standard in step and suitable for large-scale children health screening and basic-level chronic disease management.

Description

Technical Field

[0001] The present invention belongs to the technical field of the combination of medical detection and infrared spectroscopy, and particularly relates to a method and system for screening the risk of prediabetes in obese children based on mid-infrared spectroscopy. Background Art

[0002] With the changes in lifestyle and dietary structure, the incidence of obesity in children and adolescents has increased significantly, becoming a widely concerned public health issue globally. A large number of studies have confirmed that childhood obesity not only affects their normal growth and development and mental health, but is also closely related to various metabolic diseases, especially type 2 diabetes mellitus (T2DM). Moreover, T2DM shows an earlier onset trend in the pediatric population, and the detection rate of the abnormal glucose regulation state (i.e., prediabetes) in the pre-stage among obese children has also increased year by year.

[0003] Currently, the main methods used clinically for screening prediabetes include biochemical indicators such as fasting plasma glucose (FPG), oral glucose tolerance test (OGTT), glycated hemoglobin (HbA1c), fasting insulin, and insulin resistance index (HOMA-IR). Although these methods have high diagnostic efficacy in the adult population, there are still many limitations in the pediatric population. On the one hand, most of these detection methods are invasive blood sampling operations, with poor compliance in children and strong dependence on equipment and environment, which is not conducive to large-scale population screening; on the other hand, traditional single biochemical indicators are difficult to comprehensively reflect the complex metabolic characteristics of prediabetes, prone to missed detection or misjudgment, reducing the sensitivity and specificity of screening.

[0004] With the development of metabolomics and non-invasive detection technologies, the research on infrared spectroscopy, especially mid-infrared spectroscopy (MIR), in early disease screening and identification of metabolic abnormalities has gradually increased. Mid-infrared spectroscopy can comprehensively analyze various metabolic components such as proteins, lipids, carbohydrates, and nucleic acids in a sample by detecting the characteristic absorption peaks generated by the vibration of intermolecular chemical bonds (such as C–H, C=O, N–H, etc.), generating a highly specific infrared metabolic fingerprint. MIR has the advantages of fast detection speed, small sample consumption, good result repeatability, and low cost, and is theoretically very suitable for large-scale screening and long-term health management of children.

[0005] However, the current practical application of infrared spectroscopy in the field of clinical screening still faces certain technical challenges: First, the traditional detection process often requires centrifuging and layering venous blood, and repeated transfers. The process is cumbersome and has high requirements for operation, making it unsuitable for rapid screening at the grass-roots level or on-site. Second, infrared signals are easily interfered by moisture, and the standards for sample pretreatment are not unified, resulting in large variability in spectral data among different samples, which limits the generalization ability of the model. Moreover, infrared data has a high dimension and complex structure, posing high requirements for modeling methods, dimensionality reduction strategies, and the ability to interpret biomarkers.

[0006] In summary, in view of the urgent need for early screening of children with obesity-related pre-T2DM, combined with the technical potential of mid-infrared spectroscopy detection and the feasibility of rapid processing procedures, the development of a pre-diabetes risk screening program for obese children combined with mid-infrared spectroscopy has significant clinical value and industrialization prospects. Summary of the Invention

[0007] In view of the above, the purpose of the present invention is to provide a method for screening the pre-diabetes risk of obese children based on mid-infrared spectroscopy. Through a rapid and minimally invasive fingertip blood collection method, combined with a serum classification and processing process based on a microfluidic chip, and then combined with in-situ serum infrared spectroscopy analysis and multi-dimensional data modeling, it realizes the early identification and typing of the pre-diabetes state of obese children. This method has a clear structure and standardized steps, and is suitable for large-scale children's health screening and grass-roots chronic disease management.

[0008] To achieve the above invention purpose, an embodiment provides a system for screening the pre-diabetes risk of obese children based on mid-infrared spectroscopy, including the following steps: A microfluidic chip, which is used to separate serum and enrich and deposit proteins from the tissue peripheral blood of the patient dropped in, to obtain serum samples rich in different proteins; A mid-infrared spectrometer, which is used to collect in-situ mid-infrared spectra of the serum samples rich in different proteins to obtain spectral data; A risk screening module, which is used to preprocess the spectral data and input it into a risk prediction model to predict the pre-diabetes risk of obese children, and perform pre-diabetes risk screening based on the risk prediction probability value.

[0009] Preferably, the microfluidic chip includes a micron column array module, a tapered diversion channel, and a serum deposition area connected in sequence. The serum obtained by separating cells through the micron column array module from the blood flows into the tapered diversion channel, and the streamline compression of the serum is realized through a set tapered structure in the tapered diversion channel to control the flow rates of different proteins in the serum. Proteins with different flow rates are enriched and deposited in the serum deposition area to obtain serum samples rich in different proteins.

[0010] Preferably, the length of the micro-column array module is set to 100 - 300 μm, where the interval between the micro-columns is 4 - 8 μm, and further preferably 6 μm. The micro-column array is an array formed by arranging micro-columns in a periodic pattern, which can perform deterministic lateral displacement of blood, precisely exclude cell components such as red blood cells and white blood cells, and the remaining serum is automatically introduced into the tapered diversion channel. Since both the length of the micro-column array and the interval between the micro-columns will affect the effect of cell separation, through experimental exploration, it is found that when the length of the micro-column array module is set to 100 - 300 μm and the interval between the micro-columns is 4 - 8 μm, red blood cells and white blood cell components with a diameter of more than 6 μm can be effectively screened out and deflected, while particles smaller than the gap of the micro-column array (such as serum proteins, exosomes, etc.) will move straight forward, realizing efficient serum extraction. The preferred short channel size reduces pressure loss and processing difficulty, enabling blood to pass through the micro-column array in just a few seconds, and also reducing the cost of micro-nano processing.

[0011] Preferably, the length of the tapered structure in the tapered diversion channel is set to 2000 - 3000 μm, and the reduction angle of the taper is set to 5.7 - 8.5°, and further preferably the length is set to 2500 μm and the reduction angle is set to 6.8°. The tapered structure in the tapered diversion channel is actually a tapered structure with a long inlet, a short outlet and symmetric along the flow direction. The reduction degree of the outlet relative to the inlet is characterized by the reduction angle. It is found that both the length of the tapered structure and the reduction angle will affect the flow rate of different proteins. Through experimental exploration, it is found that when the length is 2500 μm and the reduction angle of the taper is 6.8°, while stabilizing the flow rate, it can ensure that the serum rich in metabolites is evenly deposited on the surface of the infrared transparent detection window, realizing in-situ retention of metabolic components such as target proteins. If the angle is too large, the channel will narrow sharply, easily generating vortices or boundary separation at the turning point, and the serum protein cannot adhere evenly; if the angle is too small, the channel contracts too slowly, the streamline focusing is insufficient, the concentration of the deposited serum protein is not concentrated enough, which will lengthen the chip size and extend the flow time.

[0012] The serum separation and protein enrichment and deposition process achieved by the above-mentioned microfluidic chip is completed automatically without centrifugation and without reagents, taking about 3 - 5 minutes, effectively solving the technical problem of the cumbersome process existing in the traditional detection process, and the microfluidic chip is for single use, effectively avoiding cross-contamination and residual interference.

[0013] Preferably, the deposition area of the microfluidic chip is exactly the surface of the detection window of the infrared spectrum, realizing in-situ retention and in-situ mid-infrared scanning of the serum sample in the deposition area, and the wave number range is 4000 - 400 cm -1. The mode is transmission mode, reflection mode or attenuated total reflection mode. Set the spectral scanning time to be within 1 minute, and collect each sample 3 times repeatedly to enhance the signal-to-noise ratio.

[0014] Since the detection window surface of the infrared spectrum is exactly the deposition area of the microfluidic chip, after the deposition is completed, the infrared spectrum can be scanned directly without transferring the sample or eluting treatment.

[0015] Preferably, preprocess the spectral data, including: Normalize the spectral data, including baseline correction, smoothing filter, first-order or second-order derivative transformation, and standard normal variate transformation; Remove the water peak and noise section from the normalized spectral data, and intercept the key band for subsequent pre-diabetes risk prediction. Among them, the lipid region corresponds to the wave number of 2800–3000 cm⁻¹, and the protein region corresponds to the wave number of 1500–1700 cm⁻¹. Based on this, intercept the key band to be detected.

[0016] Preferably, the risk prediction model is constructed based on supervised learning. Using the preprocessed spectral data as input and the pre-diabetes risk status corresponding to the spectral data as the output supervised label, realize the supervised learning of the network structure to construct the risk prediction model, where the network structure includes PLS-DA, random forest, or SVM.

[0017] Predict the pre-diabetes risk of obese children through big data analysis based on the network structure. The pre-diabetes risk of obese children here is divided into low risk, medium risk and high risk. This can ensure the accuracy and speed of risk prediction.

[0018] Preferably, screen the pre-diabetes risk according to the risk prediction probability value, including: when the risk prediction probability value is greater than the set risk warning threshold, screen high-risk patients. The risk warning threshold is obtained based on experience and generally takes a value of 0.6. Generally screen high-risk patients, so as to give guiding treatment suggestions in time.

[0019] To achieve the above invention purpose, the embodiment also provides a method for screening the pre-diabetes risk of obese children based on mid-infrared spectrum. The method uses the above system and includes the following steps: Step 1, separate the serum and enrich and deposit the protein from the peripheral blood of the patient's tissue dropped into by the microfluidic chip to obtain a serum sample rich in different proteins; Step 2, use a mid-infrared spectrometer to collect the mid-infrared spectrum in situ of the blood sample to obtain spectral data; Step 3: After preprocessing the spectral data through the risk screening module, input it into the risk prediction model for predicting the pre-diabetes risk of obese children, and conduct pre-diabetes risk screening based on the risk prediction probability value.

[0020] For the entire screening method, each step is executed sequentially in time, with a compact process. All steps can be completed within 10 minutes. Moreover, the microfluidic separation in Step 1 and the mid-infrared spectrum collection in Step 2 are the core linkage steps, ensuring the rapid detection of serum samples on the machine. The data processing and risk prediction output in Step 3 are automated, guaranteeing the consistency and accuracy of data processing and prediction.

[0021] Compared with the prior art, the beneficial effects of the present invention at least include: The present invention uses the patient's tissue peripheral blood as the test sample, with extremely low requirements for the test sample. The minimally invasive blood collection from tissue peripherals such as fingertips (only 2–5 μL of blood collection is required) is more suitable for large-scale screening of the child population. The present invention uses mid-infrared spectroscopy to obtain the overall metabolic protein map, avoiding the error of single-index judgment. Moreover, in-situ and timely testing is adopted during the testing process to avoid the problem of infrared signal interference by moisture.

[0022] The system of the present invention can also be extended to portable screening devices, suitable for hospital, school, and home scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0024] Figure 1 It is a schematic structural diagram of a pre-diabetes risk screening system for obese children based on mid-infrared spectroscopy provided by the embodiment; Figure 2 It is a schematic structural diagram of a microfluidic chip provided by the embodiment; Figure 3 It is a flowchart of a pre-diabetes risk screening method for obese children based on mid-infrared spectroscopy provided by the embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] To make the purpose, technical solutions, and advantages of the present invention clearer, the following will further elaborate on the present invention in detail with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not limit the protection scope of the present invention.

[0026] To solve the technical problems existing in the background art, a rapid detection process more suitable for the pre-diabetes risk of obese children on-site is urgently needed. Therefore, a pre-diabetes risk screening solution for obese children based on mid-infrared spectroscopy of the present invention is proposed. Whole blood is collected by means of micro-sampling, and after the serum is rapidly separated by a disposable microfluidic chip, infrared spectrum scanning is immediately carried out without chemical reagents and without markers. Combined with standardized data preprocessing and multi-omics machine learning algorithms, an integrated risk prediction model is constructed, and the detection and preliminary risk assessment can be completed within 10 minutes, providing data support for clinical decision-making. The rapid detection process of this technical solution will greatly simplify the operation steps, improve the screening efficiency, and increase the acceptance of children and their parents.

[0027] As Figure 1 shown, a pre-diabetes risk screening system for obese children based on mid-infrared spectroscopy provided by an embodiment includes a microfluidic chip, a mid-infrared spectrometer, and a risk screening module.

[0028] Among them, the microfluidic chip is mainly used to separate serum and enrich and deposit proteins from the tissue peripheral blood of the dripping patient to obtain a serum sample rich in different proteins. Specifically, a micro-sampling pen is used to collect peripheral blood from the patient's tissue, and the blood collection volume only needs 2–5 μL. The collected blood is immediately transferred into or directly dropped into the microfluidic chip for subsequent rapid processing. When collecting blood, the patient is required to be in a fasting state (8 hours) to avoid interference with the results caused by blood sugar fluctuations.

[0029] As Figure 2 shown, the microfluidic chip includes a micron column array module, a tapered diversion channel, and a serum deposition area connected in sequence. Among them, the serum obtained by separating cells through the micron column array module flows into the tapered diversion channel, and the streamline compression of the serum is realized through the set tapered structure in the tapered diversion channel to control the flow rate of different proteins in the serum. Proteins with different flow rates are enriched and deposited in the deposition area to obtain a serum sample rich in different proteins.

[0030] Considering the time and effect of cell filtration, in the example, the length of the micron column array module is set to 200 μm, the interval between micron columns is 6 μm, and the diameter of each micro-nano column is 20 μm. Blood passing through the micron column array of this size can accurately deflect and exclude cell components such as red blood cells and white blood cells, and the remaining serum is automatically introduced into the tapered diversion channel to gather the serum streamline.

[0031] The tapered diversion channel is mainly used to divert serum by controlling the flow rates of different proteins in the serum, so as to achieve the zonal enrichment deposition of different proteins in the deposition area. Here, the different proteins serve as different metabolic components for subsequent spectral acquisition and risk prediction. The tapered diversion channel directly affects the deposition result. Specifically, the length of the conical structure is set to 2500 μm, the reduction angle of the cone is set to 6.8°, the inlet width is 800 μm, and the outlet width is 200 μm. The tapered diversion channel with such dimensions can better adapt to the streamline compression of serum and the focusing of protein distribution.

[0032] The deposition area of the microfluidic chip is exactly the surface of the detection window of the infrared spectrum, that is, the serum flows through the surface with an infrared spectrum window (such as a silicon wafer or ZnSe) at the bottom, which directly serves as the deposition area. Different target proteins in the serum are naturally deposited here, and then the in-situ retention and in-situ mid-infrared scanning of the serum sample in the deposition area are directly realized.

[0033] In the embodiment, the mid-infrared spectrometer has high resolution and sensitivity, and can capture the information of various metabolic markers in the serum sample, such as the infrared absorption characteristics of molecules such as lipids, proteins, and carbohydrates. Specifically, it can be a mid-infrared Fourier transform infrared spectrometer (FTIR), which performs mid-infrared scanning on the serum sample rich in different proteins in the deposition area. Among them, the wavenumber range is 4000–400 cm - ¹, and the mode is transmission mode, reflection mode or attenuated total reflection mode. The spectral scanning time is controlled within 1 minute, and the single sample is collected 3 times repeatedly to enhance the signal-to-noise ratio.

[0034] In the embodiment, the risk screening module, as the algorithm detection part, is mainly used to preprocess the spectral data, and then input the preprocessed spectral data into the risk prediction model for the risk prediction of pre-diabetic obesity in children, and perform pre-diabetic risk screening according to the risk prediction probability value.

[0035] Among them, the preprocessing process includes normalization, removing water peaks and noise sections, and intercepting key bands, so as to extract the characteristic signals related to pre-diabetes from the complex infrared, and use the risk prediction model constructed by machine learning algorithms for risk assessment, providing the risk prediction probability value of pre-diabetes as the risk score for clinical use. This process is not only efficient and automated, but also can quickly provide results, facilitating clinical decision-making and timely intervention. After obtaining the risk score, a risk warning threshold is introduced to screen high-risk patients.

[0036] The pre-diabetes risk screening system for obese children based on mid-infrared spectroscopy provided by the above embodiments can achieve early identification and typing of pre-diabetes individuals among obese children. Specifically, by combining microfluidic technology and mid-infrared spectroscopy analysis, an efficient, minimally invasive, and rapid detection method is adopted, which can not only effectively reduce the disadvantages of traditional screening methods but also improve the efficiency and accuracy of pre-diabetes screening. In short, it has the characteristics of high sensitivity, non-invasiveness, and strong adaptability.

[0037] As Figure 3 shown, the embodiment also provides a pre-diabetes risk screening method for obese children based on mid-infrared spectroscopy, including the following steps: S1, Serum separation and protein enrichment deposition are performed on the capillary blood of the patient's tissue dropped into the microfluidic chip to obtain serum samples rich in different proteins; S2, Mid-infrared spectroscopy of the serum samples rich in different proteins is collected in situ by a mid-infrared spectrometer to obtain spectral data; S3, After preprocessing the spectral data by the risk screening module, it is input into the risk prediction model for pre-diabetes risk prediction of obese children, and pre-diabetes risk screening is performed based on the risk prediction probability value.

[0038] An actual application example of the above screening method is: The test subjects are obese children aged 6 - 14 years in a fasting state. 2 - 5 μL of capillary blood is taken with a disposable blood lancet and directly dropped into the inlet of the microfluidic chip. The microcolumn array in the chip effectively blocks components larger than 6 μm such as red blood cells and white blood cells according to particle size. The serum part travels forward along the main flow direction and enters the tapered diversion channel to focus the serum streamline and improve the adhesion efficiency of proteins at the detection window. Finally, the serum flows through the deposition area with an infrared transparent window (such as Si or ZnSe substrate) at the bottom. Due to different flow rates, protein molecules of different masses will stay in different positions of the infrared bottom layer in layers, which is conducive to capturing specific protein target proteins and forming a measurable layer. The whole process does not require centrifugation, does not require chemical reagents, is automatically completed in a closed manner, and takes about 5 minutes.

[0039] The serum sample of the measurable layer is directly scanned by the FTIR instrument without dilution or pretreatment. The instrument is set with a wavenumber range of 4000 - 400 cm - ⁻¹, a resolution of 4 cm - ⁻¹, and the integration times are 32 times.

[0040] The collected mid-infrared spectrum is denoised, SNV standardized, baseline corrected, and derivative transformed, and then the key bands are retained (such as the lipid region at 2850 - 2960 cm - ⁻¹, 1540 - 1650 cm -¹ protein region), and the whole process is automated using Python and Matlab.

[0041] Using algorithms such as PLS-DA and random forest, a risk prediction model of the preprocessed spectral data and the clinical risk diagnosis status is constructed. The model outputs the predicted probability value of prediabetes, and a high-risk warning threshold of 0.6 is set to screen the predicted probability value to obtain high-risk patients.

[0042] The above screening method also has strong scalability and can be applied to the health screening of large-scale children, especially in public health fields such as schools and communities. Since this method is non-invasive, label-free, and has low cost and high operational convenience, it will play an important role in primary healthcare, children's health management, and chronic disease prevention. Through the application of the solution of the present invention, the risk of prediabetes in the obese children group can be detected early, providing a precise health management plan and promoting the prevention and control of diabetes and other metabolic diseases.

[0043] The specific embodiments described above have detailed the technical solutions and beneficial effects of the present invention. It should be understood that the above is only the most preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, supplements, equivalent replacements, etc. made within the scope of the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A pre-diabetes risk screening system for obese children based on mid-infrared spectroscopy, characterized in that, Comprising: A microfluidic chip for separating serum and enriching and depositing proteins from the peripheral blood of a patient's tissue dropped therein to obtain serum samples rich in different proteins; A mid-infrared spectrometer for in-situ mid-infrared spectral acquisition of the serum samples rich in different proteins to obtain spectral data; A risk screening module for preprocessing the spectral data and then inputting it into a risk prediction model for pre-diabetes risk prediction of obese children, and performing pre-diabetes risk screening based on the risk prediction probability value.

2. The pre-diabetes risk screening system for obese children based on mid-infrared spectroscopy according to claim 1, wherein The microfluidic chip includes a micron column array module, a tapered diversion channel, and a serum deposition area connected in sequence. The serum obtained after separating cells by the micron column array module flows into the tapered diversion channel, and the streamline compression of the serum is realized through a set conical structure in the tapered diversion channel to control the flow rates of different proteins in the serum. Proteins with different flow rates are enriched and deposited in the serum deposition area to obtain serum samples rich in different proteins.

3. The pre-diabetes risk screening system for obese children based on mid-infrared spectroscopy according to claim 2, wherein The length of the micron column array module is set to 100 - 300 μm, and the interval between micron columns is 4 - 8 μm.

4. The pre-diabetes risk screening system for obese children based on mid-infrared spectroscopy according to claim 2, wherein The length of the conical structure in the tapered diversion channel is set to 2000 - 3000 μm, and the reduction angle of the cone is set to 5.7 - 8.5°.

5. The pre-diabetes risk screening system for obese children based on mid-infrared spectroscopy according to claim 1 or 2, characterized in that, The deposition area of the microfluidic chip is exactly the surface of the detection window of the infrared spectrum, enabling in-situ retention and in-situ mid-infrared scanning of the serum sample in the deposition area, with a wavenumber range of 4000–400 cm - ⁻¹, and the mode used is the transmission mode, reflection mode or attenuated total reflection mode.

6. The pre-diabetes risk screening system for obese children based on mid-infrared spectroscopy according to claim 1, wherein, Preprocessing the spectral data includes: Performing normalization processing on the spectral data, including baseline correction, smoothing filtering, first-order or second-order derivative conversion, and standard normal variate transformation; Removing water peaks and noise sections from the spectrally data after normalization processing, and intercepting the key bands for subsequent pre-diabetes risk prediction.

7. The pre-diabetes risk screening system for obese children based on mid-infrared spectroscopy according to claim 1, characterized in that, The risk prediction model is constructed based on supervised learning, using the preprocessed spectral data as input and the pre-diabetes risk status corresponding to the spectral data as the output supervised label to realize the supervised learning of the network structure for constructing the risk prediction model. Among them, the model includes PLS-DA, random forest, or SVM.

8. The pre-diabetes risk screening system for obese children based on mid-infrared spectroscopy according to claim 1, characterized in that, Performing pre-diabetes risk screening based on the risk prediction probability value includes: When the risk prediction probability value is greater than the set risk warning threshold, screening of high-risk patients is realized.

9. A method for screening the risk of prediabetes in obese children based on mid-infrared spectroscopy, characterized in that, The method uses the system according to any one of claims 1 - 8, and includes the following steps: Separating serum and enriching and depositing proteins from the peripheral blood of the patient's tissue dropped through the microfluidic chip to obtain serum samples rich in different proteins; Performing in-situ mid-infrared spectral acquisition on the blood sample rich in different proteins by using a mid-infrared spectrometer to obtain spectral data; Preprocessing the spectral data through the risk screening module and then inputting it into the risk prediction model for pre-diabetes risk prediction of obese children, and performing pre-diabetes risk screening based on the risk prediction probability value.

Images