Anti-interference detection method and system based on microfluidic paper-based sensor

Abstract

The invention provides an anti-interference detection method and system based on a micro-fluidic paper-based sensor, belongs to the field of intelligent medical treatment, and solves the problems that the current detection mainly depends on an immunofluorescence luminescence method, the cost of required reagents and detection machines is relatively high, and the detection cost is relatively low. And the existing paper-based microfluidic sensor capable of carrying out combined detection of biomarkers also has a certain detection precision problem. The method comprises the following steps: acquiring an electrochemical signal of a reference detection channel while acquiring an electrochemical impedance spectrum and / or volt-ampere signal by using each target marker detection channel, a pseudo target molecule which is not specifically combined with a target marker in a clinical sample in a detection concentration range is arranged on the surface of a working electrode of the reference detection channel; the electrochemical parameters of the reference detection channels serve as background items to be input into a multi-channel signal processing model, and combined deconvolution and background deduction are conducted on measurement signals of all the target marker detection channels so as to solve the corrected concentration value of at least one target marker.

Description

Technical Field

[0002] This application relates to the field of smart medical technology, and in particular to an anti-interference detection method and system based on a microfluidic paper-based sensor. Background Technology

[0004] The combined detection of three cardiovascular biomarkers—high-sensitivity troponin, D-dimer, and natriuretic peptide—can provide a multi-faceted and multi-dimensional initial assessment of patients with fatal chest pain, assisting in preliminary diagnosis and exclusion, identifying high-risk patients, performing risk stratification, and formulating rational treatment decisions. Through a series of observations, it can be further used for evaluating treatment effectiveness and adjusting treatment, as well as predicting prognosis. Currently, their detection mainly relies on immunofluorescence assays, which require expensive reagents and testing equipment. Furthermore, existing paper-based microfluidic sensors capable of combined biomarker detection also have certain accuracy issues. Summary of the Invention

[0006] This application provides an anti-interference detection method and system based on a microfluidic paper-based sensor, which can solve the problems that the current detection mainly relies on immunofluorescence luminescence method, which requires high-cost reagents and detection equipment, and the existing paper-based microfluidic sensors that can perform joint detection of biomarkers also have certain detection accuracy problems.

[0007] The first aspect of this application provides an anti-interference detection method based on a microfluidic paper-based sensor, comprising:

[0008] While acquiring electrochemical impedance spectroscopy and / or voltammetric signals using each target biomarker detection channel, an electrochemical signal of a reference detection channel is also acquired. The reference detection channel is located in an adjacent region to at least one target biomarker detection channel of the sensor. The working electrode surface of the reference detection channel is provided with pseudo-target molecules that do not specifically bind to the target biomarkers in the clinical sample within the detection concentration range.

[0009] The electrochemical parameters of the reference detection channel are used as background terms and input into the multi-channel signal processing model. The measurement signals of each target biomarker detection channel are jointly deconvolved and background subtracted to solve for the corrected concentration value of at least one target biomarker.

[0010] The detection result is determined based on the corrected concentration value.

[0011] Optionally, the structure and materials of the reference detection channel are the same as those of the target biomarker detection channel. The working electrode surface of the reference detection channel is immobilized with a pseudo-target molecule that is similar to the epitopes of at least one of high-sensitivity troponin, D-dimer, and natriuretic peptide, but does not specifically bind to the corresponding biomarker in the clinical sample. This allows the background electrochemical signal caused by non-specific protein adsorption and charge distribution in the complex matrix to be correlated between the reference detection channel and the target biomarker detection channel, serving as a background reference signal.

[0012] The method of jointly deconvolving and subtracting the measurement signals from each target marker detection channel to solve for the corrected concentration value of at least one target marker includes:

[0013] The electrochemical parameters of the reference detection channel are introduced as background terms into the multi-channel signal processing model. The measurement signals of each target marker detection channel are jointly deconvolved and background subtracted to solve for the corrected concentration value of at least one of high-sensitivity troponin, D-dimer and natriuretic peptide, thereby suppressing common-mode drift and inter-channel cross-interference caused by complex matrix.

[0014] Optionally, the pseudo-target molecule is a short peptide with site mutation, a conformationally locked nucleic acid aptamer, or a combination thereof, and the pseudo-target molecule maintains a spatial conformation and charge environment similar to that of the target marker recognition molecule.

[0015] Optionally, the liquid absorption end of the microfluidic paper-based sensor is provided with a composite liquid absorption pad, which includes a stacked fast liquid absorption layer and a variable resistance layer. The fast liquid absorption layer is made of high-porosity cellulose material to provide a reference liquid absorption driving force, and the variable resistance layer is a porous matrix with a porous network structure.

[0016] Optionally, the variable resistance layer is a composite material in which hydrophilic and hydrophobic microparticles are mixed in a porous matrix to form a wetting gradient. It is used to adjust the permeation resistance of the sample in the composite absorbent pad by the contact angle difference in different regions. When the sample hematocrit increases and the viscosity increases, the variable resistance layer is used to automatically improve the equivalent absorbent capacity by changing the wetting angle, so as to maintain the average flow rate and reaction residence time in the detection channel of each target marker within a preset range under different hematocrit conditions, and reduce the influence of inter-individual rheological differences on the corrected concentration value.

[0017] Optionally, the microfluidic paper-based sensor further includes a common working electrode. The surface of the working electrode is divided into multiple spatially separated recognition regions. Each recognition region is provided with specific recognition molecules for recognizing different target markers, so that multiple target marker detection channels share the same working electrode in physical structure, forming concentric annular regions or adjacent sector regions on a plane. The method further includes:

[0018] When acquiring the electrochemical impedance spectroscopy and / or voltammetric signal of each target biomarker detection channel, impedance test excitation is applied to the common working electrode at different frequency bands and / or voltammetric test waveforms with different pulse parameters are used to make the electron transfer kinetic response corresponding to each identification region have distinguishable characteristics in the frequency domain and / or waveform parameter domain, so as to separate and fit the electrochemical signals of different target biomarkers based on frequency band and / or waveform parameter multiplexing in the multi-channel signal processing model.

[0019] Optional, also includes:

[0020] A calibration sub-model was established for each identification region using a multi-channel signal processing model. Electrochemical parameters obtained under different frequency bands and / or different voltammetric test parameters were matched and fitted with the corresponding calibration sub-model to obtain the initial concentration estimate of each target marker.

[0021] By combining the electrochemical parameters of the reference detection channel with joint deconvolution and background subtraction, the corrected concentration values ​​of each target marker are determined.

[0022] A second aspect of this application provides an anti-interference detection system based on a microfluidic paper-based sensor, comprising:

[0023] The acquisition unit is used to acquire the electrochemical signal of a reference detection channel while acquiring electrochemical impedance spectroscopy and / or voltammetric signals using each target biomarker detection channel. The reference detection channel is disposed in an adjacent region of at least one target biomarker detection channel of the sensor. The working electrode surface of the reference detection channel is provided with pseudo-target molecules that do not specifically bind to the target biomarkers in the clinical sample within the detection concentration range.

[0024] The correction unit is used to input the electrochemical parameters of the reference detection channel as background terms into the multi-channel signal processing model, and to perform joint deconvolution and background subtraction on the measurement signals of each target marker detection channel to solve for the corrected concentration value of at least one target marker.

[0025] A determining unit is used to determine the detection result based on the corrected concentration value.

[0026] A third aspect of this application provides an electronic device, including a memory and a processor, wherein the processor is used to execute a computer program stored in the memory to implement the steps of the above-described anti-interference detection method based on a microfluidic paper-based sensor.

[0027] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described anti-interference detection method based on a microfluidic paper-based sensor.

[0028] In summary, the anti-interference detection method based on a microfluidic paper-based sensor provided in this application acquires the electrochemical signal of a reference detection channel while simultaneously acquiring electrochemical impedance spectroscopy and / or voltammetric signals using each target biomarker detection channel. The reference detection channel is located in an adjacent region to at least one target biomarker detection channel of the sensor. The working electrode surface of the reference detection channel is provided with pseudo-target molecules that do not specifically bind to the target biomarker in the clinical sample within the detection concentration range. The electrochemical parameters of the reference detection channel are input as background terms into a multi-channel signal processing model. Joint deconvolution and background subtraction are performed on the measurement signals of each target biomarker detection channel to solve for the corrected concentration value of at least one target biomarker. The detection result is determined based on the corrected concentration value. Therefore, by introducing a reference detection channel with identical structure but different molecular recognition characteristics onto a paper-based microfluidic sensor, and using the pseudo-target channel as a background signal observation window, and then combining it with a multi-channel signal processing model to perform joint deconvolution and background subtraction on the target channel signal, the corrected target biomarker concentration is obtained, and the detection result is output accordingly. This achieves highly stable and interference-resistant quantitative detection of cardiovascular biomarkers under complex clinical sample conditions. Compared with traditional single-channel, single-threshold interpretation paper-based electrochemical detection schemes, this scheme does not require additional complex hardware structures or operating steps. It significantly improves the resistance to matrix interference and cross-sample consistency mainly through structural and algorithm design, making it very suitable for rapid joint evaluation of multiple biomarkers in emergency chest pain scenarios. It helps reduce misjudgments caused by poor blood sample condition, environmental fluctuations, or minor electrode defects, and enhances the support of point-of-care testing for clinical decision-making.

[0029] Correspondingly, the systems, electronic devices, and computer-readable storage media provided in the embodiments of the present invention also have the above-mentioned technical effects. Attached Figure Description

[0031] Figure 1 A schematic flowchart illustrating a possible anti-interference detection method based on a microfluidic paper-based sensor provided in this application embodiment;

[0032] Figure 2 A schematic structural block diagram of a possible anti-interference detection system based on a microfluidic paper-based sensor provided in this application embodiment;

[0033] Figure 3 A schematic diagram of the hardware structure of a possible anti-interference detection system based on a microfluidic paper-based sensor provided in an embodiment of this application;

[0034] Figure 4 A schematic structural block diagram of a possible electronic device provided in an embodiment of this application;

[0035] Figure 5 This is a schematic structural block diagram of a possible computer-readable storage medium provided for embodiments of this application. Detailed Implementation

[0037] This application provides an anti-interference detection method and related equipment based on a microfluidic paper-based sensor, which can solve the problems that the current detection mainly relies on immunofluorescence luminescence, which requires high-cost reagents and detection machines, and the existing paper-based microfluidic sensors that can perform joint detection of biomarkers also have certain detection accuracy problems.

[0038] Please see Figure 1 The flowchart below illustrates an anti-interference detection method based on a microfluidic paper-based sensor, as provided in this application embodiment. Specifically, it may include:

[0039] S110-S130.

[0040] S110, while acquiring electrochemical impedance spectroscopy and / or voltammetric signals using each target biomarker detection channel, an electrochemical signal of a reference detection channel is acquired. The reference detection channel is located in an adjacent region to at least one target biomarker detection channel of the sensor. The working electrode surface of the reference detection channel is provided with pseudo-target molecules that do not specifically bind to the target biomarkers in the clinical sample within the detection concentration range.

[0041] S120, the electrochemical parameters of the reference detection channel are input into the multi-channel signal processing model as background terms, and the measurement signals of each target marker detection channel are jointly deconvolved and background subtracted to solve for the corrected concentration value of at least one target marker.

[0042] S130, determine the detection result based on the corrected concentration value.

[0043] For example, on the same microfluidic paper-based sensor, the target biomarker detection channel and the reference detection channel experience the same sample flow path, paper pore structure, temperature and humidity environment, and electrode material state. Therefore, background interference caused by the complex matrix, such as non-specific adsorption, conductivity changes, and double-layer capacitance changes, will be exhibited in a highly similar manner in both channels. Since spurious target molecules do not specifically bind to the target biomarker in the clinical sample within the detection concentration range, the reference detection channel mainly records the background electrochemical response brought about by the sample matrix itself, rather than the specific signal of the target biomarker. In specific implementation, the reference detection channel can be arranged in an adjacent area of ​​at least one target biomarker detection channel on the paper-based detection layer, so that the fluid distance between them is as short as possible, the paper batch is the same, the electrode printing process is consistent, and the anti-fouling layer material ratio is consistent, thereby ensuring that the background interference has good spatial correlation. On the working electrode surface of the reference detection channel, pseudo-target molecules are immobilized using the same immobilization strategy as the target channel. For example, short peptides with site mutations or conformationally locked nucleic acid aptamers can be selected to make their spatial conformation and charge environment similar to the real recognition molecule. However, by adjusting key amino acid residues or base sequences, their affinity for the target biomarker is extremely low, resulting in almost no binding within the actual detection concentration range. Thus, when complex blood samples migrate on the paper substrate, lipoproteins, albumin, various inorganic salts, and fine particles will generate similar non-specific signals in both the target and reference channels, while the additional signal generated by the binding of the target biomarker to the specific recognition molecule only appears in the target channel. This allows the system to obtain a pair of correlated signals simultaneously on the same chip: a mixed signal of background and target, and a pure background reference signal, laying the foundation for subsequent model-based separation of the true target biomarker response from the mixed signal. For example, in hyperlipidemic samples, chylomicrons can significantly alter the solution conductivity and increase electrode surface contamination. Without a reference channel, it is easy to mistake the impedance changes caused by lipemia for an increase in the target biomarker. However, with a reference channel, the impedance changes caused by lipemia will appear in both channels, but only a small amount of additional changes caused by the binding of the target biomarker will be superimposed on the target channel. This structure of common-mode and differential changes provides sufficient information for subsequent mathematical processing.

[0044] For example, the signal measured in the target biomarker detection channel can be considered as the sum of two parts: one part is the specific signal caused by the binding of the target biomarker to the specific recognition molecule, and the other part is the non-specific background signal caused by the complex matrix. The reference detection channel provides an observation that is highly correlated with the background of the target channel but basically contains no specific signal. Therefore, by establishing a multi-channel mathematical model, the observation of the target channel can be decomposed into background components and specific components, thereby obtaining the specific component corresponding to the target biomarker. In specific implementation, the impedance modulus, phase angle, peak current, peak shift, and other parameters of one or more characteristic frequency points in the electrochemical impedance spectroscopy can be selected to form a multidimensional feature vector. The feature vector of each target channel is recorded as the measurement vector, and the feature vector of the reference channel is recorded as the background vector. During the system calibration stage, a large number of tests are conducted using standard buffer, spiked plasma, samples with different degrees of lipemia and hemolysis, respectively. The relationship between the target channel signal and the background channel signal multiplied by the background coefficient plus the target concentration multiplied by the response coefficient is fitted. Mathematically, this can be understood as establishing a multivariate regression or matrix factorization model containing background and target terms for each target channel. During detection, the target channel signal and reference channel signal, simultaneously obtained on the same chip, are fed into a multi-channel signal processing model. The model first uses the reference channel signal to estimate the mapping of the corresponding background component in the target channel, then subtracts this background component from the total signal of the target channel. The remaining part is the specific signal component corresponding to the target biomarker. Based on this, the corrected concentration value of the target biomarker is calculated using a pre-calibrated response curve. Deconvolution here can be understood as the process of reverse-engineering the electrochemical response convolved with the background in the target channel back to the response solely caused by the target molecule. This significantly suppresses the influence of complex matrices, temperature and humidity fluctuations, and minor manufacturing differences between electrodes on the quantitative results, especially reducing the probability of false positives and false negatives in the low concentration range. For example, for the same concentration of high-sensitivity troponin, if one sample is clear plasma and the other is mildly hemolyzed and mildly lipemic, theoretically the specific signals are the same, but the background noise is very different. Traditional algorithms will mix the impedance changes caused by hemolysis and lipemia into the concentration calculation, which may result in the hemolyzed sample having an artificially high concentration. After introducing a reference channel and performing background subtraction through a multi-channel model, the background components of the two samples are first stripped away, and the retained specific components are closer to the true concentration. The final corrected concentration value will be more consistent, significantly improving reproducibility.

[0045] For example, after background subtraction and deconvolution processing, the obtained target biomarker concentration estimates are as close as possible to the actual physiological state. Using these corrected concentration values ​​as the basis for interpretation, a consistency evaluation consistent with or close to that of large laboratory analyzers can be reconstructed, achieving methodological alignment of the paper-based microfluidic platform in complex clinical samples. Corresponding diagnostic thresholds, risk stratification intervals, and dynamic monitoring thresholds can be established for high-sensitivity troponin, D-dimer, and natriuretic peptides, respectively. For example, acute myocardial infarction exclusion and diagnostic thresholds can be set for high-sensitivity troponin, acute pulmonary embolism exclusion thresholds can be set for D-dimer, and heart failure risk stratification thresholds can be set for natriuretic peptides, etc. These thresholds are fixed as parameters in the reading device. When the algorithm gives the corrected concentration value, the system automatically determines the interval in which the value falls and outputs textual detection conclusions based on multiple indicators. For example, if high-sensitivity troponin is below the exclusion threshold, acute myocardial infarction can be ruled out; if D-dimer is significantly elevated, acute pulmonary embolism or other hypercoagulable states should be suspected; if natriuretic peptide is significantly elevated, acute heart failure is possible, etc. More refined decision suggestions can also be given based on the combination of the three indicators. This method transforms a numerical value obtained through complex signal processing into a clinically understandable and usable conclusion. Furthermore, because the first two steps have minimized matrix interference, the reliability of the interpretation is significantly improved. For example, traditional immunochromatography often yields elevated D-dimer levels in highly inflammatory states, easily leading to numerous false positives for pulmonary embolism. This method, however, weakens the amplification of the signal by the inflammatory background through a reference channel and a multi-channel signal processing model, resulting in a corrected concentration closer to the true level. This reduces unnecessary imaging examinations and hospitalizations at the same positive threshold. Similarly, in emergency rooms during nighttime when ambient temperature fluctuates significantly, electrochemical background signals are prone to drift, and traditional equipment tends to report marginal values. This method, by correcting for the synchronous drift signal of the reference channel, provides a more stable corrected concentration, reducing inconsistent results from repeated testing of the same patient and facilitating rapid and clear treatment decisions. Therefore, by introducing a reference detection channel with identical structure but different molecular recognition characteristics onto a paper-based microfluidic sensor, and using the pseudo-target channel as a background signal observation window, and then combining it with a multi-channel signal processing model to perform joint deconvolution and background subtraction on the target channel signal, the corrected target biomarker concentration is obtained, and the detection result is output accordingly. This achieves highly stable and interference-resistant quantitative detection of cardiovascular biomarkers under complex clinical sample conditions. Compared with traditional single-channel, single-threshold interpretation paper-based electrochemical detection schemes, this scheme does not require additional complex hardware structures or operating steps. It significantly improves the resistance to matrix interference and cross-sample consistency mainly through structural and algorithm design, making it very suitable for rapid joint evaluation of multiple biomarkers in emergency chest pain scenarios. It helps reduce misjudgments caused by poor blood sample condition, environmental fluctuations, or minor electrode defects, and enhances the support of point-of-care testing for clinical decision-making.

[0046] In some examples, the structure and materials of the reference detection channel are the same as those of the target biomarker detection channel. The working electrode surface of the reference detection channel is immobilized with a pseudo-target molecule that is similar to the epitopes of at least one of high-sensitivity troponin, D-dimer, and natriuretic peptide, but does not specifically bind to the corresponding biomarker in the clinical sample. This allows the background electrochemical signal caused by non-specific protein adsorption and charge distribution in the complex matrix to be correlated between the reference detection channel and the target biomarker detection channel, serving as a background reference signal.

[0047] The method of jointly deconvolving and subtracting the measurement signals from each target marker detection channel to solve for the corrected concentration value of at least one target marker includes:

[0048] The electrochemical parameters of the reference detection channel are introduced as background terms into the multi-channel signal processing model. The measurement signals of each target marker detection channel are jointly deconvolved and background subtracted to solve for the corrected concentration value of at least one of high-sensitivity troponin, D-dimer and natriuretic peptide, thereby suppressing common-mode drift and inter-channel cross-interference caused by complex matrix.

[0049] In some examples, the pseudo-target molecule is a short peptide with site mutation, a conformationally locked nucleic acid aptamer, or a combination thereof, which maintains a spatial conformation and charge environment similar to that of the target marker recognition molecule.

[0050] For example, a control channel that senses only the background and not the actual biomarker can be artificially constructed on the same microfluidic paper-based chip: Since the reference detection channel is as consistent as possible with the target biomarker detection channel in terms of paper-based flow channel geometry, electrode substrate material, antifouling layer type and thickness, and nanostructure modification method, when complex blood samples migrate within the paper substrate, a large number of non-specific proteins, lipoproteins, inorganic ions, and trace cell debris in the plasma will produce highly similar non-specific adsorption and charge redistribution on the electrode surfaces of the target and reference channels, thus exhibiting very similar background changes in electrochemical impedance spectroscopy and voltammetric responses; simultaneously, a dummy target fixed on the working electrode surface of the reference detection channel... The target molecule is deliberately designed in terms of epitope conformation, charge environment, and spatial size to resemble at least one of the epitopes of high-sensitivity troponin, D-dimer, and natriuretic peptides. This ensures that its nonspecific adsorption behavior and double-layer formation behavior are as consistent as possible with the target channel. However, through site mutation, side-chain chemical modification, or sequence truncation, its affinity for the real biomarker in clinical samples is significantly reduced, resulting in almost no specific binding within the actual detection concentration range. This ensures that the reference channel contains almost no biomarker-specific response and mainly reflects the background electrochemical signal caused by the complex matrix. From a statistical correlation perspective, the reference detection channel records the time series of the common-mode background, while the target channel records the common-mode background. The time series of specific signals superimposed on the background can be used to create a pattern. In the electrode printing process, carbon ink, silver-silver-chlorine ink, and gold nanoparticle modification solutions from the same batch can be uniformly proportioned. First, screen printing is completed on a single substrate and uniformly cured. Then, spatially adjacent target and reference channels are formed through wax patterning or laser etching. Subsequently, anti-high-sensitivity troponin antibodies, anti-D-dimer antibodies, and anti-natriuretic peptide antibodies are immobilized on the working electrode of the target channel via EDC / NHS coupling or gold-sulfur self-assembly. On the working electrode of the reference channel, short peptides with site mutations or conformationally locked aptamers are immobilized to maintain similar charge density and hydrophobicity, making non-specific proteins more likely to adsorb in a similar manner. With this design, when the blood lipid sample... When there are abnormalities, mild hemolysis, increased inflammatory factors, or mild electrode contamination, the effects of these factors on the charge transfer resistance, double-layer capacitance, and current-current background in the low-frequency region of the impedance spectrum will be highly synchronized in the target channel and the reference channel. The reference channel can be regarded as a sensor reflecting the environmental and matrix state, thus providing a scientific basis for subsequent mathematical background removal. For example, in a patient sample with acute chest pain, there are both high triglycerides and moderate inflammation. If only the impedance change of the target channel is observed, it is difficult to distinguish whether the background drift is caused by elevated markers or by lipemia and inflammation. However, if the impedance of the reference channel is found to drift almost proportionally, it can be inferred that a part of the signal in the target channel should be classified as background component.

[0051] For example, the step of performing joint deconvolution and background subtraction on the measurement signals of each target marker detection channel to solve for the corrected concentration value of at least one target marker can be understood as constructing a multi-channel, multi-parameter linear or weakly nonlinear hybrid model, writing the electrochemical measurement signal of each target channel as a combination of background terms and specific terms, and then using the background observation given by the reference channel to decompose and solve the hybrid signal. Specifically, the impedance modulus, phase angle, peak current, and peak potential of each target channel at several characteristic frequencies, along with peak current and peak potential under one or two volt-ampere tests, can be concatenated into a high-dimensional vector. The high-dimensional vector corresponding to the reference channel is used as the background vector. During the methodological calibration phase, a series of high-sensitivity troponin, D-dimer, and natriuretic peptide standards at known concentrations are prepared and repeatedly tested under different matrix conditions, such as varying degrees of hemolysis, lipemia, bilirubin interference, and different total protein amounts. The mathematical relationship between the electrochemical characteristics of each target channel and the concentrations of the three biomarkers and the characteristics of the reference channel is fitted. This relationship can be formalized into several equations. For example, for a specific target channel, it can be expressed as a vector equation: the target channel eigenvector equals the background weight matrix multiplied by the reference channel eigenvector, plus the response coefficient matrix multiplied by the concentration vectors of the three biomarkers, plus a random noise term. In simplified cases, this can also degenerate into a multiple linear regression model or a kernel regression model. In actual detection... In the first stage, the feature vector of the reference channel acquired at the same time point from the same chip is substituted into the model. First, the part of the signal in the target channel that should belong to the background is estimated by the background weight matrix. Then, this part of the background estimate is subtracted from the measured signal of the target channel. The remaining residual is the combination of the specific responses of the three markers. Then, the residual is fitted and solved by multivariate fitting according to the pre-calibrated response coefficient matrix or nonlinear calibration curve to obtain the corrected concentration value of at least one of the markers, namely high-sensitivity troponin, D-dimer and natriuretic peptide. This process is essentially deconvolution and background subtraction of the mixed observations convolved with the background and specific signals. In the scenario of simultaneous detection of three channels, the cross-interference term between channels can also be explicitly introduced into the same model. For example, the high concentration of a certain marker has a non-specific adsorption effect on another channel. By using a multi-marker mixed solution in the calibration stage to obtain the cross-interference coefficient, this cross-influence is decoupled during detection, thereby further improving the accuracy of the concentration calculation.This multi-channel signal processing model can significantly suppress common-mode drift and inter-channel cross-interference caused by complex matrices. Therefore, firstly, in the low-concentration range, where the target signal is close to background noise, even slight changes in lipemia or electrode state can push the result up or down. Introducing a reference channel and joint deconvolution effectively improves the signal-to-noise ratio in the low-concentration range, reducing the probability of false positives and false negatives. For example, true high-sensitivity troponin is only a few nanograms per liter, easily masked by hemolysis or temperature drift in traditional electrochemical sensors without anti-interference strategies. This model can first use the reference channel to eliminate common-mode changes caused by hemolysis, retaining only the small differential signal related to troponin binding, thus providing a more accurate corrected concentration value. Secondly, in scenarios involving parallel detection of multiple biomarkers, the high concentration of one biomarker... Concentration often indirectly affects the electrochemical readings of other channels through nonspecific adsorption or matrix changes. Without cross-interference modeling, a high elevation of one biomarker may lead to a misjudgment of a slight elevation of another biomarker. Combined deconvolution explicitly introduces this cross-influence into the model and subtracts it during solution, which can reduce such misjudgments. For example, a patient suspected of having pulmonary embolism may have extremely high D-dimer levels. The natriuretic peptide channel on the same chip may be affected by some degree of matrix changes. Looking at the original signal alone, it is easy to misjudge as an increased risk of heart failure. After combining background reference and deconvolution, the model will attribute a large part of the changes to background and cross-interference. The natriuretic peptide concentration may be restored to the normal range, thus indicating a high risk of pulmonary embolism but not a high risk of heart failure. This helps doctors to accurately select imaging examinations and treatment strategies.

[0052] In some examples, the liquid absorption end of the microfluidic paper-based sensor is provided with a composite liquid absorption pad, which includes a stacked fast liquid absorption layer and a variable resistance layer. The fast liquid absorption layer is made of high-porosity cellulose material to provide a reference liquid absorption driving force, and the variable resistance layer is a porous matrix with a porous network structure.

[0053] In some examples, the variable resistance layer is a composite material in which hydrophilic and hydrophobic microparticles are mixed in a porous matrix to form a wetting gradient. It is used to adjust the permeation resistance of the sample in the composite absorbent pad by the contact angle difference in different regions. When the sample hematocrit increases and the viscosity increases, the variable resistance layer is used to automatically improve the equivalent absorbent capacity by changing the wetting angle, so as to maintain the average flow rate and reaction residence time in the detection channel of each target marker within a preset range under different hematocrit conditions, thereby reducing the impact of inter-individual rheological differences on the corrected concentration value.

[0054] It is understandable that the rheological properties of blood vary greatly among different patients, especially when hematocrit is high, which significantly increases sample viscosity. Simple paper-based capillary flow slows down and residence time increases, leading to different electrochemical responses at the same concentration in different patients, thus compromising the stability of calibration curves. The proposed solution involves stacking two layers at the aspiration end: a high-porosity cellulose layer for rapid aspiration, providing a relatively stable base suction, and a variable resistance layer that can self-adjust resistance. Since the propulsion speed of paper-based capillary flow is closely related to parameters such as pore size, contact angle, and viscosity, when hematocrit increases leading to higher viscosity, the forward propulsion will significantly slow down without intervention, and the residence time of the sample within the channel will double or even more. The variable resistance layer, by incorporating hydrophilic and hydrophobic microparticles into the porous matrix, creates a spatial wetting gradient, resulting in different contact angles for the liquid in different regions, essentially arranging a variable "capillary pump" along the aspiration path. In practice, a porous cellulose or polymer mesh can be used as a carrier, in which two types of particles are uniformly mixed: one type has hydrophilic groups on its surface, resulting in a small contact angle when in contact with blood samples and easy wetting; the other type has a hydrophobic surface, resulting in a larger initial contact angle and less wetting. By controlling the ratio and spatial distribution of the two types of particles, the variable resistance layer exhibits an equivalent wetting characteristic that varies with the sample properties when the sample flows through it. When the sample viscosity is low and the hematocrit is normal, the liquid easily penetrates the variable resistance layer. At this time, the equivalent osmotic resistance is mainly determined by the fast absorbent layer, and the overall flow rate is within the upper limit of the design. However, when the hematocrit increases and the viscosity increases, the interaction between the sample and the hydrophilic particles in the variable resistance layer is enhanced, the wetting angle in the local area decreases, and the capillary suction increases, which is equivalent to automatically "increasing the suction force" to compensate for the slowdown in flow rate caused by the increase in viscosity. In this way, the overall equivalent absorbent capacity of the composite absorbent pad is improved, thereby maintaining the average flow rate in the channel at a level close to the preset range. The variable resistance layer has a porous network structure, meaning that liquids can preferentially advance along relatively easier wetting paths. The position and shape of the wetting front automatically evolve according to the sample properties. Macroscopically, this means that under different hematocrit conditions, the overall sample advancement speed and residence time in the detection electrode area are limited to a small range. This design is equivalent to integrating a passive adaptive flow regulation unit without any feedback control or sensors within a disposable paper-based chip. This significantly reduces the impact of individual patient differences or pre-testing variations such as slight dehydration or fluid rehydration on the electrochemical response curve, allowing the calibration relationship from signal to concentration to remain stable across a wider population.For example, between a young male patient with high hematocrit and an elderly patient with low hematocrit, the sample advance speed of the former on a traditional paper-based chip may be half that of the latter, resulting in a much higher signal intensity of troponin at the same concentration. However, by using a composite absorbent pad, the variable resistance layer provides greater equivalent suction to the former, significantly reducing the difference in residence time of the two samples in the detection area. As a result, the electrochemical response curves of the two patients at the same true concentration are closer, which is beneficial for using a unified calibration model.

[0055] In some examples, the microfluidic paper-based sensor further includes a common working electrode. The surface of the working electrode is divided into multiple spatially separated recognition regions. Each recognition region is provided with specific recognition molecules for recognizing different target markers, so that multiple target marker detection channels share the same working electrode in physical structure, forming concentric annular regions or adjacent sector regions on a plane. The method further includes:

[0056] When acquiring the electrochemical impedance spectroscopy and / or voltammetric signal of each target biomarker detection channel, impedance test excitation is applied to the common working electrode at different frequency bands and / or voltammetric test waveforms with different pulse parameters are used to make the electron transfer kinetic response corresponding to each identification region have distinguishable characteristics in the frequency domain and / or waveform parameter domain, so as to separate and fit the electrochemical signals of different target biomarkers based on frequency band and / or waveform parameter multiplexing in the multi-channel signal processing model.

[0057] For example, a microfluidic paper-based sensor uses a single common working electrode with multiple spatially separated recognition regions on its surface. Each recognition region is equipped with a specific recognition molecule for recognizing different target markers. This compresses the multi-index detection that would otherwise require multiple independent electrodes onto a single working electrode, virtualizing the channels through spatial micro-patterning and surface chemical differences. The working electrode as a whole remains a conductive substrate, such as a carbon-based printed electrode or a gold-based thin film, with only one electrical connection. However, at the microscale, multiple relatively independent recognition regions appear on its surface through mask printing, photolithography, inkjet printing, or wax patterning. These regions are separated by hydrophobic insulating bands, unmodified surface areas, or highly dense inert protective layers, thereby limiting the diffusion and interference between recognition molecules and reactants in different regions. In the concentric ring scheme, the central region can serve as the first marker recognition region, and the outer rings as the second and third recognition regions. In the adjacent sector scheme, several sector-shaped recognition regions can be divided in a pie chart-like manner, with lateral mass transfer blocked at the boundaries of each region by locally hydrophobic or highly cross-linked polymer layers. Because these recognition regions share the same conductive substrate, the entire working electrode is still considered a single node in the circuit. The reference electrode and counter electrode can also maintain a single structure. This simplifies wiring, reduces the number of silver wires, and ensures that each recognition region is macroscopically under the same potential excitation and solution environment, facilitating the differentiation of responses in different regions through frequency domain or waveform design. In practice, a working electrode of sufficient area can be printed on a paper or plastic substrate using carbon or gold ink. Subsequently, several concentric or sector-shaped hydrophilic regions are formed on it using wax printing or photolithography as recognition areas, while other areas are covered with hydrophobic materials. Then, anti-high-sensitivity troponin antibodies, anti-D-dimer antibodies, anti-natriuretic peptide antibodies, or their aptamers are fixed on the surfaces of different recognition regions, giving each region the specific recognition capability for different targets. This structure achieves a planar distribution of multi-target detection channels without increasing the number of working electrodes. Furthermore, because these regions share the same substrate, their basic electrode properties, such as thickness, roughness, and conductivity, are essentially consistent, which helps reduce systematic errors caused by differences in electrode processes.

[0058] For example, when acquiring the electrochemical impedance spectroscopy or volt-ampere signal of each target marker detection channel, impedance test excitation can be applied to the common working electrode at different frequency bands, or volt-ampere test waveforms with different pulse parameters can be used to make the electron transfer kinetic response corresponding to each identification region have distinguishable characteristics in the frequency domain or waveform parameter domain. Although multiple identification regions electrically share a working electrode, the interfacial reaction kinetic constant, double-layer capacitance, diffusion layer thickness, and charge transfer path in each region are slightly different. If the excitation frequency, amplitude, and waveform are designed appropriately, the contribution of a certain region can be "amplified" and the contribution of other regions can be suppressed in the overall measured impedance or volt-ampere signal. This allows for the acquisition of multiple sets of observations with different weights for different regions in a set of test sequences, and then the information can be mathematically restored to the independent response of each region. Specifically, in the impedance spectrum, the overall interface can be equivalent to a parallel structure of multiple Randle branches, each branch corresponding to an identification region with its own charge transfer resistance, double-layer capacitance, and diffusion impedance parameters. Because different recognition molecules, marker molecular weights, and surface structures lead to different reaction rates and interface structures, some regions are more sensitive to impedance changes in the low-frequency range, while others are more sensitive to double-layer capacitance changes in the mid-to-high-frequency range. By selecting several characteristic frequency points for single-point measurement or short-sweep measurement, different weighted signal combinations for different regions can be obtained. The response sensitivity matrix of each recognition region at multiple frequency points can be determined in advance under standard samples. A set of frequencies can be selected, such that one frequency dominates the contribution of the high-sensitivity troponin recognition region, another frequency is more sensitive to the D-dimer recognition region, and a third frequency is more sensitive to the natriuretic peptide recognition region, thus forming a spectral code. In voltammetry, several differential pulse voltammeters or square wave voltammeters with different pulse parameters can be designed, such as changing the pulse height, pulse width, step potential spacing, and scan rate. Different parameter combinations will produce different amplification effects on interface processes with faster or slower reaction kinetics, causing some recognition regions to produce significant current peaks under certain waveforms, while responding weakly under other waveforms. By sequentially applying several sets of tests with different frequencies or different volt-ampere parameters on the same chip, a set of multiplexed observations can be obtained. Each set of observations is composed of the superposition of responses from multiple identification regions according to their respective weights. The advantage of this frequency band and waveform parameter multiplexing is that it obtains separable multi-channel information with limited test time and a minimal electrode structure, without the need to configure a separate working electrode and independent circuit path for each marker.For example, three frequency points can be selected: low frequency to amplify the slow charge transfer process caused by high molecular weight markers, medium frequency to reflect the medium kinetic region, and high frequency to reflect the change in double-layer capacitance. Combined with two differential pulse voltammetry tests with different parameters, five to six sets of observation data are finally formed. Among these observations, the high-sensitivity troponin recognition region contributes the most at low frequency and with the first voltammetry parameter, the D-dimer recognition region contributes significantly at medium frequency and with the second voltammetry parameter, and the natriuretic peptide recognition region is more sensitive to impedance in the high-frequency band and peak position shift in the two voltammetry tests, thus forming a mathematically solvable set of multiple equations.

[0059] For example, in a multi-channel signal processing model, electrochemical signals of different target biomarkers are separated and fitted based on frequency band and / or waveform parameter multiplexing. The principle is to treat each observed overall electrochemical signal as a linear or weakly nonlinear superposition of responses from multiple identification regions. Then, combining this with the aforementioned frequency or waveform weight matrix for each region, the independent responses of each region are reconstructed by solving this superposition relationship. These responses are then mapped to the concentrations of each biomarker through calibration curves. In practice, single-biomarker standards can be used during the calibration stage. Specifically, samples containing only high-sensitivity troponin, only D-dimer, or only natriuretic peptide can be prepared, and the overall signal at each frequency and each set of voltammetric parameters is tested gradient-wise. Since only one identification region is effective at this point, and other regions only produce baseline responses, these data can be directly used to calibrate the coefficients of the relationship between the overall signal and the response of a specific region at a certain frequency or waveform parameter. Furthermore, mixed samples of two or three biomarkers can be used to test the additivity and cross-interference of the model, and cross terms can be introduced when necessary to construct a more accurate multiple regression or matrix factorization model. During detection, the reading device applies multiple sets of frequency and waveform parameters in a preset sequence within a short period of time. The common working electrode provides an overall impedance or current-voltage characteristic vector under each set of excitations. These vectors are then sequentially input into a multi-channel signal processing model. The model uses known weight matrices and cross terms to infer the independent response of each recognition region under each excitation condition, thereby obtaining the electrochemical characteristics corresponding to the high-sensitivity troponin recognition region, D-dimer recognition region, and natriuretic peptide recognition region. These characteristics are then converted into their respective concentration values ​​using a single-channel calibration curve. If necessary, the background parameters of the reference detection channel can also be included in the model to perform background subtraction and correction on the overall signal, resulting in the corrected concentration results for the three markers. This approach significantly reduces the number of electrodes and wiring complexity, converging a structure that might otherwise require three or more working electrodes and multiple connections into a single shared working electrode and a unified test sequence, thus lowering manufacturing costs and chip size. Furthermore, by reusing frequency bands and waveform parameters, the same hardware structure can generate a sufficient number of independent observations within a limited time, allowing for mathematical separation of the responses of multiple biomarkers. This maintains the capability for joint detection of multiple indicators while reducing systematic errors caused by batch and positional differences between electrodes. For example, in a paper-based microfluidic chip used for detecting three cardiovascular markers in chest pain, only a relatively large carbon-based working electrode needs to be printed. Three sectors are then divided to immobilize recognition molecules targeting high-sensitivity troponin, D-dimer, and natriuretic peptides, respectively. A reference region is designed at the edge for a pseudo-target channel. During actual detection, the reading device first performs impedance measurements at three frequency points, then applies two sets of differential pulse voltammetric waveforms. The entire process takes, for example, tens of seconds, yielding five to six sets of overall electrochemical characteristics.The multi-channel signal processing model decomposes the overall signal into independent responses from three identification sectors and a reference region based on a pre-calibrated weight matrix. The response from the reference region is used to estimate and subtract background, and the responses from the three identification sectors are then mapped to the concentrations of three biomarkers, respectively. Clinically, doctors still receive the familiar three numbers and corresponding risk warnings. However, within the device and algorithm, through shared electrodes, frequency band reuse, and mathematical decoupling, hardware resources have been significantly saved, and the consistency and robustness of the results have been improved.

[0060] In some examples, it also includes:

[0061] A calibration sub-model was established for each identification region using a multi-channel signal processing model. Electrochemical parameters obtained under different frequency bands and / or different voltammetric test parameters were matched and fitted with the corresponding calibration sub-model to obtain the initial concentration estimate of each target marker.

[0062] By combining the electrochemical parameters of the reference detection channel with joint deconvolution and background subtraction, the corrected concentration values ​​of each target marker are determined.

[0063] For example, in the methodological calibration stage, standard sample gradients containing only high-sensitivity troponin, only D-dimer, and only natriuretic peptide are prepared separately, and repeated tests are conducted under different total protein concentrations, ionic strengths, and temperatures. This allows the microfluidic paper-based sensor to operate under preset impedance test frequency bands and different pulse parameters under volt-ampere test conditions, thereby obtaining multidimensional electrochemical feature vectors including impedance modulus, phase angle, peak current, and peak shift in each recognition region. For each type of target marker corresponding to a recognition region, based on the test data of the chip where the recognition region is located under single-marker conditions, a calibration sub-model is constructed to map the electrochemical features under different frequency bands and volt-ampere parameters to the concentration of the marker. The calibration sub-model can be a multiple linear regression model, a kernel regression model, or a simplified small neural network model, used to characterize the interfacial charge transfer resistance of the recognition region under different combinations of frequencies and waveform parameters. The study investigated the variation of double-layer capacitance and diffusion process with biomarker concentration. During the detection phase, the reading device applied multiple sets of impedance excitations and voltammetric test waveforms to the common working electrode according to the test sequence consistent with the calibration phase. The overall electrochemical characteristics acquired were decomposed by the aforementioned multi-channel signal processing framework to obtain the characteristic parameters of each identification region. These characteristic parameters were then input into the corresponding calibration sub-models for matching and fitting to solve for the initial concentration estimates of each target biomarker, such as high-sensitivity troponin, D-dimer, and natriuretic peptide. Since the calibration sub-model was trained under known real concentration and standard matrix conditions and fully utilized the differences in interface dynamics under multiple frequency bands and waveforms, it was able to perform targeted fitting of the response of each identification region under the constraints of a common working electrode and frequency band reuse, achieving preliminary quantification of each target biomarker. This resulted in high sensitivity and resolution even on a microfluidic paper-based platform with limited hardware resources.

[0064] For example, during the calibration phase, the electrochemical characteristics of the reference detection channel are simultaneously acquired under the same impedance test frequency band and voltammetric test parameters. The reference detection channel is regarded as an observation channel for the common-mode background signal caused by the complex matrix. Through large-sample calibration under sample conditions with different degrees of hemolysis, different degrees of lipemia, different bilirubin levels, and different inflammatory states, a mapping relationship between the feature vector of the reference detection channel and the background offset of each target channel is established. A joint model containing the true concentration, background offset, and noise term is constructed for each target biomarker, so that the initial concentration estimate is represented as the superposition of the true concentration and the background function of the reference channel. In the actual detection phase, the electrochemical characteristics of the reference detection channel acquired from the same chip at the same time point are substituted into the mapping relationship to estimate the background offset caused by the current matrix state to the respective channels of high-sensitivity troponin, D-dimer, and natriuretic peptide. Then, this background offset is transferred from the corresponding biomarker. The initial concentration estimate is subtracted to obtain the corrected concentration value after background correction. In embodiments where further accuracy improvement is required, the reference detection channel can be considered as a background channel and introduced into the multi-channel matrix solution process. The responses of the target channel and the reference channel are included in the solution equation. The true concentration vector of each biomarker and the background weight vector are solved simultaneously through least squares or regularized optimization algorithms, realizing joint deconvolution and background subtraction of the shared electrode multiplexed signal. Through the above steps, common mode drift caused by factors such as hemolysis, lipemia, total protein fluctuation, inflammatory state, and changes in ambient temperature in complex blood samples is effectively removed from the initial concentration estimate, making the corrected target biomarker concentration closer to the actual physiological level. This significantly reduces the risk of false positives and false negatives in the low concentration range, reduces cross-interference in the case of high concentration of multiple biomarkers, and improves the quantitative accuracy and result stability of microfluidic paper-based sensors in complex clinical scenarios such as acute chest pain.

[0065] The anti-interference detection method based on microfluidic paper-based sensor in the embodiments of this application has been described above. The anti-interference detection system based on microfluidic paper-based sensor in the embodiments of this application is described below.

[0066] Please see Figure 2 This application describes an embodiment of an anti-interference detection system based on a microfluidic paper-based sensor, which may include:

[0067] Acquisition unit 201 is used to acquire the electrochemical signal of a reference detection channel while acquiring electrochemical impedance spectroscopy and / or voltammetric signal using each target biomarker detection channel. The reference detection channel is disposed in an adjacent region of at least one target biomarker detection channel of the sensor. The working electrode surface of the reference detection channel is provided with pseudo-target molecules that do not specifically bind to the target biomarkers in the clinical sample within the detection concentration range.

[0068] The correction unit 202 is used to input the electrochemical parameters of the reference detection channel as background terms into the multi-channel signal processing model, and to perform joint deconvolution and background subtraction on the measurement signals of each target marker detection channel to solve for the corrected concentration value of at least one target marker.

[0069] The determining unit 203 is used to determine the detection result based on the corrected concentration value.

[0070] above Figure 2 The anti-interference detection system based on a microfluidic paper-based sensor in this application embodiment has been described from the perspective of modular functional entities. The following is a detailed description of the anti-interference detection system based on a microfluidic paper-based sensor in this application embodiment from the perspective of hardware processing. Please refer to... Figure 3 One embodiment of the anti-interference detection system 300 based on a microfluidic paper-based sensor in this application includes:

[0071] The system includes an input device 301, an output device 302, a processor 303, and a memory 304, wherein the number of processors 303 can be one or more. Figure 3 Taking a processor 303 as an example. In some embodiments of this application, the input device 301, output device 302, processor 303, and memory 304 can be connected via a bus or other means, wherein... Figure 3 Taking the example of a connection between China and Israel via a bus.

[0072] Specifically, the processor 303 executes the above steps by calling the operation instructions stored in the memory 304.

[0073] By calling the operation instructions stored in memory 304, processor 303 is also used to execute... Figure 1 Any of the methods in the corresponding embodiments.

[0074] Please see Figure 4 , Figure 4 A schematic diagram illustrating an embodiment of the electronic device provided in this application.

[0075] like Figure 4 As shown, this application provides an electronic device, including a memory 304, a processor 303, and a computer program 411 stored in the memory 304 and executable on the processor 303. When the processor 303 executes the computer program 411, it performs the above steps.

[0076] In practical implementation, when processor 303 executes computer program 411, it can achieve... Figure 1 Any of the corresponding implementation methods in the embodiments.

[0077] Since the electronic device described in this embodiment is the device used to implement the anti-interference detection system based on a microfluidic paper-based sensor in this application embodiment, those skilled in the art can understand the specific implementation method and various variations of the electronic device in this embodiment based on the method described in this application embodiment. Therefore, how the electronic device implements the method in this application embodiment will not be described in detail here. Any device used by those skilled in the art to implement the method in this application embodiment is within the scope of protection of this application.

[0078] Please see Figure 5 , Figure 5 This is a schematic diagram illustrating an embodiment of a computer-readable storage medium provided in this application.

[0079] like Figure 5 As shown, this embodiment provides a computer-readable storage medium 500 on which a computer program 511 is stored, which performs the above steps when executed by a processor.

[0080] By calling the operation instructions stored in memory 304, processor 303 is also used to execute... Figure 1 Any of the methods in the corresponding embodiments.

[0081] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An anti-interference detection method based on a microfluidic paper-based sensor, characterized in that, include: While acquiring electrochemical impedance spectroscopy and / or voltammetric signals using each target biomarker detection channel, an electrochemical signal of a reference detection channel is also acquired. The reference detection channel is located in an adjacent region to at least one target biomarker detection channel of the sensor. The working electrode surface of the reference detection channel is provided with pseudo-target molecules that do not specifically bind to the target biomarkers in the clinical sample within the detection concentration range. The electrochemical parameters of the reference detection channel are used as background terms and input into the multi-channel signal processing model. The measurement signals of each target biomarker detection channel are jointly deconvolved and background subtracted to solve for the corrected concentration value of at least one target biomarker. The detection result is determined based on the corrected concentration value.

2. The method according to claim 1, characterized in that, The structure and materials of the reference detection channel are the same as those of the target biomarker detection channel. The working electrode surface of the reference detection channel is immobilized with a pseudo-target molecule that is similar to the epitopes of at least one of high-sensitivity troponin, D-dimer and natriuretic peptide but does not specifically bind to the corresponding biomarker in the clinical sample. This is to make the background electrochemical signal caused by the adsorption of non-specific proteins and charge distribution in the complex matrix correlated between the reference detection channel and the target biomarker detection channel, and to serve as a background reference signal. The method of jointly deconvolving and subtracting the measurement signals of each target marker detection channel to solve for the corrected concentration value of at least one target marker includes: The electrochemical parameters of the reference detection channel are introduced as background terms into the multi-channel signal processing model. The measurement signals of each target marker detection channel are jointly deconvolved and background subtracted to solve for the corrected concentration value of at least one of high-sensitivity troponin, D-dimer and natriuretic peptide, thereby suppressing common-mode drift and inter-channel cross-interference caused by complex matrix.

3. The method according to claim 1, characterized in that, The pseudo-target molecule is a short peptide with site mutation, a conformationally locked nucleic acid aptamer, or a combination thereof, and the pseudo-target molecule maintains a spatial conformation and charge environment similar to that of the target marker recognition molecule.

4. The method according to claim 1, characterized in that, The liquid absorption end of the microfluidic paper-based sensor is provided with a composite liquid absorption pad, which includes a stacked fast liquid absorption layer and a variable resistance layer. The fast liquid absorption layer is made of high-porosity cellulose material to provide a reference liquid absorption driving force, and the variable resistance layer is a porous matrix with a porous network structure.

5. The method according to claim 4, characterized in that, The variable resistance layer is a composite material in which hydrophilic and hydrophobic microparticles are mixed in a porous matrix to form a wetting gradient. It is used to adjust the permeation resistance of the sample in the composite absorbent pad by the contact angle difference in different regions. When the sample hematocrit increases and the viscosity increases, the variable resistance layer is used to automatically improve the equivalent absorbent capacity by changing the wetting angle. This is to maintain the average flow rate and reaction residence time in the detection channels of each target marker within a preset range under different hematocrit conditions, thereby reducing the impact of inter-individual rheological differences on the corrected concentration value.

6. The method according to any one of claims 1 to 5, characterized in that, The microfluidic paper-based sensor also includes a common working electrode. The surface of the working electrode is divided into multiple spatially separated recognition regions. Each recognition region is provided with specific recognition molecules for recognizing different target markers, so that multiple target marker detection channels share the same working electrode in physical structure and form concentric annular regions or adjacent sector regions on a plane. The method further includes: When acquiring the electrochemical impedance spectroscopy and / or voltammetric signal of each target biomarker detection channel, impedance test excitation is applied to the common working electrode at different frequency bands and / or voltammetric test waveforms with different pulse parameters are used to make the electron transfer kinetic response corresponding to each identification region have distinguishable characteristics in the frequency domain and / or waveform parameter domain, so as to separate and fit the electrochemical signals of different target biomarkers based on frequency band and / or waveform parameter multiplexing in the multi-channel signal processing model.

7. The method according to claim 6, characterized in that, Also includes: A calibration sub-model was established for each identification region using a multi-channel signal processing model. Electrochemical parameters obtained under different frequency bands and / or different voltammetric test parameters were matched and fitted with the corresponding calibration sub-model to obtain the initial concentration estimate of each target marker. By combining the electrochemical parameters of the reference detection channel with joint deconvolution and background subtraction, the corrected concentration values ​​of each target marker are determined.

8. An anti-interference detection system based on a microfluidic paper-based sensor, characterized in that, include: The acquisition unit is used to acquire the electrochemical signal of a reference detection channel while acquiring electrochemical impedance spectroscopy and / or voltammetric signals using each target biomarker detection channel. The reference detection channel is disposed in an adjacent region of at least one target biomarker detection channel of the sensor. The working electrode surface of the reference detection channel is provided with pseudo-target molecules that do not specifically bind to the target biomarkers in the clinical sample within the detection concentration range. The correction unit is used to input the electrochemical parameters of the reference detection channel as background terms into the multi-channel signal processing model, and to perform joint deconvolution and background subtraction on the measurement signals of each target marker detection channel to solve for the corrected concentration value of at least one target marker. A determining unit is used to determine the detection result based on the corrected concentration value.

9. An electronic device, characterized in that, The electronic device includes at least one processor and at least one memory connected to the processor, wherein the processor is used to call program instructions in the memory to execute the anti-interference detection method based on a microfluidic paper-based sensor as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the storage medium to perform the anti-interference detection method based on a microfluidic paper-based sensor as described in any one of claims 1 to 7.

Images