PD-L1 protein high-sensitivity detection method and device based on optical fiber SPR (Surface Plasmon Resonance)

Abstract

The invention discloses a PD-L1 protein high-sensitivity detection method and device based on optical fiber SPR (Surface Plasmon Resonance), and relates to the technical field of biomolecule detection. The method comprises the following steps: preparing a disposable optical fiber sensing probe plated with a gold film; a DNA regular tetrahedron containing a PD-L1 aptamer is fixed through an Au-S bond, and a PD-L1 monoclonal antibody is used for functionalizing gold nanoparticles to amplify a signal; connecting an optical path through a self-adaptive coupler and optimizing a coupling state in real time; collecting dual-channel signals to construct a data set; optimizing a buffer system, detecting labeled serum and establishing a calibration curve; finally, the kit is applied to clinical serum sample detection. According to the present invention, the detection sensitivity and the specificity are significantly improved through the optical fiber SPR technology, and the method has advantages of real time, trace sample consumption, low detection limit, good repeatability and the like, can be widely used for the rapid detection of the clinical tumor immunotherapy related protein marker, and has important practical application value.

Description

Technical Field

[0001] This invention relates to the field of biosensor technology, specifically to a high-sensitivity detection method and device for PD-L1 protein based on fiber optic SPR. Background Technology

[0002] Lung cancer is one of the most common and deadliest malignant tumors in the world, and non-small cell lung cancer (NSCLC) is a common type that is often diagnosed at an advanced stage, leading to poor prognosis. In recent years, with advancements in anti-tumor treatments and the emergence of new drugs, immunotherapy has become the first-line standard of care for patients with advanced NSCLC. PD-L1 is currently the only recommended prognostic biomarker, and its expression level is a key indicator for assessing the efficacy of immunotherapy. However, the definition of the clinically relevant cutoff value for PD-L1 positivity remains controversial, and these differences not only hinder the establishment of universally accepted guidelines but also affect treatment decisions.

[0003] PD-L1 is an important immune checkpoint molecule in the immune system, playing a crucial role in tumor immune escape. Therefore, PD-L1 detection is of great significance for biomarker screening and personalized treatment in immunotherapy. Traditional detection methods include immunohistochemistry, enzyme-linked immunosorbent assay (ELISA), flow cytometry, and chemiluminescent immunoassay. While reliable, these methods typically suffer from large sample requirements, complex procedures, and insufficient sensitivity, making it difficult to detect low-abundance samples in real time and to accurately pinpoint PD-L1 status. Immunohistochemistry is the most commonly used clinical method for PD-L1 protein detection, achieving protein localization and semi-quantitative analysis through tissue staining. However, its results are significantly affected by antibody specificity, tissue fixation, and subjective interpretation, and it cannot achieve quantitative detection or dynamic monitoring. Although ELISA can achieve quantitative analysis, it relies on enzyme labeling reactions and multiple washing steps, making it cumbersome, time-consuming, and lacking in sensitivity, making it difficult to detect low-concentration samples. Flow cytometry is suitable for detecting PD-L1 on cell surfaces, but it requires cell separation and labeling, making it unsuitable for direct detection of soluble PD-L1 in body fluid samples. Furthermore, the equipment is expensive and the operation is complex. Chemiluminescent immunoassay, while offering high sensitivity, still requires labeling steps and complex signal amplification systems, resulting in high costs and difficulty in achieving real-time monitoring. In recent years, surface plasmon resonance (SPR) has been increasingly applied to clinical biomarker detection due to its advantages such as high sensitivity, label-free operation, and real-time dynamic monitoring. Currently, this sensor has enabled rapid and highly sensitive detection of serum, plasma, and whole blood-related disease biomarkers in immunodiagnostics. However, while traditional prism-based SPR biosensors are sensitive, reliable, and accurate, they are typically bulky and expensive, making them unsuitable for the rapidly growing demand for point-of-care diagnostics. Therefore, developing a PD-L1 detection technology that combines high sensitivity, real-time performance, portability, and low cost is extremely important.

[0004] In contrast, fiber optic SPR biosensors compete with traditional devices in terms of sensitivity and specificity, while offering the potential for developing compact, low-cost, and fully automated point-of-care testing tools. Fiber optic SPR sensors utilize optical fibers as the light transmission medium, fabricating a thin metal film in the fiber core or surface. By exciting surface plasmon waves that interact with the analyte, high-sensitivity detection of refractive index changes is achieved (Chinese Invention Patent: CN113433095A). In recent years, various fiber optic SPR structures have been proposed, such as bare-core fiber SPR (Chinese Invention Patent: CN105911025A), wedge-shaped fiber SPR (Chinese Invention Patent: CN108535220A), fiber Bragg grating SPR (Chinese Invention Patent: CN112461795A), and multimode fiber SPR (Chinese Invention Patent: CN111413298A), aiming to improve detection performance. However, existing fiber optic SPR technologies still have significant shortcomings in practical applications. First, traditional fiber optic SPR systems typically use direct fusion-splitter couplers, which in practical applications can easily lead to misalignment between the Y-shaped fiber and the fiber end face. This results in the light intensity failing to reach the specified value or showing a significant difference, affecting repeatability and consistency. Furthermore, the fixed coupling method between the fiber probe and the trunk can easily cause fluctuations in optical path loss. When disposable sensing fibers have scratches or surface contamination, the system cannot distinguish whether the signal attenuation originates from sample bonding or probe damage. Moreover, existing commercial couplers mainly rely on spectral calibration and cannot determine the fiber status in real time during detection, easily generating false positive signals. PD-L1 detection devices based on fiber optic SPR currently lack a real-time coupling quality monitoring module, making it impossible to detect the failure of disposable fiber probes and unable to adapt to the testing needs of clinical samples with large batch-to-batch variations. Therefore, existing fiber optic SPR technology still lacks a practical solution that can dynamically optimize the coupling state and achieve high-sensitivity detection in practical applications. Summary of the Invention

[0005] To address the shortcomings of existing technologies, traditional methods that rely on screws and knobs to control the wavelength data reflected by the light source have slow response speeds and cannot meet the requirements of high-throughput detection. Furthermore, manual knob operation has poor consistency. This invention provides a high-sensitivity detection method and device for PD-L1 protein based on fiber optic SPR, including an innovative optical path coupling structure and a standard method for real-time judgment of signal quality.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A high-sensitivity detection method for PD-L1 protein based on optical fiber SPR is characterized by optically connecting a TDN functionalized optical fiber and a Y-type optical fiber through an adaptive coupler, connecting a light source and stabilizing the light source and optical path, injecting an unknown sample into a liquid tank to specifically bind to the PD-L1 aptamer on the surface of the TDN functionalized optical fiber, amplifying the signal through PD-L1 antibody gold nanoparticles, and then acquiring the SPR signal through an optical signal detection system. The concentration of PD-L1 protein in the unknown sample is obtained by combining the SPR signal with a regression standard curve.

[0007] Preferably, the TDN functionalized optical fiber refers to a DNA tetrahedral framework nucleic acid containing the PD-L1 aptamer sequence that is fixed on the gold film on the surface of the optical fiber sensing through covalent bonding. The fabrication method of the TDN functionalized optical fiber is as follows: S1.1: Cut the original optical fiber to a length of 15-17cm, peel off the cladding 6-7cm from both ends of the optical fiber and retain a 3-4cm protective area in the middle, remove the cladding material in the peeled area with acetone, cut the ends of the optical fiber to prepare a disposable optical fiber with a sensing length of 6-12mm, and then deposit a gold film with a thickness of 35-55nm on the sensing area using magnetron sputtering. S1.2: Take 1 μL of each of the four DNA single strands with an original concentration of 100 μM, add 36 μL of TM buffer, 10 μL of TCEP solution with a concentration of 60 ~ 100 mM, and 50 μL of enzyme-free and pyrogen-free water to form a tube, react at 0 ~ 40℃ for more than 1 h, anneal the mixed solution at 95 ℃ for 10 min and then slowly cool it to 4 ℃ to form TDN solution; S1.3: Dilute the TDN solution to 50 nM using TM buffer, then immerse the fiber sensing area in the TDN solution and incubate overnight for at least 12 hours.

[0008] Preferably, the adaptive coupler adjusts the relative position of the TDN functionalized fiber and the Y-type fiber through an internal piezoelectric ceramic driver. At the same time, the graphene transparent electrode layer monitors the light intensity change signal during the coupling process of the two fibers in real time and outputs it to an external optical signal detection system. Based on the feedback from the optical signal detection system, the piezoelectric ceramic driver calibrates the relative position between the end faces of the two fibers to find the optimal coupling point. When the feedback reaches the optimal coupling point, the current relative position of the two fibers is locked.

[0009] Preferably, the stability of the light source and optical path specifically refers to: S2.1: Connect the TDN functionalized fiber to the Y-type fiber using the knob and adaptive coupler. A certain wavelength is generated by reflection through the light source. Adjust the light intensity to 60000±500 RIU by rotating the light source knob and stabilize for 10 minutes. Determine whether the light source is stable based on the wavelength fluctuation. S2.2: The calibration process of the adaptive coupler adopts closed-loop feedback control. By monitoring the changes in output light intensity, the driving voltage of the piezoelectric ceramic is automatically adjusted until the optimal coupling point is found, that is, the light intensity reaches the baseline position and the fluctuation is ≤±1% for 10s; 60 sets of light intensity data (I1,I2,…,I…) are continuously collected. 60 ) Calculate the coefficient of variation ( CV ),Require CV s ≤1.5% )×100%; in, The average value of 60 consecutive light intensity measurements was collected. The responsivity of the monitored output light intensity is 0.5 A / W, the driving voltage of the piezoelectric ceramic is 50±10 V, and the accuracy is 1 mV.

[0010] Preferably, the acquisition of SPR signals through the optical signal detection system specifically includes: S3.1: Calculate the properties of the substance to be tested using the formulas for the changes in reflectance and refractive index:

[0011] Where R0 is the initial reflectivity, A is the material's response constant, and θ0 is the incident angle corresponding to the SPR peak; the data are processed using adaptive Kalman filtering to optimize the signal-to-noise ratio, combined with local weighted regression to smooth the curve, ensuring that the refractive index calculation error is ≤5×10⁻ 5 RIU; S3.2: Acquire dual-channel optical response signals, extract the average response per second to construct a sample dataset containing peak-summiting data and median data; extract spectral data and calculate the average SPR resonance wavelength shift:

[0012] in, This represents the average SPR resonance wavelength shift. To collect 60 SPR resonance wavelengths within 1 second; It is the set of SPR resonance wavelengths from the 51st to the 60th within 1 second.

[0013] Preferably, the unknown sample is prepared by diluting a serum sample with MES at pH 6.0; the unknown sample is incubated with TDN functionalized optical fiber for 20-40 min, washed, and then incubated with PD-L1 antibody gold nanoparticles for 30 min. The PD-L1 antibody nanoparticles were prepared by using commercial gold nanoparticles to modify anti-PD-L1 monoclonal antibody via electrostatic adsorption. The concentration of the PD-L1 monoclonal antibody was 5 μg / mL. The binding conditions were 25°C with shaking for 1 hour. Finally, the unbound AuNP sites were blocked with BSA. The UV absorbance of the resulting product was 0.5.

[0014] Preferably, the regression standard curve is obtained by the following method: PD-L1 protein standard solutions with concentration gradients were prepared. Each standard solution was incubated with TDN functionalized optical fiber for 20-40 min. After washing, PD-L1 antibody gold nanoparticles were introduced and incubated for 30 min. The changes in SPR resonance wavelength shift were recorded. The standard solution concentration was used as the abscissa and the resonance wavelength was used as the ordinate to plot the working curve. Polynomial curve fitting was performed to obtain the regression standard curve.

[0015] Preferably, during the detection process using TDN-functionalized fiber optic probes, the following steps are required in sequence: TM buffer washing, PD-L1 dilution buffer washing and baseline stabilization, PD-L1 capture, non-specific PD-L1 adsorption washing, PBS / BSA buffer baseline stabilization, PD-L1 antibody gold nanoparticle signal amplification, and washing of unbound PD-L1 antibody gold nanoparticles.

[0016] A high-sensitivity detection device for PD-L1 protein based on fiber optic SPR for use in the method described in this invention is characterized by comprising an X, Y, Z axis displacement stage 1, a power supply 2, a Y-type optical fiber 3, a light source 4, a displacement stage controller 5, a 96-well sample well 6, an adaptive coupler 7, an optical fiber holder 8, a spectrometer 9, a detection system 10, an analysis system 11, and a TDN functionalized optical fiber 12.

[0017] Preferably, the power supply 2 is used to start the detection system 10 and the analysis system 11; in the detection system 10, an adaptive coupler 7 is fixed on the X, Y, and Z axis displacement stage 1, and multiple directions of movement are achieved through the displacement stage controller 5; the Y-type optical fiber 3 is fixed by the optical fiber fixer 8, the light source 4 consists of two LEDs, the light source 4 and the spectrometer 9 are respectively connected to the bifurcation end of the Y-type optical fiber 3, and the other end is optically connected to the TDN functionalized optical fiber 12 through the adaptive coupler 7; the 96-well sample cell 6 is placed below the adaptive coupler 7, and the TDN functionalized optical fiber 12 can be immersed in the reagent in the sample cell by the X, Y, and Z axis displacement stage 1; the spectrometer 9 is used to analyze the reflected light and calculate the SPR resonance wavelength shift in real time; the analysis system 11 processes and analyzes the detection data to realize the detection of PD-L1 protein content; The adaptive coupler 7 is based on a quartz glass frame and integrates a piezoelectric ceramic actuator, a microlens array, a graphene transparent electrode layer, and a refractive index matching liquid elastic cavity. The piezoelectric ceramic actuator is connected to the fiber optic mounting base via a flexible hinge, which can drive the two fibers to finely adjust the spacing along the optical axis. The microlens array is attached to the outer side of the fiber end face for focusing the light spot. The graphene electrode layer covers the inner wall of the elastic cavity and monitors the light intensity changes in real time. The thickness of the refractive index matching liquid elastic cavity is changed by air pressure adjustment, so that the matching liquid is tightly attached to the fiber end face to eliminate reflection loss.

[0018] Compared with the prior art, the advantages of the present invention are as follows: (1) This invention improves the shortcomings of traditional fiber optic SPR sensors, which suffer from low coupling efficiency and insufficient signal strength, by using an integrated adaptive coupler. (2) The present invention uses the naked eye to observe the light source to judge the quality of the optical fiber and the automated algorithm response evaluation system to realize real-time dynamic monitoring of the quality of the optical fiber, ensuring the consistency and repeatability of the detection. The present invention is not only fast in detection and simple in operation, but also has significant advantages in cost-effectiveness and ease of application. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of a high-sensitivity detection method for PD-L1 protein based on optical fiber surface plasmon resonance according to the present invention.

[0021] Figure 2This is a schematic diagram of the structure of a highly sensitive PD-L1 protein detection device based on optical fiber surface plasmon resonance according to the present invention. Figure 2 (a) In the middle: 1. X, Y, Z axis displacement stage; 2. Power supply; 3. Y-type optical fiber; 4. Light source; 5. Displacement stage controller; 6. 96-well sample slot; 7. Adaptive coupler; 8. Fiber optic fixture; 9. Spectrometer; 10. Detection system; 11. Analysis system; 12. TDN functionalized optical fiber; Figure 2 (b) : 701 piezoelectric ceramic actuator; 702 microlens array; 703 graphene transparent electrode layer; 704 refractive index matched hydroelastic cavity; 705 quartz glass frame; 706 pneumatic interface.

[0022] Figure 3 This is a comparison chart of the refractive index sensitivity of probes with different sensing lengths in Example 1.

[0023] Figure 4 This is a comparison chart of the FWHM of probes with different sensing lengths in Example 1.

[0024] Figure 5 This is a comparison chart of reflection intensity and resonant wavelength obtained by passing pure aqueous solutions with corresponding refractive indices through different gold plating thicknesses of optical fibers in Example 2.

[0025] Figure 6 This is a comparison chart of CVs values ​​for different numbers of data groups collected in Example 3.

[0026] Figure 7 This is a comparison chart of SPR displacements measured in different buffer systems in Example 4.

[0027] Figure 8 This is the calibration curve established in Example 4.

[0028] Figure 9 This is a comparison of the detection results of the method in Example 4 with the gold standard method ELISA.

[0029] Figure 10 This is the calibration curve established in Example 5.

[0030] Figure 11 This is a comparison of the detection results in Example 5 with the gold standard method ELISA.

[0031] Figure 12 This is a schematic diagram of the detection signal in the control group of Comparative Example 1.

[0032] Figure 13 This is a schematic diagram of the detection signal of the experimental group in Comparative Example 1.

[0033] Figure 14 The PD-L1 binding curve is shown in Comparative Example 2, which uses the adaptive coupler of the present invention.

[0034] Figure 15 The PD-L1 binding curves were detected using the comparative device in Comparative Example 2.

[0035] Figure 16 This is a comparison chart of optical power fluctuations over 10 minutes for different couplers in Comparative Example 3.

[0036] Figure 17 The original screw-nut knob coupler was used to detect PD-L1 protein in different buffer systems in Comparative Example 4.

[0037] Figure 18 The calibration curve is shown in Comparative Example 4, which uses the original nut-knob coupler for testing.

[0038] Figure 19 The calibration curve is shown in Comparative Example 5, which uses the original nut-knob coupler for testing.

[0039] Figure 20 The results of the detection using the original nut-knob coupler in Comparative Example 6 are compared with the detection results of the gold standard method ELISA. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0041] First, combine Figure 2 The detection device used in the embodiments of the present invention is described. For example... Figure 2 As shown in (a), the device includes: X, Y, Z axis displacement stage 1; power supply 2; Y-type optical fiber 3; light source 4; displacement stage controller 5; 96-well sample slot 6; adaptive coupler 7; optical fiber holder 8; spectrometer 9; detection system 10; analysis system 11; TDN functionalized optical fiber 12.

[0042] Power supply 2 is used to start detection system 10 and analysis system 11. In detection system 10, adaptive couplers 7 are fixed on X, Y, and Z axis displacement stages 1, and multiple directions of movement are achieved through displacement stage controller 5. Y-type optical fiber 3 is fixed by optical fiber fixer 8. Light source 4 consists of two LEDs. Light source 4 and spectrometer 9 are respectively connected to the bifurcation end of Y-type optical fiber 3, and the other end is optically connected to TDN functionalized optical fiber 12 through adaptive coupler 7. 96-well sample cell 6 is placed below adaptive coupler 7. The TDN functionalized optical fiber 12 can be immersed in the reagent in the sample cell by X, Y, and Z axis displacement stages 1. Spectrometer 9 is used to analyze reflected light and calculate SPR resonance wavelength shift in real time. Analysis system 11 processes and analyzes detection data to realize the detection of PD-L1 protein content.

[0043] like Figure 2 As shown in (b), the adaptive coupler 7 includes the following structure: a piezoelectric ceramic actuator 701, a microlens array 702, a graphene transparent electrode layer 703, a refractive index matching fluid elastic cavity 704, a quartz glass frame 705, and a pneumatic interface 706.

[0044] The adaptive coupler 7 is based on a high-stability quartz glass frame 705, and has an overall rectangular cavity structure. It has a collaborative architecture of "dual optical fiber-elastic cavity-control unit" inside. The core consists of four parts: piezoelectric ceramic actuator 701, microlens array 702, graphene transparent electrode layer 703, and refractive index matching liquid elastic cavity 704.

[0045] Among them, the piezoelectric ceramic actuator 701 adopts a dual-axis symmetrical distribution design. Two sets of piezoelectric ceramic plates are fixed to both ends of the frame through titanium alloy connectors. Its output end is rigidly connected to the fiber optic mounting base through a flexible hinge. The arc-shaped transition structure of the flexible hinge (curvature radius 0.5 mm) can effectively eliminate mechanical transmission errors and ensure that the actuator displacement control resolution is stably maintained at 5-10 nm. The DC voltage signal (0-100V) input by the external high-precision signal generator can drive the fiber optic mounting base to drive the fiber optic probe and the Y-shaped fiber to achieve micro-displacement adjustment within ±50 μm along the optical axis, and precisely control the distance between the two fiber end faces (adjustment accuracy 0.1 μm).

[0046] The microlens array 702 is fabricated using photolithography-etching technology. It has an overall circular thin-film structure (8 mm in diameter and 1 mm in thickness). The array unit is a regular hexagonal microlens (unit diameter 50 μm, focal length 100 μm), which is bonded to the outer side of the two optical fiber end faces using UV-cured adhesive (coaxiality error with the optical fiber ≤2 μm). Each microlens unit is precisely aligned with the corresponding optical fiber core, which can focus and compress the Gaussian beam emitted from the optical fiber, reducing the spot diameter from 10 μm to 3 μm and significantly improving the optical field coupling density. Four micro-adjustment holes (0.5 mm in diameter) are set on the edge of the array. With the help of the precision adjustment frame on the frame (adjustment accuracy 0.01 mm), the three-dimensional attitude fine adjustment of the microlens array and the optical fiber can be realized to ensure the optical axis coincidence.

[0047] A graphene transparent electrode layer 703 covers the front and rear inner walls of the refractive index matching fluid elastic cavity 704. The electrode layer thickness is controlled at 5-10 nm, and the light transmittance is ≥95%. Two sets of gold-plated electrode pins (0.2 mm in diameter) are led out from the electrode layer to form a closed loop with an external photodetector (detection accuracy 1 nW). The light intensity change signal during the coupling process of the two optical fibers can be collected in real time (sampling frequency 1 kHz). The coupling state is judged by signal feedback. Due to the atomic-level thickness characteristics of graphene, it will not cause significant attenuation or scattering of light transmission.

[0048] The core adjustment component is the refractive index matching liquid elastic cavity 704, made of highly elastic polydimethylsiloxane (PDMS) material (Shore hardness 30A). The cavity is filled with silicone oil (refractive index 1.458, 25℃) that matches the refractive index of the optical fiber. The upper and lower ends of the cavity are sealed with a sealing film (thickness 5 μm, light transmittance 98%) to achieve liquid sealing. The side of the cavity is connected to a pneumatic interface 706. Through a closed-loop control system consisting of a pneumatic pump (pressure adjustment range 0-50kPa) and a pressure sensor (accuracy 0.1 kPa), the pneumatic pressure in the cavity can be adjusted, so that the cavity thickness can be continuously varied in the range of 100-200 μm (adjustment step 1 μm). This adapts to the morphological differences of different optical fiber end faces caused by processing errors, ensuring that the matching liquid and the optical fiber end face are tightly bonded and eliminating light reflection loss caused by air gaps.

[0049] Example 1 This embodiment presents an integrated adaptive coupler refractive index sensitivity detection method based on fiber surface plasmon resonance, as follows: Step 1: One-time fiber fabrication. A 16 cm section of the original fiber is cut. Next, cladding stripping is performed at 6.1, 6.2, 6.2, 6.3, 6.4, and 6.5 cm from both ends of the fiber, retaining specific sections with protective layers in the middle at 3.0, 3.2, 3.4, 3.6, and 3.8 cm. Then, the cladding material in the stripped areas is completely removed by acetone treatment. Immediately afterwards, both ends are cut to fabricate one-time fibers with sensing areas of 4, 6, 8, 10, and 12 mm, respectively, with the other end being 0.1 cm. Finally, a 45±5 nm gold layer is formed on the sensing surface of the fiber using a magnetron sputtering device.

[0050] Step 2: Fix each probe onto the test platform in sequence and immerse it in a series of standard solutions with known refractive indices (pure water, sucrose solution (Suc), ETOH) and collect their reflectance spectra in sequence. As the sensing length increases, the resonant wavelength exhibits a redshift, and the reflectance spectra of all probes show a consistent redshift with increasing solution refractive index. By extracting the resonant wavelengths under solutions of different refractive indices, the refractive index sensitivity based on the resonant wavelength is calculated.

[0051] In the formula, This is the offset; The ratio of the refractive index change values ​​represents the sensitivity to the resonant wavelength, such as... Figure 3 The sensitivity evaluation results show that the values ​​are between 1800 nm / RIU and 2100 nm / RIU, and the probe with a sensing length of 6 mm exhibits the best sensitivity.

[0052] Furthermore, it was found that the longer the exposed sensing length, the wider the full width at half maximum (FWHM), as demonstrated by evaluating the fiber quality factor (FOM) for each sensing region.

[0053] The results are as follows Figure 4 As shown, the average FWHM values ​​for fiber optic probes with sensing lengths of 6, 8, 10, and 12 mm, corresponding to pure water with a refractive index of 1.3329, are approximately 99, 106, 114, and 120 nm, respectively.

[0054] Example 2 This embodiment is similar to Embodiment 1, except that the optimal SPR signal of the prepared sensor is determined by comparing the reflection intensity and resonant wavelength obtained from pure aqueous solutions with corresponding refractive indices for different optical fiber gold plating thicknesses (30, 50, 70, and 90 nm, respectively). The results are as follows: Figure 5As shown, the SPR signal is optimal when the gold film thickness is 50 nm, and the corresponding SPR resonance absorption peak usually appears at a wavelength of around 620 nm.

[0055] Example 3 This embodiment presents an integrated adaptive coupler stability detection method based on fiber surface plasmon resonance, the method of which is as follows: First, a 16 cm section of the original optical fiber is cut. Next, the cladding is stripped at 6.2 cm intervals at both ends of the fiber, leaving a 3.6 cm section in the middle with a protective layer. Then, the cladding material in the stripped area is completely removed by acetone treatment. Immediately afterward, both ends are cut to ensure that the length of the sensing area is 0.6 cm at one end and 0.1 cm at the other. Finally, a 50 nm gold layer is formed on the surface of the optical fiber sensing device using a magnetron sputtering device.

[0056] Then, the entire coupling system was adjusted to be near the optimal coupling point (average light intensity baseline of 100 μW), the piezoelectric ceramic driving voltage was set to a fixed value, the closed-loop feedback control was turned off, and the detection environment temperature was ensured to be 25℃.

[0057] Finally, the disposable optical fiber was immersed in pure water. Using software with a fixed time interval of 1 second, the data acquisition program was started, collecting one light intensity data point per second. 20, 40, 60, 80, and 100 sets of data were continuously collected. Repeatability experiments were performed using three optical fibers, and their CVs values ​​were calculated for each. Figure 6 As shown, the results indicate that when n=20, the CVs are all greater than 1.5%, and the conclusion is unreliable. When n=40, the CVs fluctuate around 1.5%, while when n=60, 80 and 100, the CVs value gradually stabilizes at around 0.7%. Since collecting 60 sets of data continuously is sufficient to verify the system stability, subsequent experiments all use collecting 60 sets of data continuously.

[0058] Example 4 This embodiment describes a method for detecting PD-L1 under different buffer systems based on an adaptive coupler integrated with fiber optic SPR, including the following steps: S1. Cut a 16 cm section of the original optical fiber; next, strip the cladding at 6.2 cm intervals at both ends of the fiber, leaving a 3.4 cm section in the middle with a protective layer; then thoroughly remove the cladding material from the stripped area using acetone; next, cut both ends to ensure the sensing area is 0.8 cm long at one end and 0.1 cm long at the other; finally, use a magnetron sputtering device to form a 50 nm gold layer on the optical fiber sensing surface.

[0059] S2. Prepare DNA tetrahedral framework nucleic acid (TDN). Take 1 μL of each of the four single strands (original concentration 100 μM), add 36 μL of Tris-magnesium™ buffer, 10 μL of tris(2-carboxyethyl)phosphonic acid hydrochloride (TCEP, concentration 60 mM), and 50 μL of enzyme-free and pyrogen-free water to form a tube. React at room temperature for more than 1 h. After annealing the mixture at 95 ℃ for 10 min, slowly cool it to 4 ℃ to form TDN (concentration 1 μM). Dilute the TDN to 50 nM concentration using TM buffer. The sequences of the four single strands are as follows: S1:5'-SH-TATCACCAGGCAGTTGACAGTGTAGCAAGCTGTAATAGATGCGAGGGTCCAATAC-3' S2:5'-SH-TCAACTGCCTGGTGATAAAACGACACTACGTGGGAATCTACTATGGCGGCTCTTC-3' S3:5'-SH-TTCAGACTTAGGAATGTGCTTCCCACGTAGTGTCGTTTGTATTGGACCCTCGCAT-3' S4:5'-CACCCCACCTCGCTCCCGTGACACTAATGCTATTTTTTTTTTACATTCCTAAGTCTGAAACATTACAGCTTGCTACACGAGAAGAGCCGCCATAGTA-3'.

[0060] The bare optical fiber sensing area was then immersed in the TDN solution and incubated overnight (12 h), meaning that the TDN was fixed to the gold film surface of the plasma-treated bare optical fiber through Au-S bonds (optical fiber core diameter 600 μm, gold film thickness 50 nm).

[0061] S3. Connect the TDN functionalized fiber to the Y-type fiber optic according to the knob and adaptive coupler. Generate a certain wavelength through reflection by the light source. Rotate the light source knob to adjust the light intensity to 60000±500 RIU and stabilize for 10 minutes. Determine whether the light source is stable based on the magnitude of wavelength fluctuation. S4. Based on computer software evaluation of the stability of the optical fiber for sample detection, in order to obtain the S-spectrum. Figure 1The curvature and shape of the curve are determined, and the average value is obtained through an adaptive deep learning algorithm to adjust the position of the disposable fiber and the Y-type fiber. The adaptive coupler used in this embodiment achieves nanometer-level control of the coupling gap (resolution ≤10nm) through a precision displacement mechanism. It adopts an elastic cavity design filled with refractive index matching liquid, and the cavity thickness is continuously varied at 150±20 μm by gas pressure adjustment to adapt to the morphological differences of different fiber end faces. A graphene transparent electrode layer is introduced at the coupling interface to maintain the optical transmission characteristics and monitor the coupling state in real time. In particular, the calibration process adopts closed-loop feedback control, which automatically adjusts the driving voltage of the piezoelectric ceramic (50±10 V) with an accuracy of 1 mV by monitoring the changes in output light intensity (responsivity of 0.5 A / W) until the optimal coupling point is found (the light intensity reaches the baseline position and the fluctuation is ≤±1% for 10 s). To ensure system stability (<2 h), 60 sets of light intensity data are continuously collected. I 1 , I 2 ,…,I 60 ) Calculate the coefficient of variation ( CV ), requiring light intensity fluctuation rate ( CV s ≤1.5% )×100% in, This represents the average value of 60 consecutively collected light intensity data sets.

[0062] To ensure consistent coupling efficiency during repeatability testing, after each removal and reinstallation of the disposable fiber, a laser interferometer (accuracy 0.8 nm) is used to measure the deviation between the adaptive coupler's automatic reset Y-type fiber and the disposable fiber, which must be less than 15 nm.

[0063] S5. Next, the SPR signal is acquired through an optical signal detection system, and the concentration or other properties of the analyte are calculated using the formulas for the changes in reflectance and refractive index. Environmental interference is eliminated by combining this with a real-time compensation algorithm for the refractive index matching fluid. The data processing process uses a mathematical model combined with an adaptive algorithm to dynamically optimize the data processing results and improve sensitivity.

[0064] The initial reflectivity, The response constant of the material, The incident angle corresponding to the SPR peak is given. By measuring the reflectance changes at different incident angles, the refractive index change of the analyte can be further calculated. The data are optimized for signal-to-noise ratio using adaptive Kalman filtering (AKF) (SNF improvement ≥10 dB), and combined with locally weighted regression smoothing curves to ensure that the refractive index calculation error is ≤5×10⁻⁶. -5 RIU.

[0065] Finally, the relevant performance verification needs to meet the following requirements: In the stability verification, under the condition of consistent environment and reagents, the light intensity fluctuation should be monitored under random vibration conditions. CV s ≤2%.

[0066] Then, dual-channel optical response signals are acquired, the average response per second is extracted, a sample dataset is constructed, the extracted spectral data is averaged, and the average SPR resonance wavelength shift is obtained. The formula for calculating the average SPR resonance wavelength shift is as follows:

[0067] In the formula, Average SPR resonance wavelength shift; To collect 60 SPR resonance wavelengths within 1 second; It is the set of SPR resonance wavelengths from the 51st to the 60th within 1 second.

[0068] S6: Optimize the buffer system detection conditions to detect PD-L1 protein and establish a calibration curve (PD-L1 concentration range: 0-15 ng / mL). For the capture stage, PD-L1 is captured by PD-L1 aptamers on TDN-functionalized optical fibers. For the detection stage, signal amplification is performed using commercial gold nanoparticles (AuNPs), and anti-PD-L1 monoclonal antibody is modified by electrostatic adsorption (concentration: 5 μg / mL, binding condition: 25℃ shaking for 1 hour). Finally, unbound AuNPs sites are blocked with BSA, and the UV absorbance (OD) is 0.5.

[0069] In the detection steps, the PD-L1 concentration was set to 6.25 ng / mL in different buffer systems (including 10 mM PBS, 10 mM PBST, MES, NaAc, TBST, pH range 4.0-8.0). Each sample was incubated with the fiber optic probe for 30 min, washed with PBS / BSA, and then incubated with the antibody-AuNPs complex for 30 min.

[0070] Furthermore, during the entire operation of the fiber optic probe, the movement distance of the fiber optic probe is first controlled by the X and Y axis orientation of the X, Y, and Z axis displacement stage (the movement distance of this platform is completely matched with the position of the 96-well sample slot) to reach the position directly above the sample-filled 96-well sample slot. Then, the fiber optic probe is precisely placed in each sample well by the Z axis orientation. The first step involved immersing the fiber optic probe in a sample well containing TM buffer. Unbound TDN was then washed away using TM buffer, three times for 5 seconds each time (i.e., the fiber optic probe was moved up and down three times along the Z-axis using a displacement stage to ensure the sensing area of ​​the fiber was completely immersed in and completely removed from the solution). This process was repeated. The second step involved washing away the TM buffer using a buffer diluted with PD-L1, three times for 5 seconds each time. The third step involved stabilizing the baseline using a buffer diluted with PD-L1 for 600 seconds. The fourth step involved capturing PD-L1 for 1800 seconds. The fifth step involved washing away non-specifically adsorbed PD-L1 for 60 seconds. The sixth step involved stabilizing the baseline using PBS / BSA buffer for 360 seconds. The seventh step involved signal amplification using antibody-gold nanoparticles for 1800 seconds. Finally, unbound antibody-gold nanoparticles were washed away with PBS / BSA buffer for 60 seconds. The results are as follows: Figure 7 As shown, each experiment was repeated three times. The pH 6.0 MES buffer system was optimal for detecting PD-L1 protein using this system. A calibration curve was established, and the results are shown below. Figure 8 As shown, R 2 The detection efficiency reached 0.99, with a LOD value of 80 pg / mL and a detection range of 340 pg / mL to 15 ng / mL. Through the above embodiments, this device can achieve highly sensitive detection of PD-L1 in a buffer system under low-cost, rapid, and high-precision conditions, as shown in the results. Figure 8 As shown.

[0071] S7. The constructed high-sensitivity detection device and method for PD-L1 protein based on fiber surface plasmon resonance has been applied to the detection of clinical samples.

[0072] Specifically, the clinical sample was the patient's serum. The serum fraction was separated by centrifugation at 1500 rpm and using heparinized tubes, then stored at -80°C. Before use, the sample was thawed at 4°C and brought to room temperature. Next, the serum sample underwent pretreatment by diluting it with pH 6.0 MES to adjust the ionic strength and pH. Subsequently, the pretreated serum sample was detected using a fiber optic SPR detection device. The detected portion was the sensing area of ​​the fiber optic probe, whose surface was immobilized with a framework nucleic acid containing a PD-L1 aptamer as a biorecognition element. The nucleic acid strands were firmly bound to the fiber optic metal coating via Au-S bonds, forming a stable sensing interface. When the fiber optic sensing area was immersed in the detection solution, the PD-L1 protein in the sample specifically bound to the PD-L1 aptamer. The signal was amplified by PD-L1 antibody-gold nanoparticles, and the SPR signal change curve was monitored in real time. After data acquisition, the SPR signal was analyzed using built-in software algorithms. The binding constant and PD-L1 concentration were calculated by fitting the binding curve and compared with a standard curve. During the clinical validation phase, this embodiment selected serum samples from patients with non-small cell lung cancer and compared them with the gold standard method ELISA. The results showed that the results of this method were consistent with those of the ELISA method, as follows: Figure 9 As shown in the diagram. This application approach allows the method to be directly integrated into clinical workflows, such as for preoperative screening, treatment decision-making, or efficacy monitoring, enabling bedside testing via portable devices and reducing reliance on large instruments. Overall, this method overcomes the challenges of PD-L1 detection in clinical samples, such as small sample volumes or low concentrations, through the high sensitivity and real-time performance of fiber optic SPR technology, providing a practical tool for translational medicine.

[0073] Example 5 This embodiment is a method for detecting PD-L1 in spiked serum based on an adaptive coupler integrated with fiber optic SPR. This embodiment is the same as embodiment 4, except for the following steps: In step 1, during the one-time optical fiber fabrication, while maintaining a total length of 4.3 cm, a specific 3.6 cm section with a protective layer is retained in the middle; the length of the sensing area is ensured to be 0.6 cm, and the other end is 0.1 cm; finally, a 40 nm gold layer is formed on the optical fiber sensing surface using a magnetron sputtering device.

[0074] Step 2: Prepare DNA tetrahedral framework nucleic acid (TDN). Take 1 μL of each of the four single strands (original concentration 100 μM), add 36 μL of Tris-magnesium™ buffer, 10 μL of tris(2-carboxyethyl)phosphonic acid hydrochloride (TCEP, concentration 95 mM), and 50 μL of enzyme-free and pyrogen-free water to form a tube. React at room temperature for more than 1 h. After annealing the mixture at 95 °C for 10 min, slowly cool it to 4 °C to form TDN (concentration 1 μM). Dilute the TDN to a concentration of 25 nM using TM buffer.

[0075] Next, the bare optical fiber sensing region was immersed in TDN solution and incubated overnight (12 h), meaning the TDN was fixed to the surface of the plasma-treated bare optical fiber gold film (fiber core diameter 600 μm, gold film thickness 50 nm) via Au-S bonds. A calibration curve was established in this method. Through the above embodiments, this device can achieve highly sensitive detection of PD-L1 in spiked serum under low-cost, rapid, and high-precision conditions, as shown in the results. Figure 10 As shown.

[0076] This example demonstrates the detection of PD-L1 protein in clinical samples using an adaptive sensor based on fiber optic SPR integration, and compares it with the ELISA method. Similar to Example 4, this implementation constructs a calibration curve, and the results are as follows. Figure 11 As shown, this enables highly sensitive quantitative detection of PD-L1.

[0077] Comparative Example 1 This comparative example is a device for high-sensitivity detection of PD-L1 based on an adaptive coupler integrated with fiber optic SPR and a conventional coupler. The device used is basically the same as in Example 4, except that the adaptive coupler in the control group is replaced with a conventional coupler (THORLABS, 10440A).

[0078] Specifically, to verify the advantages of the adaptive coupler proposed in this invention compared to traditional ordinary couplers, key indicators such as signal-to-noise ratio (SNR), system stability (through error bar analysis), and spectral wavelength adjustment accuracy were examined, including the following steps: First, using the same batch of PD-L1 standard samples, tests were conducted on both an ordinary coupler (THORLABS, 10440A) system (control group) and an adaptive coupler system (experimental group). In the baseline stability test, the fiber optic sensing probe was immersed in water, a 5% sucrose solution, a 10% sucrose solution, and ETOH, respectively, and spectral data were continuously acquired for 2 hours at a sampling frequency of 1 Hz.

[0079] Then, the experimental group used an adaptive coupler to monitor the fiber coupling efficiency in real time. When the axial deviation between the disposable fiber and the Y-type fiber was detected to exceed 5 μm, the piezoelectric ceramic was automatically triggered to perform three-dimensional position compensation. In contrast, the control group used a mechanically fixed structure for the ordinary coupler, and the optical path was only ensured by initial manual adjustment.

[0080] Furthermore, such as Figure 12 As shown, when the resonance wavelength is maintained at approximately 620-650 nm and the dip value is approximately 60-80, the control group exhibits significant wavelength instability; for example... Figure 13 As shown, the wavelength curve of the control group was significantly stabilized after adjustment by the adaptive coupler.

[0081] Then, in the stability test, the two systems ran continuously for 2 hours to monitor the resonant wavelength drift. The target resonant wavelength was set at 632.8 nm (the theoretical SPR wavelength when the gold film thickness is 50 nm). The experimental group stabilized to 632.8 ± 0.2 nm within 12 seconds through the closed-loop feedback system, and automatically identified and eliminated 3 fiber ends with damage (judged by Rayleigh scattering intensity > 5%). The control group was unable to compensate for the 8° tilt error of the fiber end face, and finally stabilized at 630.4 ± 2.1 nm, and required manual replacement of 2 fibers with insufficient coupling efficiency.

[0082] As shown in Table 1, in the detection of 5 ng / mL PD-L1 samples, the average signal-to-noise ratio of the experimental group reached 42.5 dB (±1.1 dB standard deviation), significantly higher than the 18.7 dB (±2.3 dB) of the control group. This was mainly attributed to the adaptive coupler controlling the light intensity fluctuation within 3.0% (compared to 12.3% in the control group). The error bars of the experimental group were all less than 0.1 nm, while those of the control group were all greater than 0.1 nm. This indicates that the adaptive coupler performed well and had high detection stability.

[0083] Table 1. Performance Comparison of Conventional Couplers and Adaptive Couplers

[0084] Finally, in the wavelength positioning accuracy test, the target resonant wavelength was set to 620 nm. Three types of optical fibers of different qualities were used for testing: premium type (650 nm wavelength transmission loss ≤0.2 dB / km, refractive index uniformity error ≤1 × 10⁻⁻⁻⁶). 6 Medium (650nm wavelength transmission loss ≤0.5dB / km, refractive index uniformity error ≤5×10⁻) 6 ); Inferior type (transmission loss ≤1.0dB / km, no obvious fiber breaks or severe dispersion, only meeting the basic requirement of "optical path continuity"). The above classification is based on the international standard IEC 60793-1-40:2024.

[0085] Furthermore, while conventional couplers require repeated manual adjustments (averaging 10 min per adjustment), the adaptive coupler automatically identifies the fiber grade through a built-in quality assessment algorithm (based on insertion loss and spectral half-width at half-maximum) and drives a three-dimensional micro-displacement stage (stroke ±2 mm, repeatability 0.5 μm) for compensation and adjustment.

[0086] Furthermore, as shown in Table 2, only high-quality optical fibers achieved the target (100% success rate), medium-quality optical fibers had a success rate of 42%, and low-quality optical fibers failed completely. All optical fibers reached the target wavelength within 120 seconds, with a wavelength positioning accuracy of 650 ± 0.12 nm (k=3).

[0087] Table 2. Debugging results of different types of optical fibers under adaptive coupler and conventional coupler operation

[0088] In this comparative example, the adaptive coupler, through closed-loop control, improves the signal-to-noise ratio by 127%, increases wavelength positioning accuracy by 3 times, reduces the error bar by 63%, and completely eliminates the need for manual adjustment. These advantages directly translate into improved detection sensitivity—specifically, the detection limit for PD-L1 protein. This demonstrates that the improved device of this invention can achieve highly sensitive detection of PD-L1 protein.

[0089] Comparative Example 2 This comparative example compares the impact of differences in the degrees of freedom of the adaptive coupler on detection performance. Specifically, it compares movement in only the X and Y directions versus movement with all degrees of freedom, as follows: Step 1: The proportional coupler contains only two precision electric displacement stages, which control the X-axis (forward and backward) and Y-axis (left and right) movement respectively. The clamps used to fix the disposable optical fiber and the Y-type optical fiber are connected by a nut and knob. It does not have any active adjustment function in the Z-axis (up and down) direction, nor does it have any angle fine adjustment function. The end faces of the two optical fibers are considered to have been roughly adjusted to approximately the same horizontal plane during initial assembly, but real-time fine calibration cannot be performed in the experiment. This fully free degree adaptive coupler can independently or in conjunction control the linear displacement of the X, Y, and Z axes as well as the rotation around the X, Y, and Z axes. The clamp design allows for all-round precise adjustment of the spatial attitude of the two optical fibers (mainly through integrated control software and optical power feedback system, which can automatically perform "finding the optimal coupling point").

[0090] Step 2: The proportional device control software drives the X and Y platform to perform a region scan, with a scanning range of ±10 μm and a step size of 0.5 μm; the output light intensity value corresponding to each coordinate point is recorded in real time; the light intensity baseline value (RIU=60000) is determined as the "optimal coupling position", and the platform is moved and fixed at this position; the device of this invention first performs a coarse scan of the XY plane to find the region with RIU around 60000, starts the Z-axis scan, and near the optimal XY coordinates, controls the Z-axis to move in 0.1 μm steps within a range of ±2 μm to find the light intensity peak; then, under the optimal XY and Z coordinates, the yaw and pitch angles (±0.5°) are finely adjusted to further optimize the optical path accuracy.

[0091] Step 3: In MES buffer, at the optimal coupling position, acquire and record a stable SPR resonance spectrum to determine the initial resonance wavelength.

[0092] Step 4 involves analyzing a series of PD-L1 protein standard solutions of different concentrations (1000 ng / mL, 100 ng / mL, 10 ng / mL, 1 ng / mL, and 0.1 ng / mL) for 30 min at each concentration, followed by MES washing for 10 min. Spectrometers continuously acquire spectra throughout the binding process.

[0093] The coupling position of the comparative amplifier remains fixed at the point determined in the second step; the adaptive coupler is always in active closed-loop feedback mode. After acquiring several spectra, the system slightly adjusts all degrees of freedom (with extremely small amplitudes, approximately tens of nm) to detect the trend of light intensity changes and makes real-time reverse adjustments to dynamically maintain optimal coupling, compensating for potential coupling efficiency degradation caused by liquid flow, minor temperature fluctuations, or mechanical creep. The results are as follows: Figure 14 , 15 As shown, the device of the present invention detects a smooth PD-L1 binding curve with a high signal-to-noise ratio, a clear binding / dissociation process, and a rapid and stable baseline fall back; the comparative device detects a PD-L1 binding curve with many spikes and fluctuating signals during the binding process.

[0094] The results in Table 3 show that the device of the present invention, due to its ability to optimize the Z-axis and angle, finds a global optimum with much higher coupling efficiency, improving the initial signal strength by about 70%. The comparative device, due to its non-optimal and fixed coupling state, is sensitive to environmental disturbances and has high baseline noise. The device of the present invention, through dynamic maintenance, achieves extremely stable baseline. Although the present invention takes slightly longer, it yields a better and more stable coupling state, and this time cost is completely acceptable in high-sensitivity detection.

[0095] Table 3 Comparison of Coupling Efficiency and Signal Stability

[0096] The device of this invention significantly outperforms the comparative device in all key indicators, including detection limit, signal strength, repeatability, and signal quality. In particular, the detection limit is improved from ng / mL to pg / mL, and the stability RSD is optimized from 0.85% to 0.12%.

[0097] Comparative Example 3 This comparative study compares the impact of the adaptive coupler's degree of freedom of movement on the PD-L1 detection performance. Similar to Comparative Example 2, the difference lies in that the coupler in this comparative device can only move along the Z-axis. Other positions require manual control to find the maximum optical power point at the current XY-plane position, record this power value, and maintain this state while continuously monitoring and recording the optical power fluctuations over 10 minutes. The results are as follows: Figure 16 As shown.

[0098] The results in Table 4 show that the device of the present invention can find a global optimum with much higher coupling efficiency because it can perform XY axis and angle optimization, and the initial signal strength is improved by about 75%. The comparative device is sensitive to environmental perturbations and has baseline noise because the coupling state is not optimal and is fixed.

[0099] Table 4 Comparison of Coupling Efficiency and Signal Stability

[0100] Comparative Example 4 This comparative example is the same as Example 4, except that the original screw-nut knob coupler was used to detect PD-L1 protein in different buffer systems. The results are as follows: Figure 17 , 18 As shown.

[0101] Comparative Example 5 This comparative example is the same as Example 5, except that the original screw-nut knob coupler was used to detect spiked serum PD-L1 protein. The results are as follows: Figure 19 As shown.

[0102] Comparative Example 6 This comparative example is the same as Example 4, except that the original screw-nut knob coupler was used to detect PD-L1 protein in clinical samples. The results are as follows: Figure 20 As shown.

Claims

1. A method for high-sensitivity detection of PD-L1 protein based on optical fiber SPR, characterized in that, The TDN functionalized fiber and the Y-type fiber are optically connected by an adaptive coupler. After the light source is connected and the light source and optical path are stabilized, the unknown sample is injected into the liquid tank and specifically binds to the PD-L1 aptamer on the surface of the TDN functionalized fiber. The signal is amplified by PD-L1 antibody gold nanoparticles. Then, the SPR signal is collected by the optical signal detection system and combined with the regression standard curve to obtain the PD-L1 protein concentration in the unknown sample. 2.The high-sensitivity PD-L1 protein detection method based on fiber SPR according to claim 1, wherein, The TDN functionalized optical fiber refers to a DNA tetrahedral framework nucleic acid containing the PD-L1 aptamer sequence that is fixed in a gold film on the surface of the optical fiber sensing through covalent bonding. The fabrication method of the TDN functionalized optical fiber is as follows: S1.1: Cut the original optical fiber to a length of 15-17cm, peel off the cladding 6-7cm from both ends of the optical fiber and retain a 3-4cm protective area in the middle, remove the cladding material in the peeled area with acetone, cut the ends of the optical fiber to prepare a disposable optical fiber with a sensing length of 6-12mm, and then deposit a gold film with a thickness of 35-55nm on the sensing area using magnetron sputtering. S1.2: Take 1 μL of each of the four DNA single strands with an original concentration of 100 μM, add 36 μL of TM buffer, 10 μL of TCEP solution with a concentration of 60~100 mM, and 50 μL of enzyme-free and pyrogen-free water to form a tube, react at 0~40℃ for more than 1 h, anneal the mixed solution at 95 ℃ for 10 min and then slowly cool it to 4 ℃ to form TDN solution; S1.3: Dilute the TDN solution to 50 nM using TM buffer, then immerse the fiber sensing area in the TDN solution and incubate overnight for at least 12 hours. 3.The high-sensitivity PD-L1 protein detection method based on fiber SPR of claim 1, wherein, The adaptive coupler adjusts the relative position of the TDN functionalized fiber and the Y-type fiber through an internal piezoelectric ceramic driver. At the same time, the graphene transparent electrode layer monitors the light intensity change signal during the coupling process of the two fibers in real time and outputs it to an external optical signal detection system. Based on the feedback from the optical signal detection system, the piezoelectric ceramic driver calibrates the relative position between the end faces of the two fibers to find the optimal coupling point. When the feedback reaches the optimal coupling point, the current relative position of the two fibers is locked. 4.The high-sensitivity PD-L1 protein detection method based on fiber SPR of claim 3, wherein, The stability of the light source and optical path specifically refers to: S2.1: Connect the TDN functionalized fiber to the Y-type fiber using the knob and adaptive coupler. A certain wavelength is generated by reflection through the light source. Adjust the light intensity to 60000±500 RIU by rotating the light source knob and stabilize for 10 minutes. Determine whether the light source is stable based on the wavelength fluctuation. S2.2: The calibration process of the adaptive coupler adopts closed-loop feedback control. By monitoring the change in output light intensity, the driving voltage of the piezoelectric ceramic is automatically adjusted until the optimal coupling point is found, that is, the light intensity reaches the baseline position and the fluctuation is ≤±1% for 10s. Continuous collection of 60 groups of light intensity data (I1, I2, …, I 60 ) to calculate the coefficient of variation ( CV ), requirements CV s ≤1.5%: )×100%； wherein, is the average of 60 groups of light intensity collected continuously; The responsivity of the monitored output light intensity is 0.5 A / W, the driving voltage of the piezoelectric ceramic is 50±10 V, and the accuracy is 1mV. 5.The high-sensitivity PD-L1 protein detection method based on fiber SPR of claim 1, wherein, The acquisition of SPR signals through the optical signal detection system specifically includes: S3.1: Calculate the characteristics of the measured substance according to the formula of the change of reflectivity and refractive index: Wherein, R0 is the initial reflectivity, A is the response constant of the material, θ0 is the incident angle corresponding to the SPR peak; the data is optimized by adaptive Kalman filtering to improve the signal-to-noise ratio, and the curve is smoothed by local weighted regression to ensure that the calculation error of the refractive index is less than or equal to 5 x 10⁻ 5 RIU; S3.2: Collect the double-channel optical response signal, extract the average response value per second to construct a sample data set containing peak data and median data; extract the spectral data and calculate the average SPR resonance wavelength shift: wherein, is the average SPR resonance wavelength shift; is a set of 60 SPR resonance wavelengths collected in 1 second; is a set of the 51st to 60th SPR resonance wavelengths collected in 1 second. 6.The high-sensitivity PD-L1 protein detection method based on fiber SPR according to any one of claims 1-5, characterized in that, The unknown sample is serum sample diluted by pH 6.0 MES; the unknown sample is incubated with TDN functionalized optical fiber for 20-40 min, and after washing, PD-L1 antibody nanogold is introduced for incubation for 30 min; The PD-L1 antibody nanogold is: commercial gold nanoparticles are used, and anti-PD-L1 monoclonal antibody is modified by electrostatic adsorption, the concentration of PD-L1 monoclonal antibody is 5 μg / mL, the binding condition is 25℃ shaking for 1 hour, and finally BSA is used to block the unbound AuNPs sites, and the obtained product has a UV absorption value of 0.

5. 7.The high-sensitivity PD-L1 protein detection method based on fiber SPR of claim 6, wherein, The regression standard curve is obtained by the following method: A standard solution of PD-L1 protein with a concentration gradient is prepared, each standard solution is incubated with TDN functionalized optical fiber for 20-40 min, and after washing, PD-L1 antibody nanogold is introduced for incubation for 30 min, and the change of SPR resonance wavelength shift is recorded, the concentration of standard solution is taken as the abscissa, and the resonance wavelength is taken as the ordinate, a working curve is drawn, and polynomial curve fitting is performed to obtain the regression standard curve. 8.The high-sensitivity PD-L1 protein detection method based on fiber SPR of claim 7, wherein, During the detection process using the TDN functionalized optical fiber probe, TM buffer cleaning, dilution of PD-L1 buffer cleaning and baseline stabilization, PD-L1 capture, non-specific adsorption of PD-L1 cleaning, PBS / BSA buffer baseline stabilization, PD-L1 antibody nanogold signal amplification and unbound PD-L1 antibody nanogold cleaning operations need to be performed in sequence.

9. A high-sensitivity PD-L1 protein detection device based on fiber-optic SPR for use in the method of any one of claims 1-8, characterized in that, It comprises an X, Y, Z axis displacement table (1), a power supply (2), a Y-type optical fiber (3), a light source (4), a displacement table control instrument (5), a 96-well sample tank (6), an adaptive coupler (7), an optical fiber fixer (8), a spectrometer (9), a detection system (10), an analysis system (11), and a TDN functionalized optical fiber (12).

10. The apparatus of claim 9, wherein, The power supply (2) is used to start the detection system (10) and the analysis system (11); in the detection system (10), the adaptive coupler (7) is fixed on the X, Y, Z axis displacement table (1), and can move in multiple directions through the displacement table control instrument (5); the Y-type optical fiber (3) is fixed through the optical fiber fixer (8), the light source (4) is composed of two LED lamps, the light source (4) and the spectrometer (9) are connected with the bifurcated end of the Y-type optical fiber (3) respectively, the other end is optically connected with the TDN functionalized optical fiber (12) through the adaptive coupler (7); the 96-well sample tank (6) is placed below the adaptive coupler (7), and the TDN functionalized optical fiber (12) can be immersed in the reagent in the sample tank through the X, Y, Z axis displacement table (1); the spectrometer (9) is used to analyze the reflected light and calculate the SPR resonance wavelength shift in real time; the analysis system (11) processes and analyzes the detection data to realize the detection of the content of PD-L1 protein; The adaptive coupler (7) takes a quartz glass frame as a main body, and internally integrates a piezoelectric ceramic driver, a microlens array, a graphene transparent electrode layer and a refractive index matching liquid elastic cavity; the piezoelectric ceramic driver is connected with a fiber fixed seat through a flexible hinge, and can drive two optical fibers to finely adjust the spacing along an optical axis; The microlens array is attached to the outside of the fiber end face and is used for focusing a light spot; the graphene electrode layer is covered on the inner wall of the elastic cavity and is used for monitoring light intensity changes in real time; the refractive index matching liquid elastic cavity changes the thickness through air pressure adjustment, so that the matching liquid is closely attached to the fiber end face, and reflection loss is eliminated.

Images