A kit and method for rapid detection of klebsiella pneumoniae

Abstract

The application discloses a kit and a method for rapidly detecting Klebsiella pneumoniae, and comprises a microfluidic detection chip, reagent components and auxiliary components, wherein the microfluidic detection chip comprises a substrate, and a sample loading area, a mixing reaction area, a Raman detection area and a waste liquid recovery area are arranged on the substrate; the method comprises the following steps: step one, sample pretreatment; step two, automatic mixing reaction; step three, Raman spectrum signal acquisition; step four, intelligent analysis and interpretation; and step five, result output; the application adopts an MXene-Ag composite enhanced substrate and an integrated microfluidic detection chip, mixes and reacts the to-be-tested alveolar lavage fluid sample and the enhanced substrate automatically, collects surface enhanced SERS signals through a portable Raman spectrometer, and intelligently analyzes and identifies the spectrum with the aid of an interpretable Attention-CNN model, so that the method has the advantages of simple operation, rapid detection, high sensitivity, good specificity and suitability for clinical on-site instant detection.

Description

Technical Field

[0001] This invention relates to the field of microbial detection technology, specifically to a kit and method for rapid detection of Klebsiella pneumoniae. Background Technology

[0002] Klebsiella pneumoniae is one of the most common Gram-negative opportunistic pathogens in clinical practice. It is a core pathogen in hospital-acquired pneumonia, ventilator-associated pneumonia, and pulmonary infections, characterized by rapid onset, rapid progression, high drug resistance, and high mortality. Bronchoalveolar lavage fluid (BAL) is a commonly used sample for the etiological diagnosis of pulmonary infections. It directly obtains pathogens from the lower respiratory tract infection site, avoiding contamination by upper respiratory tract flora. Rapid and accurate detection of Klebsiella pneumoniae in BAL is of great significance for clinical diagnosis, treatment, and medication. However, current routine clinical methods for detecting Klebsiella pneumoniae in bronchoalveolar lavage fluid generally have the following drawbacks: traditional bacterial culture methods require steps such as sample inoculation, enrichment, culture, and biochemical identification, resulting in a long culture cycle and a low positive detection rate, which cannot provide timely evidence for early clinical diagnosis and treatment; although technologies such as real-time quantitative PCR and matrix-assisted laser desorption / ionization time-of-flight mass spectrometry have shortened the detection time, they have drawbacks such as complex sample pretreatment, reliance on expensive large benchtop equipment, the need for professional personnel to operate, and the inability to achieve point-of-care testing, making it difficult to promote their application.

[0003] Surface-enhanced Raman scattering (SERS) technology, with its advantages of fingerprint specificity, fast detection speed, high sensitivity, and no need for complex labeling, has shown great application potential in the field of rapid pathogen detection. However, existing SERS-based Klebsiella pneumoniae detection methods still have the following drawbacks:

[0004] 1. The overall performance of SERS-enhanced substrates cannot meet clinical needs: Existing silver-based enhanced substrates are prone to aggregation, have poor stability, and weak signal reproducibility; some MXene-based composite substrates, such as Ag / MXene-Ti3C2 composites, require a 50-70℃ water bath reaction for 24-45 hours, which results in a long preparation cycle; some ternary composite systems incorporating MOFs, such as Ag / MXene / Fe-Co MOF, have problems with complicated processes and high costs, and their structural stability in the complex matrix of bronchoalveolar lavage fluid is insufficient, making it impossible to balance high enhancement performance, rapid preparation, and clinical applicability.

[0005] 2. Extremely low integration and automation of detection systems: Existing technologies mostly focus on optimizing material preparation or single detection steps, failing to form a complete detection system. Most solutions employ manual sample addition, oven / natural drying, and manual selection of detection sites, resulting in cumbersome procedures prone to human error and cross-contamination. Label-based detection solutions require multiple magnetic separation and washing steps, making the process complex and requiring various laboratory equipment such as magnetic separation and centrifugation equipment, which is completely unsuitable for clinical field testing scenarios.

[0006] 3. Lack of specificity optimization for clinical samples, resulting in insufficient anti-interference ability and detection accuracy: Existing SERS bacterial detection technologies mostly use laboratory-cultured bacterial suspensions as the detection target, without considering the interference from complex matrices such as proteins, cell debris, and mucins in bronchoalveolar lavage fluid; broad-spectrum bacterial detection protocols lack specific design targeting the biomolecular characteristics on the surface of Klebsiella pneumoniae, which easily leads to false positive and false negative results in clinical samples, and the detection accuracy cannot meet the requirements of clinical diagnosis;

[0007] 4. Poor equipment adaptability and lack of clinical field testing capability: Existing high-performance SERS detection mostly relies on large benchtop micro Raman spectrometers, which are bulky and expensive, and can only be used in laboratory environments, making it impossible to achieve point-of-care testing; portable Raman spectrometers lack suitable and efficient SERS enhancement systems, making it difficult to achieve accurate detection of low concentrations of pathogens in clinical settings.

[0008] 5. Insufficient intelligence and interpretability in spectral data analysis: Current detection methods mostly rely on manual experience to identify characteristic peaks, which is inefficient and subject to significant subjective interference; some methods use traditional machine learning algorithms such as PCA and SVM, which cannot automatically focus on the characteristic peak regions of Klebsiella pneumoniae, have weak resistance to matrix interference, and cannot provide interpretable decision-making basis, resulting in a black box problem and making it difficult to gain clinical acceptance and promotion.

[0009] In summary, existing technologies cannot simultaneously achieve rapid and low-cost preparation of SERS substrates, highly integrated and automated detection systems, high specificity detection of complex clinical samples, portability, and interpretable intelligent analysis, which severely restricts the practical application of SERS technology in point-of-care clinical testing of Klebsiella pneumoniae. Summary of the Invention

[0010] The purpose of this invention is to provide a kit and method for rapid detection of Klebsiella pneumoniae, in order to solve the problems mentioned in the background art.

[0011] To achieve the above objectives, the present invention provides the following technical solution: a kit for rapid detection of Klebsiella pneumoniae, comprising a microfluidic detection chip, reagent components, and auxiliary components. The microfluidic detection chip includes a substrate, on which a sample loading area, a mixing reaction area, a Raman detection area, and a waste liquid recovery area are disposed. The reagent components include a sample dilution buffer, a positive control, and a negative control. The auxiliary components include a sterile pipette tip, a silver-plated aluminum slide with a focusing groove, and a disposable sample processing tube.

[0012] The substrate is made of transparent PDMS material; the sample loading area is used to add the bronchoalveolar lavage fluid sample to be tested; the mixing reaction area is used to mix the sample with the reinforcing substrate, and a micro-vortex structure is set in the mixing reaction area, and an MXene-Ag composite reinforcing substrate is pre-cured; the Raman detection area is an optically transparent detection window, with a SERS active layer formed by the MXene-Ag composite reinforcing substrate pre-laid for Raman spectroscopy scanning, and a silver-plated aluminum slide with a focusing groove is set below the detection window to achieve focusing of laser and scattered light; the waste liquid recovery area is equipped with a water-absorbing filter membrane to collect waste liquid after detection.

[0013] The microfluidic detection chip relies on capillary force to automatically drive the sample flow.

[0014] The MXene-Ag composite reinforced substrate is Ti3C2T. x The composite colloidal cured layer with uniformly loaded silver nanoparticles on the surface of MXene two-dimensional sheets, and the preparation method of the MXene-Ag composite reinforced substrate are as follows: Ti3C2T... x MXene powder was dispersed in sterile deionized water to prepare a dispersion with a concentration of 0.5-2 mg / mL. The dispersion was sonicated for 30-60 minutes to fully exfoliate the crystal layer. The dispersion was then centrifuged, and the supernatant was collected. The supernatant was placed in a water bath at 30±5℃ under magnetic stirring. Sodium citrate solution and AgNO3 solution were added sequentially, followed by the addition of NaBH4 reducing agent. The mixture was stirred at room temperature for 10-30 minutes. The reaction product was centrifuged and washed to obtain the MXene-Ag composite colloid.

[0015] The sample dilution buffer is used to dilute and adjust the pH of bronchoalveolar lavage fluid samples, maintain the stability of bacterial morphology and biomolecular structure, reduce sample matrix viscosity, and reduce non-specific adsorption; the positive control is used to verify the effectiveness of the kit, and the negative control is used to eliminate contamination interference.

[0016] The sample dilution buffer contained 0.01 mol / L potassium chloride, 0.01 mol / L calcium chloride, and 0.005 mol / L Tris-HCl buffer, with a pH of 7.2-7.4, and was sterilized by filtration through a 0.22 μm filter membrane. The positive control contained a suspension of inactivated Klebsiella pneumoniae standard strain at a concentration of 1 × 10³ CFU / mL, with the above sample dilution buffer as the solvent and 0.01% thimerosal added as a preservative. The negative control used sterile 0.01 mol / L PBS buffer, with a pH of 7.2-7.4, and was sterilized by filtration through a 0.22 μm filter membrane.

[0017] A method for rapid detection of Klebsiella pneumoniae includes the following steps: Step 1, sample pretreatment; Step 2, automated mixing reaction; Step 3, Raman spectral signal acquisition; Step 4, intelligent analysis and interpretation; and Step 5, result output.

[0018] In step one above, 50-100 μL of the bronchoalveolar lavage fluid sample to be tested is taken with a pipette and added to the sample processing tube. 40-70 μL of sample dilution buffer is added and mixed well. 50-100 μL of the diluted sample is taken and added to the sample loading area of ​​the microfluidic detection chip. The microfluidic detection chip is then inserted into the detection slot of the portable Raman spectrometer.

[0019] In step two above, the sample enters the mixing reaction zone under capillary force and comes into full contact with the MXene-Ag composite reinforced substrate, reacting for 5-10 minutes.

[0020] In step three above, after the reaction in step two is completed, the portable Raman instrument is started, automatically scanning the Raman detection area, collecting Raman spectra, and transmitting the collected spectral data to the smart terminal.

[0021] In step four above, the smart terminal performs background subtraction and normalization on the received spectral data, and then inputs it into the Attention-CNN model. The Attention-CNN model outputs the detection results, and the determination of the bronchoalveolar lavage fluid sample is completed based on the detection results.

[0022] In step three, the acquisition parameters of the portable Raman spectrometer are as follows: excitation wavelength 532nm, laser power adaptively adjusted from 0.1 to 1.0mW, integration time 10s, spectral acquisition range 800-1800cm⁻¹, automatic scanning of 3 sites in each detection window, and taking the average spectrum as the final detection spectrum.

[0023] In step four, the Attention-CNN model extracts global and local spectral features through multi-scale convolutional layers, automatically focuses on the characteristic peak regions of Klebsiella pneumoniae through spatial and channel attention modules, amplifies the weights of the characteristic peak regions, suppresses background stray signal interference, and completes automatic spectral classification and recognition.

[0024] In step four, the detection results include classification results, classification confidence, and feature importance heatmap.

[0025] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention uses an MXene-Ag composite enhancement substrate and an integrated microfluidic detection chip to automatically mix and react the bronchoalveolar lavage fluid sample to be tested with the enhancement substrate. The surface-enhanced SERS signal is collected by a portable Raman spectrometer, and the spectrum is intelligently analyzed and identified by an interpretable Attention-CNN model. It has the advantages of simple operation, fast detection, high sensitivity and good specificity, and is suitable for real-time detection in clinical settings. Attached Figure Description

[0026] Figure 1 This is a flowchart of the method of the present invention;

[0027] Figure 2 Transmission electron microscopy image of an MXene-Ag composite reinforced substrate;

[0028] Figure 3 This is a schematic diagram of a microfluidic detection chip structure;

[0029] Figure 4 This is a schematic diagram of the Attention-CNN model structure;

[0030] Figure 5 Comparison of SERS spectra of test results for healthy negative samples and Klebsiella pneumoniae positive samples;

[0031] Figure 6 This is a schematic diagram of a portable Raman spectrometer. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] Please see the appendix Figure 1 - Appendix Figure 6This invention provides an embodiment of a kit for rapid detection of Klebsiella pneumoniae, comprising a microfluidic detection chip, reagent components, and auxiliary components. The microfluidic detection chip includes a substrate with a sample loading area, a mixing reaction area, a Raman detection area, and a waste liquid recovery area. The reagent components include a sample dilution buffer, a positive control, and a negative control. The auxiliary components include a sterile pipette tip, a silver-plated aluminum slide with a focusing groove, and a disposable sample processing tube. The substrate is made of transparent PDMS material. The sample loading area is used to add the bronchoalveolar lavage fluid sample to be tested. The mixing reaction area is used to... The sample is mixed with the reinforcing substrate, and a micro-vortex structure is set in the mixing reaction zone, which is pre-cured with an MXene-Ag composite reinforcing substrate. The Raman detection zone is an optically transparent detection window, pre-coated with a SERS active layer formed by the MXene-Ag composite reinforcing substrate for Raman spectroscopy scanning. A silver-plated aluminum slide with a focusing groove is placed below the detection window to focus the laser and scattered light. A water-absorbing filter membrane is set in the waste liquid recovery zone to collect waste liquid after detection. The microfluidic detection chip automatically drives the sample flow by capillary force. The MXene-Ag composite reinforcing substrate is Ti3C2T. x The composite colloidal cured layer with uniformly loaded silver nanoparticles on the surface of MXene two-dimensional sheets, and the preparation method of the MXene-Ag composite reinforced substrate are as follows: Ti3C2T... x MXene powder was dispersed in sterile deionized water to prepare a dispersion with a concentration of 0.5-2 mg / mL. The dispersion was sonicated for 30-60 minutes to fully exfoliate the crystal layer. The dispersion was then centrifuged, and the supernatant was collected. Under magnetic stirring, the supernatant was placed in a water bath at 30±5℃, and sodium citrate solution and AgNO3 solution were added sequentially. Then, NaBH4 reducing agent was added dropwise, and the reaction was stirred at room temperature for 10-30 minutes. The reaction product was centrifuged and washed to obtain the MXene-Ag composite colloid. The sample dilution buffer was used to dilute and adjust the pH of bronchoalveolar lavage fluid samples, maintain the stability of bacterial morphology and biomolecular structure, reduce sample matrix viscosity, and reduce nonspecificity. Adsorption; the positive control is used to verify the effectiveness of the kit, and the negative control is used to eliminate contamination interference; the sample dilution buffer contains 0.01 mol / L potassium chloride, 0.01 mol / L calcium chloride, and 0.005 mol / L Tris-HCl buffer, pH 7.2-7.4, and is sterilized by filtration through a 0.22 μm filter membrane; the positive control contains a suspension of inactivated Klebsiella pneumoniae standard strain at a concentration of 1 × 10³ CFU / mL, the solvent being the above sample dilution buffer, with 0.01% thimerosal added as a preservative; the negative control uses sterile 0.01 mol / L PBS buffer, pH 7.2-7.4, and is sterilized by filtration through a 0.22 μm filter membrane.

[0034] A method for rapid detection of Klebsiella pneumoniae includes the following steps: Step 1, sample pretreatment; Step 2, automated mixing reaction; Step 3, Raman spectral signal acquisition; Step 4, intelligent analysis and interpretation; and Step 5, result output.

[0035] In step one above, 50-100 μL of the bronchoalveolar lavage fluid sample to be tested is taken with a pipette and added to the sample processing tube. 40-70 μL of sample dilution buffer is added and mixed well. 50-100 μL of the diluted sample is taken and added to the sample loading area of ​​the microfluidic detection chip. The microfluidic detection chip is then inserted into the detection slot of the portable Raman spectrometer.

[0036] In step two above, the sample enters the mixing reaction zone under capillary force and comes into full contact with the MXene-Ag composite reinforced substrate, reacting for 5-10 minutes.

[0037] In step three above, after the reaction in step two is completed, the portable Raman spectrometer is activated, automatically scanning the Raman detection area, acquiring Raman spectra, and transmitting the acquired spectral data to the smart terminal. The acquisition parameters of the portable Raman spectrometer are: excitation wavelength 532nm, laser power adaptively adjustable from 0.1 to 1.0mW, integration time 10s, spectral acquisition range 800-1800cm⁻¹, automatic scanning of 3 sites in each detection window, and taking the average spectrum as the final detection spectrum.

[0038] In step four above, the smart terminal performs background subtraction and normalization on the received spectral data, and then inputs it into the Attention-CNN model. The Attention-CNN model extracts global and local spectral features through multi-scale convolutional layers, automatically focuses on the characteristic peak regions of Klebsiella pneumoniae through spatial and channel attention modules, amplifies the weights of the characteristic peak regions, suppresses background stray signal interference, and completes automatic spectral classification and recognition. The Attention-CNN model outputs detection results including classification results, classification confidence, and feature importance heatmaps, and the determination of bronchoalveolar lavage fluid samples is completed based on the detection results.

[0039] Experimental Example 1:

[0040] To verify the effectiveness of the present invention, the following four samples were tested according to the method described in the embodiments: Sample A was bronchoalveolar lavage fluid from a pneumonia patient, which was cultured and confirmed to contain Klebsiella pneumoniae; Sample B was bronchoalveolar lavage fluid from a healthy person, which did not contain Klebsiella pneumoniae, but was mixed with Escherichia coli and Pseudomonas aeruginosa to verify specificity; Sample C contained 1 ppb (10⁻ 9Standard bronchoalveolar lavage fluid samples of Klebsiella pneumoniae (g / mL); Sample D is a mixed suspension containing 100 ppb Klebsiella pneumoniae and 100 ppb Staphylococcus aureus. The experimental results are shown in Table 1. The experimental results of Sample A show that this method takes less time than traditional culture identification. The experimental results of Sample B show that this method has high selectivity for Klebsiella pneumoniae and is not affected by other common bacteria. The experimental results of Sample C show that this method can accurately identify trace infection samples. The experimental results of Sample D show that this method can specifically identify Klebsiella pneumoniae under the condition of interference from coexisting bacteria.

[0041] Comparative Example 1:

[0042] The Ag / MXene-Ti3C2 composite material was prepared by a dual reduction method using the technical solution disclosed in CN109827945A. Sodium citrate and MXene-Ti3C2 were used as dual reducing agents, and the substrate was prepared by reacting in a water bath at 60℃ for 36 hours. Sample A was taken and manually dropped onto the substrate-modified silicon wafer. After natural drying, the spectrum was collected by a tabletop Raman spectrometer, and the results were determined by manual peak matching. The results are shown in Table 2.

[0043] Comparative Example 2:

[0044] Using the technical solution disclosed in CN202510732193.8, MXene nanosheets and Fe-Co MOF nanomaterials were synthesized sequentially. An Ag / MXene / Fe-Co MOF ternary composite structure was constructed through a multi-step reaction and modified onto a DD-PDMS flexible substrate. Sample A was taken, manually dropped onto the sensor surface, dried in an oven at 60℃, and then Raman spectroscopy was performed. The results were determined by manual peak matching. The results are shown in Table 2.

[0045] Comparative Example 3:

[0046] Using the technical solution disclosed in CN110702662A, IgG@Fe3O4 magnetic beads and boronized SERS tags were prepared. Sample A was taken and subjected to magnetic bead incubation, three magnetic separations and washings, SERS tag labeling, and a second magnetic separation and washing. Finally, the enriched product was transferred to a capillary for Raman detection, and the results were determined by the tag signal. The results are shown in Table 2.

[0047] Comparative Example 4:

[0048] Using the technical solution disclosed in CN120690298A, silver nanoparticles with a diameter of <11nm were prepared as an reinforcing substrate by sodium citrate reduction method. Sample A was taken, manually dropped onto a silicon wafer, and after natural drying, spectra were collected by manually selecting 45 detection sites using an inVia™ benchtop micro Raman spectrometer. The spectra were analyzed by traditional machine learning algorithms such as PCA and SVM. The results are shown in Table 2.

[0049] Table 1. Detection results for different samples

[0050]

[0051] Table 2. Overall Performance Comparison of the Invention and Comparative Examples

[0052]

[0053] Based on the above, the advantages of this invention are as follows: This invention employs an MXene-Ag composite reinforced substrate, where MXene two-dimensional materials provide chemical reinforcement through charge transfer resonance, and silver nanoparticles provide electromagnetic reinforcement through localized surface plasmon resonance, achieving over 10 8 With a high SERS enhancement factor, combined with an integrated microfluidic detection chip featuring a pre-modified enhanced substrate, the sample and substrate are automatically mixed. A portable Raman spectrometer acquires the signal, and a CNN deep learning model based on an attention mechanism precisely focuses on the characteristic peaks of Klebsiella pneumoniae while suppressing interference from the complex matrix of bronchoalveolar lavage fluid. The system boasts high efficiency (rapid room temperature substrate preparation, detection in just 7-10 minutes), ultra-high sensitivity (detection limit as low as 0.5 ppb), high specificity (specifically identifying target bacteria without interference), high stability (fully automated from sample input to result output without cross-contamination), and interpretability (outputting characteristic heatmaps). It is suitable for point-of-care testing in clinical settings, completely solving the core shortcomings of traditional testing methods such as long testing cycles, cumbersome substrate preparation, bulky equipment, and reliance on manual analysis. This fully meets the needs of rapid and accurate screening for pathogens causing lung infections in clinical settings.

[0054] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A kit for rapid detection of Klebsiella pneumoniae, comprising a microfluidic detection chip, reagent components, and auxiliary components, characterized in that: The microfluidic detection chip includes a substrate with a sample loading area, a mixing reaction area, a Raman detection area, and a waste liquid recovery area. The reagent components include a sample dilution buffer, a positive control, and a negative control. The auxiliary components include a sterile pipette tip, a silver-plated aluminum slide with a focusing groove, and a disposable sample processing tube.

2. The kit for rapid detection of Klebsiella pneumoniae according to claim 1, characterized in that: The substrate is made of transparent PDMS material; the sample loading area is used to add the bronchoalveolar lavage fluid sample to be tested; the mixing reaction area is used to mix the sample with the reinforcing substrate, and a micro-vortex structure is set in the mixing reaction area, and an MXene-Ag composite reinforcing substrate is pre-cured; the Raman detection area is an optically transparent detection window, with a SERS active layer formed by the MXene-Ag composite reinforcing substrate pre-laid for Raman spectroscopy scanning, and a silver-plated aluminum slide with a focusing groove is set below the detection window to achieve focusing of laser and scattered light; the waste liquid recovery area is equipped with a water-absorbing filter membrane to collect waste liquid after detection.

3. The kit for rapid detection of Klebsiella pneumoniae according to claim 1, characterized in that: The microfluidic detection chip relies on capillary force to automatically drive the sample flow.

4. The kit for rapid detection of Klebsiella pneumoniae according to claim 1, characterized in that: The MXene-Ag composite reinforced substrate is Ti3C2T. x The composite colloidal cured layer with uniformly loaded silver nanoparticles on the surface of MXene two-dimensional sheets, and the preparation method of the MXene-Ag composite reinforced substrate are as follows: Ti3C2T... x MXene powder was dispersed in sterile deionized water to prepare a dispersion with a concentration of 0.5-2 mg / mL. The dispersion was sonicated for 30-60 minutes to fully exfoliate the crystal layer. The dispersion was then centrifuged, and the supernatant was collected. The supernatant was placed in a water bath at 30±5℃ under magnetic stirring. Sodium citrate solution and AgNO3 solution were added sequentially, followed by the addition of NaBH4 reducing agent. The mixture was stirred at room temperature for 10-30 minutes. The reaction product was centrifuged and washed to obtain the MXene-Ag composite colloid.

5. A kit for rapid detection of Klebsiella pneumoniae according to claim 1, characterized in that: The sample dilution buffer is used to dilute and adjust the pH of bronchoalveolar lavage fluid samples, maintain the stability of bacterial morphology and biomolecular structure, reduce sample matrix viscosity, and reduce non-specific adsorption; the positive control is used to verify the effectiveness of the kit, and the negative control is used to eliminate contamination interference.

6. A kit for rapid detection of Klebsiella pneumoniae according to claim 5, characterized in that: The sample dilution buffer contained 0.01 mol / L potassium chloride, 0.01 mol / L calcium chloride, and 0.005 mol / L Tris-HCl buffer, with a pH of 7.2-7.4, and was sterilized by filtration through a 0.22 μm filter membrane. The positive control contained a suspension of inactivated Klebsiella pneumoniae standard strain at a concentration of 1 × 10³ CFU / mL, with the above sample dilution buffer as the solvent and 0.01% thimerosal added as a preservative. The negative control used sterile 0.01 mol / L PBS buffer, with a pH of 7.2-7.4, and was sterilized by filtration through a 0.22 μm filter membrane.

7. A method for rapid detection of Klebsiella pneumoniae, comprising: step one, sample pretreatment; step two, automated mixing reaction; step three, Raman spectral signal acquisition; step four, intelligent analysis and interpretation; and step five, result output; characterized in that: In step one above, 50-100 μL of the bronchoalveolar lavage fluid sample to be tested is taken with a pipette and added to the sample processing tube. 40-70 μL of sample dilution buffer is added and mixed well. 50-100 μL of the diluted sample is taken and added to the sample loading area of ​​the microfluidic detection chip. The microfluidic detection chip is then inserted into the detection slot of the portable Raman spectrometer. In step two above, the sample enters the mixing reaction zone under capillary force and comes into full contact with the MXene-Ag composite reinforced substrate, reacting for 5-10 minutes. In step three above, after the reaction in step two is completed, the portable Raman instrument is started, automatically scanning the Raman detection area, collecting Raman spectra, and transmitting the collected spectral data to the smart terminal. In step four above, the smart terminal performs background subtraction and normalization on the received spectral data, and then inputs it into the Attention-CNN model. The Attention-CNN model outputs the detection results, and the determination of the bronchoalveolar lavage fluid sample is completed based on the detection results.

8. A method for rapid detection of Klebsiella pneumoniae according to claim 7, characterized in that: In step three, the acquisition parameters of the portable Raman spectrometer are as follows: excitation wavelength 532nm, laser power adaptively adjusted from 0.1 to 1.0mW, integration time 10s, spectral acquisition range 800-1800cm⁻¹, automatic scanning of 3 sites in each detection window, and taking the average spectrum as the final detection spectrum.

9. A method for rapid detection of Klebsiella pneumoniae according to claim 7, characterized in that: In step four, the Attention-CNN model extracts global and local spectral features through multi-scale convolutional layers, automatically focuses on the characteristic peak regions of Klebsiella pneumoniae through spatial and channel attention modules, amplifies the weights of the characteristic peak regions, suppresses background stray signal interference, and completes automatic spectral classification and recognition.

10. A method for rapid detection of Klebsiella pneumoniae according to claim 7, characterized in that: In step four, the detection results include classification results, classification confidence, and feature importance heatmap.

Images