Microbial drug sensitivity detection method

Abstract

The invention discloses a microbial drug sensitivity detection method, which belongs to the field of microbial detection, and comprises the following steps: acquiring a dropping liquid image, identifying the image, extracting microbial growth characteristics, counting microbial growth phenotypes, obtaining a drug sensitivity detection result and the like, and comparing a confidence interval lower limit a2 of an image P2 with a confidence interval b1 of an image P1 to obtain the drug sensitivity detection result. The sensitivity of the microbial medicine is rapidly and accurately detected without marks, so that the problem of low detection speed of the sensitivity of the microbial medicine in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of microbial detection, and in particular to a method for detecting microbial drug susceptibility. Background Technology

[0002] The global spread of antibiotic resistance (AMR) and the resulting infections have become a major threat to human health. Rapid and reliable identification of antibiotic resistance is crucial for the proper treatment of bacterial infections. Currently, in clinical practice, pathogen identification and antibiotic susceptibility testing (AST) still rely on time-consuming culture-based techniques. This method involves placing a patient's sample in a nutrient-rich culture medium, utilizing the nutrients to promote bacterial growth, and finally observing the presence of bacterial colonies in the petri dish to determine the type and quantity of infecting bacteria. For positive samples identified as bacterial infections, further susceptibility testing with different antibiotics is required to determine the optimal treatment. However, because the culture-based testing process is time-consuming—sample culture typically takes 24 to 48 hours or even longer to produce results—it can delay treatment and is unsuitable for emergency situations or situations requiring frequent testing. Therefore, there is an urgent need for techniques that can rapidly determine the antibiotic susceptibility of microorganisms in the shortest possible time.

[0003] Droplet microfluidics is a high-throughput, miniaturized analytical technique that shows great potential in advancing bacterial identification and drug susceptibility testing. However, current droplet microfluidics detection methods mostly employ flow cytometry or low-magnification microscopy, which suffer from slow drug susceptibility testing speeds. Summary of the Invention

[0004] In order to overcome the shortcomings of the prior art, one of the objectives of this invention is to provide a method for detecting microbial drug susceptibility with fast detection speed.

[0005] One of the objectives of this invention is achieved through the following technical solution:

[0006] A method for detecting microbial drug susceptibility includes the following steps:

[0007] Droplet image acquisition: Prepare multiple microbial droplets containing the drug and multiple microbial droplets without the drug, and acquire image P1 of the multiple microbial droplets containing the drug and image P2 of the multiple microbial droplets without the drug.

[0008] Image recognition: Identify each image P1 and each image P2, and identify the droplets and microorganisms in each image;

[0009] Microbial growth feature extraction: Based on the images of microorganisms inside the labeled droplets, analyze the growth-related features of the microorganisms;

[0010] Microbial growth phenotypic statistics: The mean value of the statistical measure L of microbial growth characteristics in multiple droplets in images P1 and P2. Standard deviation S(L), based on the mean The growth rate confidence interval [a,b] is calculated using the standard deviation S(L), where a and b are the lower and upper limits of the confidence interval, respectively.

[0011] Drug sensitivity test results: When the lower limit of the confidence interval a2 of multiple images P2 is higher than the confidence interval b1 of multiple images P1, it is determined that the microorganism is sensitive to the added drug; when the lower limit of the confidence interval a2 of multiple images P2 is not higher than the confidence interval b1 of multiple images P1, the detection time is extended and a2 and b1 are compared again until the detection time is greater than the longest detection time. When the detection time is greater than the longest detection time, a2 is not higher than b1, and it is determined that the microorganism is not sensitive to the added drug.

[0012] Furthermore, in the microbial growth feature extraction step, the microbial growth-related features include any one or more of the following: the ratio of the pixel area of ​​the microbial image to the droplet image, the number of microorganisms in the droplet image, and the proportion of droplets with a certain growth characteristic among all droplets.

[0013] Furthermore, the proportion of droplets with a certain growth characteristic in all droplets is specifically defined as: the ratio of droplets containing one or more microorganisms to all droplets containing microorganisms, or the ratio of the area of ​​microorganism pixels in a droplet exceeding one or more microorganism pixels to all droplets or all droplets containing microorganisms.

[0014] Furthermore, in the droplet image acquisition step, the total time for acquiring the image is the detection time. The shortest detection time is obtained by plotting the upper confidence limit curve of the growth rate of the sample containing the drug and the lower confidence limit curve of the growth rate of the sample without the drug. The time at the intersection of the upper limit curve and the lower limit curve is the shortest detection time.

[0015] Furthermore, in the drug sensitivity detection result step, the longest detection time is 4h-5h.

[0016] Furthermore, in the drug sensitivity detection result step, after extending the detection time, droplet images are acquired at fixed intervals, with an interval of 5 seconds to 30 minutes.

[0017] Furthermore, in the droplet image acquisition step, the droplet is prepared using a droplet generation chip.

[0018] Furthermore, in the image recognition step, identifying droplets in each image specifically involves using a computer vision-based target detection method to remove invalid droplets that are blurred or overlap.

[0019] Furthermore, in the image recognition step, the computer vision-based target detection method is YOLO target detection network deep learning technology, DETR target detection network deep learning technology, or Hough circle detection technology.

[0020] Compared with existing technologies, the microbial drug susceptibility detection method of the present invention, through steps such as droplet image acquisition, image recognition, microbial growth feature extraction, microbial growth phenotype statistics, and drug susceptibility detection results, compares the lower limit of the confidence interval a2 of image P2 with the confidence interval b1 of image P1, which can quickly detect the sensitivity of microorganisms to added drugs, and the detection speed is fast. Attached Figure Description

[0021] Figure 1 This is a flowchart of the microbial drug susceptibility detection method of the present invention;

[0022] Figure 2 This is a schematic diagram of droplet identification in the microbial drug susceptibility detection method of the present invention;

[0023] Figure 3 This is a schematic diagram of microbial identification in the microbial drug susceptibility detection method of the present invention;

[0024] Figure 4 This is a schematic diagram of the microbial pixel size in the microbial drug susceptibility detection method of the present invention. Detailed Implementation

[0025] 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.

[0026] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0027] Please see Figure 1 The present invention discloses a method for detecting microbial drug susceptibility, comprising the following steps:

[0028] A method for detecting microbial drug susceptibility includes the following steps:

[0029] Droplet image acquisition: Prepare multiple microbial droplets containing the drug and multiple microbial droplets without the drug, and acquire image P1 of the multiple microbial droplets containing the drug and image P2 of the multiple microbial droplets without the drug.

[0030] Image recognition: Identify each image P1 and each image P2, and identify the droplets and microorganisms in each image;

[0031] Microbial growth feature extraction: Based on the images of microorganisms inside the labeled droplets, analyze the growth-related features of the microorganisms;

[0032] Microbial growth phenotypic statistics: The mean value of the statistical measure L of microbial growth characteristics in multiple droplets in images P1 and P2 at multiple time points was calculated. Standard deviation S(L), based on the mean The growth rate confidence interval [a,b] is calculated using the standard deviation S(L), where a and b are the lower and upper limits of the confidence interval, respectively.

[0033] Drug sensitivity test results: When the lower limit of the confidence interval a2 of multiple images P2 is higher than the confidence interval b1 of multiple images P1, it is determined that the microorganism is sensitive to the added drug; when the lower limit of the confidence interval a2 of multiple images P2 is not higher than the confidence interval b1 of multiple images P1, the detection time is extended and a2 and b1 are compared again until the detection time is greater than the longest detection time. When the detection time is greater than the longest detection time, a2 is not higher than b1, and it is determined that the microorganism is not sensitive to the added drug.

[0034] The specific steps for acquiring droplet images are as follows:

[0035] The droplet generation chip prepares droplets by adding a microbial solution to be detected to the sample input port of the chip, and adding culture media containing and without drugs to the culture media input port of the chip, thereby generating multiple microbial droplets containing the drug and multiple microbial droplets without the drug. In this application, the drug is an antibiotic. Images are shown below. Figure 2As shown, the image contains multiple droplets. The image capture time is the detection time. The shortest detection time is obtained by plotting the upper confidence limit curve of the growth rate of the drug-containing sample and the lower confidence limit curve of the growth rate of the drug-free sample. The time of the intersection of the upper and lower limit curves is the shortest detection time. The maximum detection time is 4-5 hours. When a microorganism is detected to be sensitive to the sample, the detection result is obtained and the detection is stopped. When no microorganism is detected to be sensitive to the sample, if the microorganism is sensitive to the sample, the drug will affect the growth of the microorganism, causing changes in the sample's growth rate. Different types of microorganisms have different sensitivities to different drugs, and the effects of some drugs on microbial growth may not be immediately apparent. Therefore, it is necessary to extend the detection time and continue detection. During the continued detection, images are captured at fixed intervals, with an interval of 5 seconds to 30 minutes. The images are analyzed to determine drug sensitivity. This process continues until the maximum detection time is reached. If no microorganism is detected to be sensitive to the sample after the maximum detection time, the microorganism is considered to be insensitive to the sample (drug resistance).

[0036] The specific steps for image recognition acquisition are as follows:

[0037] Please continue reading. Figure 2 To identify droplets in each image, a computer vision-based object detection method is used to remove invalid droplets that are blurred or overlap. This method includes YOLO object detection network deep learning, DETR object detection network deep learning, or Hough circle detection. Please continue reading. Figure 3 To identify microorganisms in the droplets in each image, semantic segmentation networks such as UNet and CellPose, or object detection networks such as YOLO series and DETR, are used to identify the microorganisms in the droplets.

[0038] The specific steps for extracting microbial growth characteristics are as follows:

[0039] Based on the extracted images of microorganisms within the droplets, morphological analysis is used to identify growth-related features such as the number of microorganisms within the droplets and the pixel area occupied by the microorganism images. Specifically, the growth-related features of microorganisms include any one or more of the following: the ratio of the pixel area of ​​the microorganism image to that of the droplet image, the number of microorganisms in the droplet image, and the proportion of droplets with a certain growth characteristic among all droplets. The proportion of droplets with a certain growth characteristic among all droplets specifically refers to: the ratio of droplets containing one or more microorganisms to all droplets containing microorganisms, or the ratio of the pixel area of ​​more than one microorganism in a droplet to the pixel area of ​​all droplets or all droplets containing microorganisms. In this embodiment, the extracted growth feature is the ratio of the pixel area of ​​the microorganism image to that of the droplet image.

[0040] The specific steps for statistical analysis of microbial growth phenotypes are as follows:

[0041] Mean value of the phenotypic statistic L of microbial growth characteristics in droplets The standard deviation S(L) and 99% confidence interval [a, b] are used to provide an overall assessment of the microbial growth status in the sample. L is the growth phenotypic characteristic at a certain time point, such as the ratio of the pixel area of ​​the microbial image to the droplet image in all or part of the microbial-containing droplets. a and b are the lower and upper limits of the 99% confidence interval, respectively.

[0042] Specifically, the mean of the statistic L1 corresponding to image P1 of the microbial droplet containing the drug. Standard deviation S(L1) and 99% confidence interval [a1, b1]. Mean of statistic L2 corresponding to image P2 of drug-free microbial droplets. Standard deviation S(L2) and 99% confidence interval [a2, b2].

[0043] The following uses Escherichia coli standards and clinical urine samples as examples to illustrate the use of the microbial drug susceptibility testing method in this application.

[0044] Sensitivity test of Escherichia coli to the antibiotic ciprofloxacin:

[0045] (1) Take samples and dilute the original bacterial culture to 5×10⁻⁶ using sterile liquid culture medium such as Luria-Bertani medium or Müller-Hinton medium. 5 CFU / mL.

[0046] (2) Add 10 μL of Escherichia coli culture to the sample input port of the droplet generation chip, and add culture medium containing the antibiotic ciprofloxacin and culture medium without the antibiotic to the culture medium input port to generate microdroplets.

[0047] (3) Transfer the microdroplets to the detection chamber and place them on the microscope stage for imaging. If necessary, place the detection chamber in a constant temperature incubator for incubation. The imaging parameters vary depending on the growth rate. The magnification of the imaging system is 10X or 20X, the imaging interval is 5s-30min, and the total duration is 30min-4h. Depending on performance requirements, other volumes of bacterial solution can also be used, such as (1-1000uL).

[0048] (4) Based on the captured droplet images, microbial growth characteristics are extracted, and microbial growth phenotypes are statistically analyzed to determine the drug sensitivity of *E. coli* to the antibiotic ciprofloxacin. For example... Figure 4As shown, the horizontal axis represents time, and the vertical axis represents the number of pixels in the microbial image. The blue lines in the figure represent the image pixels of E. coli without the antibiotic ciprofloxacin; the orange lines in the figure represent the image pixels of E. coli with the antibiotic ciprofloxacin. The E. coli strain ATCC 25922 can be observed to have a difference in the mean of growth phenotypic statistics 50 minutes after droplet generation, and the strain can be determined to be sensitive to antibiotics within 90 minutes.

[0049] Sensitivity testing of clinical urine samples for the antibiotic ciprofloxacin:

[0050] (1) Use a filter with a membrane pore size of 5 μm or 2.5 μm to remove a large number of non-bacterial impurities from the urine sample. If necessary, bacterial enrichment can also be performed to obtain pretreated urine. The pretreated urine includes Escherichia coli, Proteus mirabilis, Klebsiella pneumoniae, Streptococcus saprophyticus, Enterobacter faecalis, etc.

[0051] (2) Add 10 μL of pretreated urine to the sample input port of the droplet generation chip, and add culture medium containing antibiotics and culture medium without antibiotics to the culture medium input port to generate microdroplets. The antibiotics are any one of amoxicillin, ciprofloxacin, fosfomycin, nitrofurantoin, and trimethoprim.

[0052] (3) Transfer the microdroplets to the detection chamber and place them on the microscope stage for imaging. If necessary, place the detection chamber in a constant temperature incubator for incubation. Imaging parameters may vary depending on the growth rate. The imaging system magnification is 10X or 20X, the imaging interval is 5s-30min, and the total duration is 30min-4h. Depending on performance requirements, other volumes of bacterial culture can also be used, such as (1-1000uL).

[0053] (4) Based on the captured droplet images, microbial growth characteristics are extracted, microbial growth phenotypes are statistically analyzed, and the drug sensitivity of microorganisms in urine to antibiotics is determined.

[0054] Compared with existing technologies, the microbial drug susceptibility detection method of the present invention, through steps such as droplet image acquisition, image recognition, microbial growth feature extraction, microbial growth phenotype statistics, and drug susceptibility detection results, compares the lower limit of the confidence interval a2 of image P2 with the confidence interval b1 of image P1 to perform label-free, rapid, and accurate detection of microbial drug susceptibility, thereby solving the problem of slow detection speed of microbial drug susceptibility in existing technologies.

[0055] The above embodiments merely illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the invention patent. It should be noted that, for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present invention. These are all equivalent modifications and improvements made to the above embodiments based on the essential technology of the present invention, and all of these fall within the protection scope of the present invention.

Claims

1. A method for detecting microbial drug susceptibility, characterized in that, Includes the following steps: Droplet image acquisition: Prepare multiple microbial droplets containing the drug and multiple microbial droplets without the drug, and acquire image P1 of the multiple microbial droplets containing the drug and image P2 of the multiple microbial droplets without the drug. Image recognition: Identify each image P1 and each image P2, and identify the droplets and microorganisms in each image; Microbial growth feature extraction: Based on the images of microorganisms inside the labeled droplets, analyze the growth-related features of the microorganisms; Microbial growth phenotypic statistics: The mean value of the statistical measure L of microbial growth characteristics in multiple droplets in images P1 and P2. Standard deviation S(L), based on the mean The growth rate confidence interval [a,b] is calculated using the standard deviation S(L), where a and b are the lower and upper limits of the confidence interval, respectively. Drug sensitivity test results: When the lower limit of the confidence interval a2 of multiple images P2 is higher than the confidence interval b1 of multiple images P1, it is determined that the microorganism is sensitive to the added drug; when the lower limit of the confidence interval a2 of multiple images P2 is not higher than the confidence interval b1 of multiple images P1, the detection time is extended and a2 and b1 are compared again until the detection time is greater than the longest detection time. When the detection time is greater than the longest detection time, a2 is not higher than b1, and it is determined that the microorganism is not sensitive to the added drug.

2. The method for detecting microbial drug susceptibility according to claim 1, characterized in that: In the microbial growth feature extraction step, the microbial growth-related features include any one or more of the following: the ratio of the pixel area of ​​the microbial image to the droplet image, the number of microorganisms in the droplet image, and the proportion of droplets with a certain growth characteristic among all droplets.

3. The method for detecting microbial drug susceptibility according to claim 2, characterized in that: The specific percentage of droplets with a certain growth characteristic in all droplets is: the ratio of droplets containing one or more microorganisms to all droplets containing microorganisms, or the ratio of the area of ​​microbial pixels in droplets exceeding one or more microorganism pixels to all droplets or all droplets containing microorganisms.

4. The method for detecting microbial drug susceptibility according to claim 1, characterized in that: In the droplet image acquisition step, the total time for acquiring the image is the detection time. The shortest detection time is obtained by plotting the upper confidence limit curve of the growth rate of the sample containing the drug and the lower confidence limit curve of the growth rate of the sample without the drug. The time at the intersection of the upper limit curve and the lower limit curve is the shortest detection time.

5. The method for detecting microbial drug susceptibility according to claim 1, characterized in that: In the drug sensitivity detection result step, the longest detection time is 4h-5h.

6. The method for detecting microbial drug susceptibility according to claim 1, characterized in that: In the drug sensitivity detection result step, after extending the detection time, droplet images are acquired at fixed intervals, with an image capture interval of 5s-30min.

7. The method for detecting microbial drug susceptibility according to claim 1, characterized in that: In the droplet image acquisition step, droplets are prepared using a droplet generation chip.

8. The method for detecting microbial drug susceptibility according to claim 1, characterized in that: In the image recognition step, identifying droplets in each image specifically involves using a computer vision-based target detection method to remove invalid droplets that are blurred or overlap.

9. The method for detecting microbial drug susceptibility according to claim 8, characterized in that: In the image recognition step, the computer vision-based target detection method is YOLO target detection network deep learning technology, DETR target detection network deep learning technology, or Hough circle detection technology.

Images