Electronic device and method for performing antibiotic susceptibility test of microorganism

Abstract

An electronic device for performing an antibiotic susceptibility test of a microorganism, according to one embodiment of the present invention, may comprise a processor, which: obtains antibiotic data obtained by measuring the change in electrical characteristics according to antibiotic treatment of a target microorganism; inputs, into a deep learning model trained to determine whether the antibiotic data is positive or negative by using positive control group data, negative control group data and antibiotic data of the microorganism, the positive control group data, the negative control group data and the antibiotic data of the target microorganism, thereby determining whether the antibiotic data of the target microorganism is positive or negative; and uses the determination result so as to identify antibiotic susceptibility information about the target microorganism according to antibiotic type and concentration.

Description

Electronic device and method for performing antibiotic susceptibility testing of microorganisms

[0001] The present invention relates to an electronic device and method for performing antibiotic susceptibility testing of microorganisms.

[0002] Among the various diseases that affect the human body, there are those caused by microbial infections, including urinary tract infections and sepsis. Treating these diseases requires antibiotics, which inhibit microbial growth or kill them. However, over a hundred different antibiotics have been developed to date, and their effectiveness varies depending on the type of microorganism and its antibiotic resistance. Therefore, the most effective antibiotic must be selected for each patient, and the diagnostic testing method for this purpose is called antibiotic susceptibility testing.

[0003] One antibiotic susceptibility test utilizes changes in electrical characteristics based on microbial growth. Data representing these changes, such as electrical capacitance graphs, are analyzed using principal component analysis (PCA) to quantify similarity with the control group.

[0004] However, many errors occurred when data values ​​differed between the time-dependent capacitance graph of the control group and the time-dependent capacitance graph of antibiotic-treated microorganisms in the differential graph obtained during the PCA process, resulting in judgments that differed from the actual results. For example, there may be differences in the time at which the peak capacitance appears or in the peak value between the time-dependent capacitance graph of the control group and the time-dependent capacitance graph of antibiotic-treated microorganisms.

[0005] Therefore, there is a need to find a new analysis method that reduces judgment errors while using data measuring electrical characteristic changes.

[0006] The purpose of the present invention is to provide an electronic device and method for determining whether a microorganism is susceptible to antibiotics, which can significantly reduce determination errors.

[0007] In an electronic device for performing an antibiotic susceptibility test of a microorganism according to one embodiment of the present invention, a processor may be included that obtains antibiotic data by measuring changes in electrical characteristics according to antibiotic treatment of a target microorganism, inputs positive control data, negative control data, and antibiotic data of the target microorganism into a deep learning model trained to determine positivity or negativity of the antibiotic data using the positive control data, negative control data, and antibiotic data of the microorganism, and determines positivity or negativity of the antibiotic data of the target microorganism, and identifies antibiotic susceptibility information of the target microorganism according to the type and concentration of the antibiotic using the determination result.

[0008] The above deep learning model can be trained to extract feature vectors of positive control data, negative control data, and antibiotic data of microorganisms, respectively, and to make the feature vector of the antibiotic data similar to the feature vector of the negative control data or the positive control data based on the labeling of the positive or negative of the antibiotic data.

[0009] The processor extracts feature vectors of positive control data, negative control data, and antibiotic data of the target microorganism, respectively, and measures the similarity between the feature vectors of the antibiotic data and the feature vectors of the positive control data and the negative control data using the deep learning model, and determines whether the antibiotic data of the target microorganism is positive or negative based on the similarity.

[0010] The above processor can obtain the antibiotic data through a biosensor including a plurality of cross-electrodes implemented to measure in real time changes in electrical characteristics according to whether microorganisms proliferate or not according to antibiotic treatment.

[0011] The above processor can calculate the minimum inhibitory concentration (MIC) for each antibiotic type of the target microorganism by matching the antibiotic type and concentration information corresponding to each of the plurality of cross electrodes to the determination result.

[0012] The above processor can obtain the antibiotic susceptibility information by determining one of the susceptibility, intermediate resistance, and resistance of the target microorganism to each type of antibiotic based on the minimum inhibitory concentration.

[0013] In a method for testing antibiotic susceptibility of a microorganism performed by an electronic device according to one embodiment of the present invention, the method may include: a step of obtaining antibiotic data by measuring changes in electrical characteristics according to antibiotic treatment of a target microorganism; a step of inputting positive control data, negative control data, and antibiotic data of the target microorganism into a deep learning model trained to determine positivity or negativity of the antibiotic data of the microorganism using the positive control data, negative control data, and antibiotic data of the microorganism, thereby determining positivity or negativity of the antibiotic data of the target microorganism; and a step of identifying antibiotic susceptibility information of the target microorganism according to the type and concentration of the antibiotic using the determination result.

[0014] According to one embodiment of the present invention, the sensitivity can be accurately determined according to the antibiotic concentration, thereby dramatically increasing the accuracy of the sensitivity test for each type of antibiotic.

[0015] According to one embodiment of the present invention, by training a deep learning model using a feature vector of data, it is possible to accurately determine whether it is positive or negative without being affected by differences in time-series electrical characteristic changes between antibiotic data and control data.

[0016] According to one embodiment of the present invention, by using a deep learning model trained to make feature vectors similar between data, it is possible to more accurately determine whether data whose time-series electrical characteristic changes are unclear are positive or negative.

[0017] FIG. 1 is a schematic diagram illustrating an antibiotic susceptibility testing system according to one embodiment of the present invention.

[0018] FIG. 2 is a block diagram illustrating the configuration of an electronic device according to one embodiment of the present invention.

[0019] FIG. 3 is a diagram illustrating an operation flow chart of an electronic device according to one embodiment of the present invention.

[0020] FIG. 4 is a drawing illustrating a biosensor according to one embodiment of the present invention.

[0021] FIG. 5 is a diagram illustrating a learning process of a deep learning model according to one embodiment of the present invention.

[0022] FIG. 6 is a drawing showing an operation flow chart of an electronic device according to a first embodiment of the present invention.

[0023] FIG. 7 is a diagram illustrating an operation flow chart of an electronic device according to a second embodiment of the present invention.

[0024] Figure 8 is a diagram illustrating performance criteria for an antibiotic susceptibility test according to one embodiment of the present invention.

[0025] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. The detailed description set forth below, together with the accompanying drawings, is intended to explain exemplary embodiments of the present invention and is not intended to represent the only embodiments in which the present invention may be practiced. In the drawings, portions irrelevant to the description may be omitted for clarity in describing the present invention, and the same reference numerals may be used throughout the specification for identical or similar components.

[0026] FIG. 1 is a schematic diagram illustrating an antibiotic susceptibility testing system according to one embodiment of the present invention.

[0027] An antibiotic susceptibility test system (1) (hereinafter, also referred to as system (1)) according to one embodiment of the present invention may include a biosensor (10) and an electronic device (100).

[0028] A biosensor (10) according to one embodiment of the present invention is a sensor that measures changes in electrical characteristics and may be comprised of a plurality of interdigitated electrodes (IDEs). The biosensor (10) may be implemented to measure changes in electrical characteristics in real time according to whether or not microorganisms proliferate by treating microorganisms with antibiotics at each interdigitated electrode.

[0029] At this time, the electrical characteristics may be capacitance, impedance, resistance, reactance, etc. Therefore, when measuring changes in electrical characteristics, there is an advantage in that sensitivity can be increased even with a lower concentration and a smaller amount of sample compared to when measuring changes in optical characteristics. The biosensor (10) can measure according to the type and concentration of antibiotics treated to the target microorganism, and an example of the biosensor (10) is illustrated in FIG. 4.

[0030] An electronic device (100) according to one embodiment of the present invention is a device for performing antibiotic susceptibility testing of microorganisms, and can be implemented as a computer, server, smart phone, tablet PC, smart pad, laptop, etc.

[0031] The electronic device (100) can use antibiotic data received from the biosensor (10) to calculate the minimum inhibitory concentration (MIC) for each type of antibiotic for the target microorganism and perform a sensitivity test. The minimum inhibitory concentration refers to the lowest concentration of the antibiotic required to inhibit visible growth of the microorganism.

[0032] Antibiotic susceptibility test results can be divided into, for example, Resistance (R), Intermediate Resistance (I), and Susceptibility (S). For example, to treat an infection caused by a microorganism, an antibiotic to which the microorganism is susceptible must be used. A microorganism identified as susceptible can be treated with antibiotics prescribed at the recommended dosage for the microorganism and the site of infection. Furthermore, a microorganism identified as intermediate resistant may have a lower therapeutic effect because the minimum inhibitory concentration of the antibiotic for the target microorganism is similar to the maximum concentration of the drug that can be prescribed. Meanwhile, a microorganism identified as resistant cannot be treated with the maximum concentration of the drug that can be prescribed.

[0033] As described above, when conducting a susceptibility test by analyzing antibiotic data using principal components, there were cases where the data values ​​differed from the control group despite the graph pattern being similar, and cases where the results were judged differently from the actual results.

[0034] In the present invention, a method for performing susceptibility testing through pattern analysis of time series of antibiotic data using a deep learning model is proposed.

[0035] Hereinafter, the configuration and operation of an electronic device according to one embodiment of the present invention will be specifically described with reference to the drawings.

[0036] FIG. 2 is a block diagram illustrating the configuration of an electronic device according to one embodiment of the present invention.

[0037] An electronic device (100) according to one embodiment of the present invention may include an input unit (110), a communication unit (120), a display unit (130), a storage unit (140), and a processor (150).

[0038] The input unit (110) generates input data in response to user input of the electronic device (100). For example, the user input may be a user input for starting the operation of the electronic device (100), a user input for acquiring antibiotic data, a user input for labeling antibiotic data for learning, a user input for entering the type and concentration of antibiotics to be processed for each cross-electrode of the biosensor (10), etc. In addition, if it is a user input required in the process of performing deep learning model learning, antibiotic susceptibility testing, etc., it may be applied without limitation.

[0039] The input unit (110) includes at least one input means. The input unit (110) may include a keyboard, a key pad, a dome switch, a touch panel, a touch key, a mouse, a menu button, etc.

[0040] The communication unit (120) can perform communication with external devices such as a biosensor (10) and a server to transmit and receive antibiotic data, positive control data, negative control data, antibiotic type and concentration information, antibiotic susceptibility information, deep learning models, etc.

[0041] To this end, the communication unit (120) can perform wireless communication such as 5G (5th generation communication), LTE-A (long term evolution-advanced), LTE (long term evolution), Wi-Fi (wireless fidelity), Bluetooth, or wired communication such as LAN (local area network), WAN (Wide Area Network), and power line communication.

[0042] The display unit (130) displays display data according to the operation of the electronic device (100). The display unit (130) may display a screen displaying at least one of antibiotic data, positive control data, and negative control data, a screen displaying antibiotic type and concentration information, a screen displaying antibiotic susceptibility information, a screen receiving user input, etc.

[0043] The display unit (130) includes a liquid crystal display (LCD), a light emitting diode (LED) display, an organic light emitting diode (OLED) display, a micro electro mechanical systems (MEMS) display, and an electronic paper display. The display unit (130) may be implemented as a touch screen by being combined with the input unit (110).

[0044] The storage unit (140) can store operation programs of the electronic device (100). The storage unit (140) can include non-volatile storage that can preserve data (information) regardless of whether power is supplied, and volatile memory that cannot preserve data when power is not supplied and into which data to be processed by the processor (150) is loaded. The storage includes flash memory, hard-disk drive (HDD), solid-state drive (SSD), read-only memory (ROM), etc., and the memory includes buffer, random access memory (RAM), etc.

[0045] The storage unit (140) can store antibiotic data, positive control data, negative control data, antibiotic type and concentration information, antibiotic susceptibility information, deep learning models, etc. The storage unit (140) can store operation programs, etc. required in the process of acquiring data, determining positive / negative antibiotic data of target microorganisms, identifying antibiotic susceptibility information, extracting feature vectors from time series data, measuring similarity between feature vectors, etc.

[0046] The processor (150) can control at least one other component (e.g., hardware or software component) of the electronic device (100) by executing software such as a program, and can perform various data processing or operations.

[0047] A processor (150) according to one embodiment of the present invention obtains antibiotic data by measuring changes in electrical characteristics according to antibiotic treatment of a target microorganism, inputs positive control data, negative control data, and antibiotic data of the target microorganism into a deep learning model trained to determine positivity or negativity of the antibiotic data using positive control data, negative control data, and antibiotic data of the microorganism, thereby determining positivity or negativity of the antibiotic data of the target microorganism, and using the determination result, identifies antibiotic susceptibility information of the target microorganism according to the type and concentration of antibiotic.

[0048] A processor (150) according to one embodiment of the present invention learns a deep learning model that determines whether antibiotic data is positive or negative using positive control data, negative control data, and antibiotic data of microorganisms, or receives and stores a deep learning model generated through prior learning from the outside and uses it, but is not limited to any one of them.

[0049] Meanwhile, the processor (150) may perform at least a portion of the data analysis, processing, and result information generation for performing the above operations using at least one of a machine learning, neural network, or deep learning algorithm as a rule-based or artificial intelligence algorithm. Examples of the neural network may include models such as a CNN (Convolutional Neural Network), a DNN (Deep Neural Network), an RNN (Recurrent Neural Network), and a Vision Transformer.

[0050] FIG. 3 is a diagram illustrating an operation flow chart of an electronic device according to one embodiment of the present invention.

[0051] According to one embodiment of the present invention, the processor (150) can obtain antibiotic data measuring changes in electrical characteristics according to antibiotic treatment of a target microorganism (S10).

[0052] The target microorganism may be a bacterium, and the bacterium is preferably, but not limited to, a gram-positive bacterium, a gram-negative bacterium, and an antibiotic-resistant strain thereof. Specifically, the gram-positive bacterium may be at least one selected from the group consisting of Bacillus subtilis, Staphylococcus aureus, Enterococcus faecalis, and Staphylococcus epidermidis, and the gram-negative bacterium may be at least one selected from the group consisting of Escherichia coli, Psedomonas aeruginosa, Acinetobacter baumannii, and Salmonella typhimurium.

[0053] The above antibiotics are not specifically limited in type, and any antibiotic whose sensitivity can be measured by the biosensor (10) according to the present invention is included. Specifically, the antibiotics include Ampicillin, Tetracycline, Gentamicin, Erythromycin, Vancomycin, Linezolid, Methicillin, Oxacillin, Cefotaxime, Rifampicin, Amikacin, Amikacin, Kanamycin, Tobramycin, Neomycin, Ertapenem, Doripenem, Imipenem / Cilastatin, Meropenem, Ceftazidime, Cefepime, It may be selected from the group consisting of Ceftaroline, Ceftobiprole, Aztreonam, Piperacillin, Polymyxin B, Colistin, Ciprofloxacin, Levofloxacin, Moxifloxacin, Gatifloxacin, Tigecycline, combinations thereof and derivatives thereof, preferably Ampicillin or Tetracycline, but not limited thereto.

[0054] Antibiotic data refers to data measuring changes in electrical characteristics over time after antibiotic treatment of microorganisms. For example, if a microorganism is resistant to the antibiotic, the antibiotic data will show an increase in electrical capacity over time as the bacteria proliferate. Conversely, if the microorganism is susceptible to the antibiotic, the antibiotic will inhibit or kill the bacteria, resulting in no change in bacterial concentration. Consequently, the antibiotic data will show a constant electrical capacity over time.

[0055] The processor (150) can obtain antibiotic data from the biosensor (10), but is not limited thereto, and the acquisition path is not limited to any one, such as receiving antibiotic data from an external device that receives the data through the biosensor (10).

[0056] According to one embodiment of the present invention, the processor (150) inputs positive control data, negative control data, and antibiotic data of a target microorganism into a deep learning model trained to determine whether the antibiotic data is positive or negative using the positive control data, negative control data, and antibiotic data of the microorganism, thereby determining whether the antibiotic data of the target microorganism is positive or negative (S20).

[0057] Positive control data refers to data measuring changes in electrical characteristics over time by growing only microorganisms without antibiotics, and negative control data refers to data measuring changes in electrical characteristics over time in a culture medium (a liquid or solid material used for growing, preserving, etc. microorganisms) without microorganisms.

[0058] Qualitatively, if the antibiotic data is similar to the positive control data, the microorganism is likely to be resistant to the antibiotic, and if it is similar to the negative control data, the microorganism is likely to be susceptible to the antibiotic.

[0059] The present invention utilizes a deep learning model trained to determine whether antibiotic data is positive or negative by leveraging the time-series characteristics of these data. This allows for quantitative analysis of whether antibiotic data is similar to positive or negative control data. If the antibiotic data is similar to the positive control data, it is determined as positive (P), and if it is similar to the negative control data, it is determined as negative (N).

[0060] The specific learning process of the deep learning model is explained with reference to Fig. 5, and the judgment process using the deep learning model is explained with reference to Fig. 6.

[0061] According to one embodiment of the present invention, the processor (150) can identify antibiotic susceptibility information of a target microorganism according to the type and concentration of antibiotic using the judgment result (S30).

[0062] The processor (150) can obtain a judgment result that determines whether the antibiotic data of the target microorganism is positive or negative. At this time, the judgment result is obtained in response to the case where a specific antibiotic is treated at a specific concentration to the target microorganism, and the processor (150) can obtain a judgment result according to the antibiotic concentration for each type of antibiotic. The processor (150) can synthesize the judgment results to calculate the minimum inhibitory concentration for each type of antibiotic for the target microorganism.

[0063] The processor (150) can obtain antibiotic susceptibility information by determining one of susceptibility, intermediate resistance, and resistance for each antibiotic type of the target microorganism based on the minimum inhibitory concentration. At this time, the processor (150) can determine one of susceptibility, intermediate resistance, and resistance for each antibiotic type by applying the antibiotic susceptibility test performance standards (e.g., CLSI guidelines, etc.) of the target microorganism.

[0064] According to one embodiment of the present invention, the sensitivity can be accurately determined according to the antibiotic concentration, thereby dramatically increasing the accuracy of the sensitivity test for each type of antibiotic.

[0065] FIG. 4 is a drawing illustrating a biosensor according to one embodiment of the present invention.

[0066] A biosensor (10) according to one embodiment of the present invention is provided with a plurality of cross electrodes (11), and may further include a storage compartment (not shown) capable of storing the cross electrodes (11), antibiotics, and microorganisms therein. Specifically, when measuring the antibiotic susceptibility of microorganisms using the biosensor (10), microorganisms or antibiotics, etc., may be processed inside the storage compartment. The storage compartment may also be of an open top type.

[0067] FIG. 5 is a diagram illustrating a learning process of a deep learning model according to one embodiment of the present invention.

[0068] As previously described with reference to S20 of FIG. 3, a deep learning model (500) can be used to determine whether antibiotic data of a target microorganism is positive or negative. The following description assumes that the processor (150) trains the deep learning model (500).

[0069] The deep learning model (500) can utilize frameworks such as, for example, Support Vector Machine (SVM), Long Short-term Model (LSTM), Transformer, and Recurrent Neural Network (RNN), but is not limited to any one of them.

[0070] First, the processor (150) can extract feature vectors of positive control data (510), negative control data (520), and antibiotic data (530) prepared for learning (S210).

[0071] At this time, the positive control data (510), negative control data (520), and antibiotic data (530) are configured as a single set prepared under identical conditions (same microorganism, same culture medium, same processing method, etc.). In addition, the antibiotic data (530) is prepared with a correct answer as to whether it is positive or negative, i.e., labeled (540).

[0072] The processor (150) can train a deep learning model (500) so that the feature vector of the antibiotic data becomes similar to the feature vector of the positive control data (510) or the feature vector of the negative control data (520) based on the labeling (540) of the positive or negative antibiotic data (S220).

[0073] For example, for antibiotic data (530) labeled as positive, the processor (150) can train the deep learning model (500) so that the feature vector of the antibiotic data (530) becomes similar to the feature vector of the corresponding positive control data (510).

[0074] Conversely, for negatively labeled antibiotic data (530), the processor (150) can train the deep learning model (500) so that the feature vector of the antibiotic data (530) becomes similar to the feature vector of the corresponding negative control data (520).

[0075] According to one embodiment of the present invention, by training a deep learning model (500) using the feature vector of data, it is possible to accurately determine whether it is positive or negative without being affected by the difference in time-series electrical characteristic changes between antibiotic data and control data.

[0076] FIG. 6 is a drawing showing an operation flow chart of an electronic device according to a first embodiment of the present invention.

[0077] As previously described with reference to S20 of FIG. 3 and FIG. 5, a process of using a trained deep learning model (500) to determine whether antibiotic data of a target microorganism is positive or negative is described.

[0078] The processor (150) can extract feature vectors of the positive control data, negative control data, and antibiotic data of the target microorganism (S21). Similarly, for the target microorganism, the positive control data, negative control data, and antibiotic data are prepared as a single set under identical conditions (same culture medium, same processing method, etc.).

[0079] The processor (150) can measure the similarity between the feature vector of antibiotic data and the feature vectors of positive and negative control data using a deep learning model (500) (S22). At this time, any algorithm for measuring similarity, such as cosine similarity measurement, can be employed without limitation.

[0080] The processor (150) can determine whether the antibiotic data of the target microorganism is positive or negative based on the similarity (S23).

[0081] For example, if the feature vector of antibiotic data has a high similarity to the feature vector of positive control data, the processor (150) may determine that the target microorganism is resistant to the antibiotic. Conversely, if the feature vector of antibiotic data has a high similarity to the feature vector of negative control data, the processor (150) may determine that the target microorganism is susceptible to the antibiotic.

[0082] According to one embodiment of the present invention, by using a deep learning model (500) trained to make feature vectors similar between data, it is possible to more accurately determine whether data whose time-series electrical characteristic change pattern is unclear is positive or negative.

[0083] FIG. 7 is a diagram illustrating an operation flow chart of an electronic device according to a second embodiment of the present invention.

[0084] As previously described with reference to S30 of FIG. 3, the process of obtaining antibiotic susceptibility information based on the results determined by the deep learning model (500) is described.

[0085] According to one embodiment of the present invention, the processor (150) can calculate the minimum inhibitory concentration for each antibiotic type of the target microorganism by matching the antibiotic type and concentration information corresponding to each of the plurality of cross electrodes to the determination result (S31).

[0086] Each cross electrode of the biosensor (10) corresponds to one channel, and information mapping the type of antibiotic and the concentration of antibiotic to each channel (referred to as antibiotic type and concentration information) can be provided.

[0087] For example, if the target microorganism is E. coli, the judgment result of determining whether it is positive or negative according to each antibiotic data of E. coli using a deep learning model (500) is as shown in Table 1 below. The judgment result is indicated as P if it is positive, and N if it is negative.

[0088] CH001CH002CH003CH004CH005CH006CH007CH008CH009CH010CH011CH012Ampicillin0.25μg / mLAmpicillin0.5μg / mLAmpicillin1μg / mLAmpicillin2μg / mLAmpicillin4μg / mLAmpicillin8μg / mLAmpicillin16 μg / mLAmpicillin32μg / mLErtapenem0.5 μg / mLErtapenem1μg / mLErtapenem2μg / mLErtapenem4μg / mLPPPPPPPPNNNNCH013CH014CH015CH016CH017CH018CH019CH020CH021CH022CH023CH024Meropenem0.25μg / mLMeropenem0.5μg / mLMeropenem1μg / mLMeropenem2μg / mLMeropenem4μg / mLMeropenem8μg / mLMeropenem16μg / mLCefoxitin2μg / mLCefoxitin4μg / mLCefoxitin8μg / mLCefoxitin16μg / mLCefoxitin32μg / mLPPNNNNNNNNNNCH025CH026CH027CH028CH029CH030CH031CH032CH033CH034CH035CH036Cefuroxime 1μg / mLCefuroxime 2μg / mLCefuroxime 4μg / mLCefuroxime 8μg / mLCefuroxime 16μg / mLCefuroxime 32μg / mLAmoxicillin / clavulanate 2 / 1μg / mLAmoxicillin / clavulanate 4 / 2μg / mLAmoxicillin / clavulanate 8 / 4μg / mLAmoxicillin / clavulanate 16 / 8μg / mLAmoxicillin / clavulanate 32 / 16μg / mLAmoxicillin / clavulanate 64 / 32μg / mLPPPPPPPPPPPPCH037CH038CH039CH040CH041CH042CH043CH044CH045CH046CH047CH048Amoxicillin2μg / mLAmoxicillin4μg / mLAmoxicillin8μg / mLPenicillin 0.12μg / mLPenicillin 0.25μg / mLPenicillin 0.5μg / mLPenicillin 1μg / mLPenicillin 2μg / mLPenicillin 4μg / mLPenicillin 8μg / mLPenicillin 16μg / mLCefepime 0.5μg / mLPPPPPPPPPPPPCH049CH050CH051CH052CH053CH054CH055CH056CH057CH058CH059CH060Cefepime 1μg / mLCefepime 2μg / mLCefepime 4μg / mLCefepime 8μg / mLCefepime 16μg / mLCefepime 32μg / mLPiperacillin8μg / mLPiperacillin 16 μg / mLPiperacillin 32 μg / mLPiperacillin 64 μg / mLPiperacillin 128 μg / mLCefotaxime 0.5μg / mLNNNNNNPPPPPPCH061CH062CH063CH064CH065CH066CH067CH068CH069CH070CH071CH072Cefotaxime1μg / mLCefotaxime2μg / mLCefotaxime4μg / mLCefotaxime8μg / mLCefotaxime 16μg / mLCefotaxime 32μg / mLCeftazidime / clavulanate 1 / 4μg / mLCeftazidime / clavulanate 2 / 4μg / mLCefotaxime / clavulanate 0.25 / 4 μg / mLCefotaxime / clavulanate 0.5 / 4 μg / mLCeftriaxone0.5μg / mLCeftriaxone1μg / mLPPPPNNPPPPPPCH073CH074CH075CH076CH077CH078CH079CH080CH081CH082CH083CH084Ceftriaxone2μg / mLCeftriaxone4μg / mLCeftriaxone8μg / mLCeftriaxone16μg / mLCeftriaxone 32μg / mLOxacillin 0.25μg / mLOxacillin 0.5μg / mLCefazolin2μg / mLCefazolin4μg / mLCefazolin8μg / mLCefazolin 16μg / mLCefazolin 32μg / mLPPPPPPPPPPPPCH085CH086CH087CH088CH089CH090CH091CH092CH093CH094CH095CH096Aztreonam1μg / mLAztreonam2μg / mLAztreonam4μg / mLAztreonam8μg / mLAztreonam 16μg / mLAztreonam 32μg / mLImipenem1μg / mLImipenem2μg / mLImipenem4μg / mLImipenem8μg / mLImipenem 16μg / mLGentamicin2μg / mLPPPNNNPNNNNNCH097CH098CH099CH100CH101CH102CH103CH104CH105CH106CH107CH108Gentamicin4μg / mLGentamicin8μg / mLGentamicin 16μg / mLAmikacin8μg / mLAmikacin 16μg / mLAmikacin 32μg / mLAmikacin 64μg / mLStreptomycin 500μg / mLGentamicin1,000μg / mLDoxycycline2μg / mLDoxycycline4μg / mLDoxycycline8μg / mLNNNNNNNNNNNNCH109CH110CH111CH112CH113CH114CH115CH116CH117CH118CH119CH120Doxycycline 16μg / mLTetracycline1μg / mLTetracycline2μg / mLTetracycline4μg / mLTetracycline8μg / mLTetracycline 16μg / mLTigecycline0.5μg / mLTigecycline1μg / mLTigecycline2μg / mLTigecycline4μg / mLMinocycline4μg / mLMinocycline8μg / mLNNNNNNNNNNNNCH121CH122CH123CH124CH125CH126CH127CH128CH129CH130CH131CH132Minocycline 16μg / mLCiprofloxacin 0.06 μg / mLCiprofloxacin 0.12 μg / mLCiprofloxacin 0.25 μg / mLCiprofloxacin 0.5 μg / mLCiprofloxacin 1μg / mLCiprofloxacin 2μg / mLCiprofloxacin 4 μg / mLLevofloxacin 0.12μg / mLLevofloxacin 0.25μg / mLLevofloxacin 0.5μg / mLLevofloxacin1μg / mLNPPPPPPPNNNNCH133CH134CH135CH136CH137CH138CH139CH140CH141CH142CH143CH144Levofloxacin2μg / mLLevofloxacin4μg / mLLevofloxacin8μg / mLMoxifloxacin 0.5 μg / mLMoxifloxacin 1μg / mLMoxifloxacin 2μg / mLMoxifloxacin 4μg / mLTeicoplanin8μg / mLTeicoplanin 16μg / mLTeicoplanin 32μg / mLVancomycin1μg / mLVancomycin2μg / mLNNNNNNNPPPPPCH145CH146CH147CH148CH149CH150CH151CH152CH153CH154CH155CH156Vancomycin4μg / mLVancomycin8μg / mLVancomycin16μg / mLVancomycin32μg / mLLinezolid2μg / mLLinezolid4μg / mLLinezolid8μg / mLICR 1 / 0.5μg / mLICR 4 / 0.5μg / mLErythromycin 0.25μg / mLErythromycin 0.5μg / mLErythromycin1μg / mLPPPPPPPPPPPPCH157CH158CH159CH160CH161CH162CH163CH164CH165CH166CH167CH168Erythromycin2μg / mLErythromycin4μg / mLErythromycin8μg / mLClindamycin 0.25μg / mLClindamycin 0.5μg / mLClindamycin1μg / mLClindamycin2μg / mLClindamycin4μg / mLTrimethoprim / sulfamethoxazole0.5 / 9.5μg / mLTrimethoprim / sulfamethoxazole1 / 19μg / mLTrimethoprim / sulfamethoxazole2 / 38μg / mLTrimethoprim / sulfamethoxazole4 / 76μg / mLPPPPPPPPNNNNCH169CH170CH171CH172CH173CH174CH175CH176CH177CH178CH179CH180Trimethoprim / sulfamethoxazole8 / 152μg / mLNitrofurantoin32μg / mLNitrofurantoin64μg / mLNitrofurantoin128μg / mLMupirocin 128μg / mLMupirocin 256μg / mLNitroxoline 16μg / mLNitroxoline 32μg / mLColistin 1μg / mLColistin 2μg / mLColistin 4μg / mLColistin 8μg / mLNNNNNNNNNNNNCH181CH182CH183CH184CH185CH186CH187CH188CH189CH190CH191CH192Positive ControlNegative Control.

[0089] Looking at Table 1, channels 1 to 8 represent the judgment results of antibiotic data obtained by treating the antibiotic ampicillin at various concentrations (0.25, 0.5, 1, 2, 4, 8, 16, 32 μg / mL). The E. coli in question was judged to be positive for ampicillin up to a concentration of 32 μg / mL. Therefore, although the minimum inhibitory concentration of the E. coli in question for ampicillin is not precisely known, it can be calculated to exceed at least a concentration of 32 μg / mL.

[0090] In addition, channels 61 to 66 in Table 1 represent the judgment results of antibiotic data obtained by treating with cefotaxime at different concentrations (1, 2, 4, 8, 16, 32 μg / mL). The E. coli in question was judged to be positive up to a concentration of 8 (μg / mL), but negative at concentrations of 16 μg / mL and 32 μg / mL. Therefore, the minimum inhibitory concentration of the E. coli in question against cefotaxime can be calculated to be 16 μg / mL.

[0091] According to one embodiment of the present invention, the processor (150) can obtain antibiotic susceptibility information by determining one of the susceptibility (S), intermediate resistance (I), and resistance (R) of the target microorganism to the antibiotic type based on the minimum inhibitory concentration (S32).

[0092] At this time, the processor (150) can determine one of susceptibility, intermediate resistance, and resistance for each type of antibiotic by applying the antibiotic susceptibility test performance standards (e.g., CLSI guidelines, etc.) of the target microorganism. The antibiotic susceptibility test performance standards are established for each microorganism, and an example of the CLSI guidelines is illustrated in FIG. 8.

[0093] Table 2 below is an example of antibiotic susceptibility information obtained based on the results in Table 1 above.

[0094] No.AntibioticsMIC (μg / mL)S / I / R1Ampicillin, Amp> 32R2Ertapenem, Erta≤ 0.5S3Meropenem, Mero1*4Cefoxitin, Ctin≤ 2S5Cefuroxime, Cefu> 32*6Amoxicillin-clavulanate, Amx-c> 64 / 32R7Amoxicillin, Amx> 8*8Penicillin, Peni> 16*9Cefepime, Pime1S10Piperacillin, Pipe> 128R11Cefotaxime, Taxi16R12Ceftazidime-Clavulanate, Dime-cPos.13Cefotaxime-Clavulanate, Taxi-cPos.14Ceftriaxone, Axone> 32R15Oxacillin, Oxa> 0.5*16Cefazolin, Zolin> 32R17Aztreonam, Azt8I18Imipenem, Imi2I19Gentamicin, Gen≤ 2S20Amikacin, Amk≤ 8I21High level Streptomycin, HLSMNeg.22High level Gentamicin, HLGMNeg.23Doxycycline, Doxy≤ 2S24Tetracycline, Tetra≤ 1S25Tigecycline, Tige≤ 0.5*26Minocycline, Mino≤ 4S27Ciprofloxacin, Cipro> 4R28Levofloxacin, Levo≤ 0.12S29Moxifloxacin, Moxi≤ 0.5*30Teicoplanin, Tei> 32*31Vancomycin, Van> 32*32Linezolid, Line> 8*33Induced Clindamycin Resistance, ICRPos.34Erythromycin, Ery> 8*35Clindamycin, Clinda> 4*36Trimethoprim / sulfamethoxazole, Tri / Sul≤ 0.5 / 9.5S37Nitrofurantoin, Nifur≤ 32S38Mupirocin, Mupi≤ 128*39Nitroxoline, Nirox≤ 16*40Colistin, Col≤ 1I.

[0095] Looking at Table 2, the E. coli in question was determined to be positive for ampicillin up to a concentration of 32 μg / mL. Based on the guidelines of FIG. 8, the processor (150) can determine that the E. coli in question is resistant (R) to ampicillin.

[0096] In addition, since the minimum inhibitory concentration of the E. coli to the cytotoxic agent is 16 μg / mL, the processor (150) can determine that the E. coli to be resistant (R) to the cytotoxic agent based on the guidelines of FIG. 8.

[0097] The national research and development projects that supported this application are as follows.

[0098] [Project ID] 2420005260

[0099] [Project Number] RS-2024-00442440

[0100] [Ministry Name] Ministry of SMEs and Startups

[0101] [Name of Project Management (Specialist) Agency] Small and Medium Business Technology Information Promotion Agency

[0102] [Research Project Name] Small and Medium Enterprise Technology Innovation Development Project Export-Oriented (Export)

[0103] [Research Project Name] Commercialization of a Rapid, Fully Automated Antibiotic Susceptibility Testing System Using New Capacitance Measurement Technology

[0104] [Contribution rate] 1 / 1

[0105] [Name of the project performing organization] Protia Co., Ltd.

[0106] Research Period: August 1, 2024 - July 31, 2028

Claims

1. In an electronic device that performs antibiotic susceptibility testing of microorganisms, Obtain antibiotic data by measuring changes in electrical characteristics of target microorganisms following antibiotic treatment, By inputting the positive control data, negative control data, and antibiotic data of the target microorganism into a deep learning model trained to determine the positivity or negativity of the antibiotic data using the positive control data, negative control data, and antibiotic data of the microorganism, the positivity or negativity of the antibiotic data of the target microorganism is determined. An electronic device including a processor that identifies antibiotic susceptibility information of the target microorganism according to the type and concentration of antibiotic using the above judgment result.

2. In paragraph 1, The above deep learning model is, An electronic device characterized in that feature vectors of positive control data, negative control data, and antibiotic data of microorganisms are extracted, and the feature vector of the antibiotic data is trained to be similar to the feature vector of the negative control data or the positive control data based on the labeling of the positive or negative of the antibiotic data.

3. In paragraph 2, The above processor, Extract feature vectors of positive control data, negative control data, and antibiotic data of the target microorganisms, respectively. Using the above deep learning model, the similarity between the feature vector of the antibiotic data and the feature vector of the positive control data and the negative control data is measured, An electronic device that determines whether the antibiotic data of the target microorganism is positive or negative based on the above similarity.

4. In paragraph 1, The above processor, An electronic device that acquires antibiotic data through a biosensor including a plurality of cross-electrodes implemented to measure in real time changes in electrical characteristics according to whether microorganisms proliferate or not according to antibiotic treatment.

5. In paragraph 4, The above processor, An electronic device that calculates the minimum inhibitory concentration (MIC) for each antibiotic type of the target microorganism by matching the antibiotic type and concentration information corresponding to each of the plurality of cross electrodes to the judgment result.

6. In paragraph 5, The above processor, An electronic device that obtains antibiotic susceptibility information by determining one of susceptibility, intermediate resistance, and resistance of the target microorganism to each type of antibiotic based on the minimum inhibitory concentration.

7. In a method for testing antibiotic susceptibility of microorganisms performed by an electronic device, A step of acquiring antibiotic data by measuring changes in electrical characteristics of target microorganisms following antibiotic treatment; A step of inputting positive control data, negative control data, and antibiotic data of a target microorganism into a deep learning model trained to determine whether the antibiotic data of a microorganism is positive or negative using positive control data, negative control data, and antibiotic data of the microorganism, thereby determining whether the antibiotic data of the target microorganism is positive or negative; A method comprising a step of identifying antibiotic susceptibility information of the target microorganism according to the type and concentration of antibiotic using the above judgment result.

Images