Method, device and system for dynamically monitoring living cells

Abstract

The invention discloses a dynamic living cell monitoring method which comprises the following steps: acquiring a cell scanning picture group of living cells growing in a culture container under specified conditions, and preprocessing the pictures; inputting the preprocessed cell scanning picture group into the trained cell recognition model, and outputting a cell growth state picture group; inputting the cell growth state diagram group into a trained survival rate calculation model, and outputting the survival rate of cells in each cell growth state diagram; drawing the obtained cell survival rate values into a cell survival rate-time curve graph; and combining the generated cell survival rate-time curve graph with a nonlinear regression statistical method, and selecting specified parameters to calculate the half-inhibition concentration value of the drug effect. Monitoring the growth state of the cells under specified conditions, and calculating the survival rate of the cells and the half-inhibition concentration value of the drug effect based on the data; the efficiency and the accuracy in the fields of medicine research and development, medicine screening, treatment scheme effects and the like can be remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of cell detection, and in particular relates to a method, device and system for dynamic monitoring of living cells. Background Art

[0002] The latest global cancer data in 2021 shows that about 90% of cancer incidence is caused by solid tumors. The treatment of solid tumors is generally surgery combined with postoperative chemotherapy (radiotherapy). At present, the sensitivity test technology for selecting therapeutic drugs after surgery is limited. In clinical tests, no relevant in vitro diagnostic products (IVD) have been used for chemotherapy drug sensitivity tests on patients' autologous cells. The inability to obtain drug sensitivity test results in the first time is a major constraint in the process of individualized treatment and precision treatment. Summary of the invention

[0003] The purpose of the present invention is to address the problems existing in the prior art and to provide a method, device and system for dynamic monitoring of living cells. By continuously and visually monitoring the growth state of cells under the action of drugs, and calculating the cell survival rate and the half-inhibitory concentration values ​​of various drugs in a specific time period, the effects of medication regimens on target cells can be quickly compared, providing researchers with data support for subsequent research and assisting in clinical precision treatment.

[0004] In one aspect, the present invention provides a method for dynamic monitoring of living cells, comprising the following steps: Acquire a group of cell scanning images of living cells grown in a culture container under specified conditions, and pre-process the images; The pre-processed cell scan image group is input into the trained cell recognition model, and a cell growth state image group is output; Input the cell growth state graph group into the trained survival rate calculation model, and output the survival rate of the cells in each cell growth state graph; The obtained cell survival rate values ​​are plotted into a cell survival rate-time curve graph; The generated cell survival rate-time curve was combined with the nonlinear regression statistical method, and the specified parameters were selected to calculate the half-inhibitory concentration value of the drug effect.

[0005] Through the above technical scheme, through the cell recognition model and the survival rate calculation model, the cell survival rate and drug efficacy can be automatically and accurately calculated, human errors can be reduced, research efficiency can be improved, the cell growth state under specified conditions can be monitored, and the cell survival rate and the half-inhibitory concentration value of the drug efficacy can be calculated based on these data, which can significantly improve the efficiency and accuracy in the fields of drug development and cell biology research, and has broad application prospects; for example, the collected cells are derived from primary cells obtained from autologous surgery of tumor patients, which can represent individual genetic information and pathological information, and are of great significance for the pre-screening of precise tumor drug regimens; the above method can exclude ineffective medication for patients with newly diagnosed tumors, with an accuracy of more than 95%, avoiding economic waste and toxic side effects caused by ineffective treatment; and it can also find available drugs or regimens for patients with recurrent, drug-resistant, metastatic, and refractory tumors.

[0006] Furthermore, the specified conditions include drug type, drug concentration, cell type, shooting start time, and shooting end time. The clarification of these conditions helps the repeatability of the experiment and the reliability of the results, and also provides researchers with more flexibility in experimental design.

[0007] Further, the training of the cell recognition model includes: Construct a sample information set, including cell images and cell growth status images; construct a convolutional neural network model, where the input of the convolutional neural network is the cell image and the output is the cell growth status image; The convolutional neural network model is trained using the constructed sample information set, and the trained cell recognition model is output. This step is the basis for the subsequent cell survival rate calculation, and its accuracy directly affects the reliability of the final result.

[0008] Further, the training of the survival rate calculation model includes: Construct a sample data set of cell growth status diagrams and corresponding cell survival rates under different experimental conditions; Annotate the cell growth state diagram in the sample data set; Construct a convolutional neural network model whose input is the labeled image and output is the survival rate of cells in the image; The convolutional neural network model is trained using the loss function, and the trained survival rate calculation model is output. This process ensures that the survival rate calculation model can accurately calculate the cell survival rate, providing strong support for drug efficacy evaluation.

[0009] Furthermore, the growth status of each cell is annotated, including marking the living cells or dead cells in each picture. After the cell pictures are annotated, it is convenient for the convolutional neural network model to identify living cells or dead cells, and then calculate the cell survival rate of the picture.

[0010] In one aspect, a living cell dynamic monitoring device is provided, comprising: a first main module for acquiring a group of cell scanning pictures taken of living cells grown in a culture container under specified conditions, and preprocessing the pictures; The second main module is used to input the pre-processed cell scanning image group into the trained cell recognition model and output the cell growth state image group; The third main module is used to input the cell growth state diagram group into the trained survival rate calculation model, and output the survival rate of the cells in each cell growth state diagram; The fourth main module is used to plot the obtained cell survival rate values ​​into a cell survival rate-time curve graph; The fifth main module is used to combine the generated cell survival rate-time curve with the nonlinear regression statistical method and select the specified parameters to calculate the half-inhibitory concentration value of the drug efficacy.

[0011] In one aspect, the present invention provides a living cell dynamic monitoring system, comprising a culture medium module, an imaging module and a main control module; The culture carrier module is configured to carry a cell culture carrier container; The imaging module is configured to acquire images of living cells grown in a culture container; The main control module is configured to implement the method described in any one of claims 1 to 5.

[0012] Furthermore, the imaging module includes an LED cold light source, an objective lens module, and a camera module, wherein the objective lens module is located between the observed object and the camera; The LED cold light source comprises an LED light source, a light collecting lens group, a color filter, a field lens group, an aperture grating, a field grating and a filter which are arranged in sequence from top to bottom.

[0013] Through precise power design and optical path control, the LED cold light source ensures that only the imaging area and its vicinity will receive sufficient illumination, while other non-imaging areas are almost unaffected. This ability to precisely control the illumination range not only improves the clarity and quality of imaging, but also greatly reduces the phototoxic effects on non-imaging areas. This structural design helps to provide stable and uniform illumination conditions, reduce the interference of illumination on cell growth status, and improve the accuracy of monitoring results. It can only provide sufficient illumination to meet imaging needs, and can also effectively reduce the phototoxic effects on non-imaging areas, protecting the safety of sensitive biological samples being observed.

[0014] Furthermore, it also includes a communication module, a storage module, and a display module, which are respectively used for data storage, data storage, and result display.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. Monitor the cell growth state under specified conditions, and calculate the cell survival rate and half-inhibitory concentration value of the drug efficacy based on these data. This technical solution can significantly improve the efficiency and accuracy of drug development, cell biology research and other fields, and has broad application prospects; 2. The collected cells are all derived from primary cells of tumor patients' own sources, which can represent individual genetic information and pathological information, and are of great significance for accurate tumor medication; 3. Through the above method, ineffective medication can be excluded for patients with primary tumors, with an accuracy of more than 95%, avoiding economic waste and toxic side effects caused by ineffective treatment; 4. Available drugs or solutions can be found for patients with recurrent, drug-resistant, metastatic, and refractory tumors. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a schematic diagram of a method for dynamic monitoring of living cells according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a system for dynamic monitoring of living cells according to an embodiment of the present invention; Figure 3 A schematic diagram showing a comparison of cell images before and after preprocessing in a living cell dynamic monitoring system according to an embodiment of the present invention; Figure 4 A schematic diagram of various cancer cells selected in a method for dynamic monitoring of living cells in an embodiment of the present invention; Figure 5 Schematic diagram of growth curves of different cell areas in the embodiment of the present invention; Figure 6 This is a schematic diagram of the state of liver cancer cells before drug addition in the experimental example of the present invention; Figure 7 Schematic diagram of the state of liver cancer cells in the experimental example of the present invention at 24 hours, 48 ​​hours and 72 hours after adding 50 μM of the drug; Figure 8 Schematic diagram of the state of liver cancer cells in the experimental example of the present invention after adding 10 μM of the drug for 24 hours, 48 ​​hours and 72 hours; Fig. 9 Schematic diagram of the state of liver cancer cells in the experimental example of the present invention at 24 hours, 48 ​​hours and 72 hours after adding 2 μM of the drug; Fig.10 Schematic diagram of the state of liver cancer cells in the experimental example of the present invention after adding 0.4 μM of the drug for 24 hours, 48 ​​hours and 72 hours; Fig.11 Schematic diagram of the state of liver cancer cells in the experimental example of the present invention after adding 0.08 μM of the drug for 24 hours, 48 ​​hours and 72 hours; Fig.12Schematic diagram of the state of liver cancer cells in the experimental example of the present invention after adding 0.016 μM of the drug for 24 hours, 48 ​​hours and 72 hours; Fig.13 Schematic diagram of the state of the liver cancer cell control group in the experimental example of the present invention after 24 hours, 48 ​​hours and 72 hours; Fig.14 Schematic diagram of the state of the liver cancer cell control group negative for 24 hours, 48 ​​hours and 72 hours in the experimental example of the present invention; Fig.15 Schematic diagram of the state of liver cancer cells in the experimental example of the present invention at 24 hours, 48 ​​hours and 72 hours after no drug is added; Fig.16 This is a schematic diagram of the analysis results of liver cancer cells in the experimental example of the present invention; Fig.17 This is a schematic diagram of the state of prostate cancer cells before drug addition in the experimental example of the present invention; Fig.18 Schematic diagram of the state of prostate cancer cells in the experimental example of the present invention after adding 50 μM of the drug for 24 hours, 48 ​​hours and 72 hours; Fig.19 Schematic diagram of the state of prostate cancer cells 24 hours, 48 ​​hours and 72 hours after adding 10 μM of the drug in the experimental example of the present invention; Fig. 20 Schematic diagram of the state of prostate cancer cells 24 hours, 48 ​​hours and 72 hours after adding 2 μM of the drug in the experimental example of the present invention; Fig.21 Schematic diagram of the state of prostate cancer cells in the experimental example of the present invention after adding 0.4 μM of the drug for 24 hours, 48 ​​hours and 72 hours; Fig. 22 Schematic diagram of the state of prostate cancer cells in the experimental example of the present invention after adding 0.08 μM of the drug for 24 hours, 48 ​​hours and 72 hours; Fig.23 Schematic diagram of the state of prostate cancer cells in the experimental example of the present invention after adding 0.016 μM of the drug for 24 hours, 48 ​​hours and 72 hours; Fig.24 Schematic diagram of the state of the prostate cancer cell control group in the experimental example of the present invention after 24 hours, 48 ​​hours and 72 hours; Fig.25 Schematic diagram of the state of the negative prostate cancer cell control group at 24 hours, 48 ​​hours and 72 hours in the experimental example of the present invention; Fig.26 Schematic diagram of the state of prostate cancer cells in the experimental example of the present invention at 24 hours, 48 ​​hours and 72 hours after no drug is added; Fig. 27 It is a schematic diagram of the analysis results of prostate cancer cells in the experimental example of the present invention; Fig.28 This is a schematic diagram of the structure of a living cell dynamic monitoring device according to an embodiment of the present invention; Fig.29 This is a schematic diagram of the structure of an LED cold light source in a living cell dynamic monitoring device according to an embodiment of the present invention.

[0017] In the figure: 1. Shell; 2. LED cold light source; 3. Cell stage; 4. Objective lens module; 5. Camera module; 6. Main control module. DETAILED DESCRIPTION

[0018] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] In the description of the present invention, it should be noted that the directions or positional relationships indicated by terms such as "middle", "upper", "lower", "left", "right", "inside" and "outside" are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention.

[0020] like Figure 1 , Figure 3 , Figure 4 , Figure 5 As shown, on the one hand, the present invention provides a method for dynamic monitoring of living cells, comprising the following steps: Acquire a group of cell scanning images of living cells grown in a culture container under specified conditions, and pre-process the images; The pre-processed cell scan image group is input into the trained cell recognition model, and a cell growth state image group is output; Input the cell growth state graph group into the trained survival rate calculation model, and output the survival rate of the cells in each cell growth state graph; The obtained cell survival rate values ​​are plotted into a cell survival rate-time curve graph; The generated cell survival rate-time curve was combined with the nonlinear regression statistical method, and the specified parameters were selected to calculate the half-inhibitory concentration value of the drug effect.

[0021] Through the above technical solution, through the cell recognition model and the survival rate calculation model, the cell survival rate and drug efficacy can be automatically and accurately calculated, reducing human errors and improving research efficiency.

[0022] In this embodiment, the half-inhibitory concentration value of the drug effect is calculated by nonlinear regression statistical method using existing software.

[0023] Specifically: After obtaining the cell survival rate-time curve graph, import the graph directly into the Graphpad Prism software, and then select the Dose-response-Inhibition parameter in the Nonlinear regression (Curve fit) function under the analyses item of the software to calculate the half-inhibitory concentration value of the drug efficacy (IC50 value).

[0024] The detection results of the present invention output the IC50 value of the drug efficacy of the corresponding cells, and compare it with the IC50 reference value in the literature to automatically judge the strength of the drug efficacy, especially the experimental results of obviously insensitive drugs, to provide data support and reference standards for doctors' clinical medication.

[0025] Preprocessing operations include denoising, contrast enhancement, and normalization of images. These operations help improve image quality and provide a more favorable data basis for subsequent identification, thereby improving the accuracy of overall monitoring.

[0026] The specified conditions include drug type, drug concentration, cell type, shooting start time, and shooting end time. The clarity of these conditions helps the repeatability of the experiment and the reliability of the results, while also providing researchers with more flexibility in experimental design.

[0027] The training of the cell recognition model includes: Construct a sample information set, including common cell image data; construct a convolutional neural network model. Specifically, in this embodiment, a Mask R-CNN model is used. The input of Mask R-CNN is a clear cell image, and the output is a cell growth status image; The Mask R-CNN model is trained using the constructed sample information set, and the trained cell recognition model is output.

[0028] By building a Mask R-CNN model and training it with sample information sets, we obtained a cell recognition model that can output cell image data. This step is the basis for subsequent cell survival rate calculations, and its accuracy directly affects the reliability of the final results.

[0029] The Mask R-CNN model is an existing network model. The algorithm steps of the Mask R-CNN model are as follows: First, input an image you want to process, and then perform the corresponding preprocessing operation, or the preprocessed image; Then, input it into a pre-trained neural network (ResNeXt, etc.) to obtain the corresponding feature map; Next, a predetermined number of ROIs are set for each point in the feature map, thereby obtaining multiple candidate ROIs; Next, these candidate ROIs are sent to the RPN network for binary classification (foreground or background) and BB regression to filter out some candidate ROIs; Next, perform ROIAlign operation on the remaining ROIs (i.e. first align the pixels of the original image and feature map, and then align the feature map with the fixed features); Finally, these ROIs are classified (N-category classification), BB regression and MASK generation (FCN operation is performed in each ROI).

[0030] The Mask R-CNN model consists of three modules: Faster-rcnn, ROIAlign, and FCN.

[0031] Among them, Faster R-CNN (Regional Convolutional Neural Network) is a deep learning model widely used in target detection tasks. It combines the advantages of RPN (Region Proposal Network) and Fast R-CNN. The following is the main structure and working principle of Faster R-CNN: Basic network: usually use pre-trained convolutional neural networks (such as VGG, ResNet, etc.) to extract image features.

[0032] Region Proposal Network (RPN): This is a small convolutional network that runs in parallel and generates a set of candidate object regions (called "RoIs" - region of interest) for the input image. RPN predicts whether each candidate region contains an object and its category probability.

[0033] RoI Pooling: For each candidate region, features are sampled from the original feature map through a fixed-size window of the RoI pooling layer. This step ensures that all RoIs have the same size for further processing.

[0034] ROIAlign or RoIPooling: Replaces the earlier version of RoIPooling, processes the position information of the bounding box more accurately and reduces the loss of accuracy.

[0035] Classification and regression: The features of the RoI are passed to the fully connected layer to determine whether the area contains an object (binary classification problem) and determine the exact location of the object (bounding box coordinate regression).

[0036] Non-Maximum Suppression (NMS): In order to remove highly similar region proposals, the candidate region with the highest confidence is retained as the final result.

[0037] The key to Faster R-CNN is its end-to-end design, which can simultaneously complete the two steps of region proposal and target recognition, significantly improving the speed and accuracy of target detection.

[0038] Secondly, ROIAlign (Region of Interest Alignment) is a technology in Faster R-CNN. It is an improvement over the earlier RoIPooling. It is mainly used to process the region proposal (RoI, Regions of Interest) operation of target detection tasks in convolutional neural networks. The core idea of ​​ROIAlign is to perform pixel-level interpolation according to the bounding box of each RoI before feature aggregation (pooling) to ensure that each pixel of the output feature map is a linear transformation of the actual position on the original image.

[0039] Finally, the FCN algorithm is a classic semantic segmentation algorithm that can accurately segment objects in images. Its overall architecture is shown in the figure above. It is an end-to-end network. The main modules include convolution and deconvolution, that is, first convolve and pool the image to continuously reduce the size of its feature map; then perform deconvolution, that is, perform interpolation operations to continuously increase its feature map, and finally classify each pixel value. This achieves accurate segmentation of the input image.

[0040] The training of the survival rate calculation model includes: Construct a sample data set of cell growth status diagrams and corresponding cell survival rates under different experimental conditions (such as cells under different drugs, different drug concentrations, different placement times, etc.); Annotate the cell growth state diagram in the sample data set; Construct a convolutional neural network model whose input is the labeled image and output is the survival rate of cells in the image; The convolutional neural network model is trained using the loss function, and the trained survival rate calculation model is output. This process ensures that the survival rate calculation model can accurately calculate the cell survival rate, providing strong support for drug efficacy evaluation.

[0041] Among them, labeling each image includes marking the living cells or dead cells in each image. After the cell images are labeled, it is convenient for the convolutional neural network model to identify living cells or dead cells, and then calculate the cell survival rate of the image.

[0042] In this embodiment, the survival rate calculation model is used to simultaneously identify the number of surviving cells and the total number of cells in the image. Then, the survival rate calculation model substitutes the two parameters of the number of surviving cells and the total number of cells into formula (1) to obtain the cell survival rate of the image.

[0043] Formula (1): Cell survival rate = number of surviving cells / total number of cells*100% Experimental Example 1: Liver cancer cells were used as primary cells to administer drugs for monitoring.

[0044] Culture system: flat-bottom 96-well plate, liver cancer cell culture medium, seeding density 8×103 cells / well.

[0045] Experimental procedure: 24 hours after inoculation of cells, add corresponding compounds according to the experimental design and observe for 72 hours.

[0046] The working concentration of drugs is shown in Table 1: Table 1 Liver cancer cell monitoring statistics

[0047] Cell pictures such as Figure 6-15 As shown: Specifically, (healthy cells) before drug addition (Group 1) Figure 6 As shown; Test results: After adding the drug (6 groups) Figure 7-15 , killing effectiveness determination: (5-fluorouracil) The recommended therapeutic dose for patients with good hematopoietic function and nutritional status is 12 mg / kg per day by intravenous injection, and the maximum daily dose is 800 mg. After 4 days of injection, if no toxicity is found, the dose is then changed to 6 mg / kg, once every other day, for a total of 4 times. The literature recommends an experimental dose of 1 μM.

[0048] The test results are shown in Table 2: Table 2 Statistical table of survival rate of liver cancer cells

[0049] like Fig.17As shown, the report output results: half inhibition concentration test value 0.7686μM, half inhibition concentration reference value 0.8μM.

[0050] Experimental Example 2: Prostate cancer cells were used as primary cells to administer drugs for monitoring.

[0051] Culture system: flat-bottom 96-well plate, prostate cancer cell culture medium, seeding density 5×103 cells / well.

[0052] Experimental procedure: 24 hours after inoculation of cells, add corresponding compounds according to the experimental design and observe for 72 hours.

[0053] The working concentration comparison of drugs is shown in Table 3: Table 3 Prostate cancer cell monitoring statistics

[0054] Prostate cancer cell pictures Figure 17-26 As shown: Specifically, (healthy cells) before drug addition (Group 1) Fig.17 As shown; Test results: After adding the drug (6 groups) Figure 18-26 , killing effectiveness determination.

[0055] Killing effectiveness determination: (Etoposide) The treatment dose in the instructions is 60-100 mg / m2 (body surface) per day for 3-5 consecutive days, and repeated every 3-4 weeks. The literature recommends an experimental dose of 1-100 μM.

[0056] The test results are shown in Table 4: Table 4 Statistical table of survival rate of prostate cancer cells

[0057] like Fig. 27 As shown, the report output results: half inhibition concentration test value 4.351μM, IC50 reference value 3.56μM.

[0058] Through the above method, compared with the problem of lack of individual differences in traditional tumor cell line culture, in this application, the collected cells are all derived from primary cells of the tumor patient's own source, which can represent individual genetic information and pathological information, and are of great significance for the precise use of tumor drugs. At the same time, the above method can exclude ineffective drugs for patients with initial tumors, with an accuracy of more than 95%, avoiding economic waste and toxic side effects caused by ineffective treatment. Secondly, it can find available drugs or schemes for patients with recurrent, resistant, metastatic, and refractory tumors. At the same time, it provides a scientific basis for the use of off-label guidelines, such as the use of targeted drugs for multiple purposes, and provides high-throughput in vitro drug sensitivity testing for patients who lack effective treatment drugs in the guidelines, including a drug library approved by the FDA, which can screen hundreds of drugs and schemes in about two weeks.

[0059] In addition, this application can use primary cells to establish a drug genome database for different tumors, providing support for the development of new drugs and target discovery for related cancer types.

[0060] like Fig.28 As shown, based on the same technical concept as the above embodiment, on one hand, a living cell dynamic monitoring device is provided, wherein the first main module is used to obtain a group of cell scanning pictures taken of living cells grown in a culture container under specified conditions, and to pre-process the pictures; The second main module is used to input the pre-processed cell scanning image group into the trained cell recognition model and output the cell growth state image group; The third main module is used to input the cell growth state diagram group into the trained survival rate calculation model, and output the survival rate of the cells in each cell growth state diagram; The fourth main module is used to plot the obtained cell survival rate values ​​into a cell survival rate-time curve graph; The fifth main module is used to combine the generated cell survival rate-time curve with the nonlinear regression statistical method and select the specified parameters to calculate the half-inhibitory concentration value of the drug efficacy.

[0061] Based on the same technical concept of the above embodiments, on one hand, a living cell dynamic monitoring system is provided, including a loading module, an imaging module and a main control module; The culture carrier module is configured to carry a cell culture carrier container; The imaging module is configured to acquire images of living cells grown in a culture container; The main control module is configured to scan the cell picture group taken by the imaging unit, identify the cell picture group in turn, calculate the cell survival rate, generate a cell survival rate and time curve graph, and calculate the half-inhibitory concentration value of the corresponding drug.

[0062] The loading module is arranged at the center of the top of the housing 1 , and specifically includes a cell loading platform 2 and living cells grown in a culture container.

[0063] like Fig.28 As shown, the imaging module includes an LED cold light source 2, an objective lens module 4 and a camera module 5; The LED cold light source 2 is arranged above the cell stage 3, and the objective lens module 4 and the camera module 5 are arranged below the cell stage 3 from top to bottom.

[0064] like Fig.29 As shown, the LED cold light source includes an LED light source, a light collecting lens group, a color filter, a field lens group, an aperture grating, a field grating and a filter which are arranged in sequence from top to bottom.

[0065] Through precise power design and optical path control, the LED cold light source 2 ensures that only the imaging area and its vicinity will receive sufficient illumination, while other non-imaging areas are almost unaffected. This ability to precisely control the illumination range not only improves the clarity and quality of imaging, but also greatly reduces the phototoxic effects on non-imaging areas. This structural design helps to provide stable and uniform illumination conditions, reduce the interference of illumination on cell growth status, and improve the accuracy of monitoring results. It can only provide sufficient illumination to meet imaging requirements, and can also effectively reduce the phototoxic effects on non-imaging areas, protecting the safety of sensitive biological samples being observed.

[0066] It should be noted that in the continuous observation of living cells, the culture area is much larger than the imaging area of ​​a single field of view. The traditional shooting sampling method is to use multi-image stitching to image the entire culture area. This technology is mature, but the data throughput is too large, the equipment requirements are too high, and the amount of redundant data generated is too much, which not only causes unnecessary interference to the result data judgment, but also wastes data equipment processing and storage resources, increases unnecessary burdens on clinical treatment and research, and is not conducive to the implementation of technology and large-scale promotion. Therefore, in this embodiment, the objective lens module and the camera module are integrated and can be moved on x and y on a surface through a driving structure such as an electric slide rail. When shooting and sampling, refer to the classic blood cell counting method and adopt a single-hole multi-field sampling and counting sampling method. Its advantages are as follows: (1) Comprehensiveness: By sampling the entire culture area in multiple fields of view, the comprehensiveness and representativeness of the data can be ensured, reducing the deviation caused by local sampling.

[0067] (2) Accuracy: Multi-field sampling can capture more cell status and distribution information, thereby improving the accuracy of counting. At the same time, random errors can be further reduced by taking multiple samples and calculating the average value.

[0068] (3) Reliability: This method avoids the possible randomness of single-field sampling, making the counting results more reliable and stable.

[0069] Single-hole multi-field sampling, collecting data from the entire culture area, can effectively avoid random field of view errors.

[0070] In this embodiment, the living cell dynamic monitoring device also includes a communication module, a storage module, and a display module, which are respectively used for data transmission, data storage, and result display. The main control module interacts with the storage unit and the display unit at the same time to realize data storage and result display. All data will be automatically uploaded to the cloud through the communication module regularly to reduce the hardware storage pressure, and data from different sources can also be integrated and shared, which is convenient for users to view and analyze at any time. Provide data export and report generation functions to facilitate users to share results with clinicians or other researchers.

[0071] The above equipment combined with the above monitoring method can fully support high-throughput chemotherapy drug sensitivity detection technology, and can select accurate medication plans for clinical tumor patients. During the clinical diagnosis and treatment process, fresh tumor samples such as intraoperative tissue, pleural effusion, puncture samples, and biopsy tissues of the patient are obtained, and the new conditional reprogramming cell (CRC) technology is used to rapidly expand and enrich primary cells within 72 hours. On the premise of preserving the genetic, pathological and other biological characteristics of primary cells and original tissues, more than 100 drugs or combination plans are quickly screened within 1-2 weeks to help patients exclude drugs that are already resistant or ineffective (more than 90% compliance rate), and provide medication references for tumor patients in the middle and late stages, drug-resistant recurrence, and lack of guideline reference medication, and formulate accurate diagnosis and treatment plans to further reduce the incidence of ineffective treatment in clinical tumor patients and improve the success rate of medication.

[0072] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for dynamic monitoring of living cells, characterized in that: The following steps are involved: Acquire a group of cell scanning images of living cells grown in a culture container under specified conditions, and pre-process the images; The pre-processed cell scan image group is input into the trained cell recognition model, and a cell growth state image group is output; Input the cell growth state graph group into the trained survival rate calculation model, and output the survival rate of the cells in each cell growth state graph; The obtained cell survival rate values ​​are plotted into a cell survival rate-time curve graph; The generated cell survival rate-time curve was combined with the nonlinear regression statistical method, and the specified parameters were selected to calculate the half-inhibitory concentration value of the drug effect.

2. A method for dynamic monitoring of living cells according to claim 1, characterized in that: The specified conditions include drug type, drug concentration, cell type, shooting start time, and shooting end time.

3. A method for dynamic monitoring of living cells according to claim 1, characterized in that: The training of the cell recognition model includes: Construct a sample information set, including cell images and cell growth status images; construct a convolutional neural network model, where the input of the convolutional neural network is the cell image and the output is the cell growth status image; The convolutional neural network model is trained using the constructed sample information set, and the trained cell recognition model is output.

4. The method for dynamic monitoring of living cells according to claim 1, characterized in that: The training of the survival rate calculation model includes: Construct a sample data set of cell growth status diagrams and corresponding cell survival rates under different experimental conditions; Annotate the cell growth state diagram in the sample data set; Construct a convolutional neural network model whose input is the labeled image and output is the survival rate of cells in the image; The convolutional neural network model is trained using the loss function, and the trained survival rate calculation model is output.

5. A method for dynamic monitoring of living cells according to claim 4, characterized in that: Annotate each cell growth status image, including marking the living cells or dead cells in each image.

6. A living cell dynamic monitoring device, characterized in that: include: The first main module is used to obtain a group of cell scanning pictures taken of living cells growing in a culture container under specified conditions, and to pre-process the pictures; The second main module is used to input the pre-processed cell scanning image group into the trained cell recognition model and output the cell growth state image group; The third main module is used to input the cell growth state diagram group into the trained survival rate calculation model, and output the survival rate of the cells in each cell growth state diagram; The fourth main module is used to plot the obtained cell survival rate values ​​into a cell survival rate-time curve graph; The fifth main module is used to combine the generated cell survival rate-time curve with the nonlinear regression statistical method and select the specified parameters to calculate the half-inhibitory concentration value of the drug efficacy.

7. A living cell dynamic monitoring system, characterized in that: It includes a culture and loading module, an imaging module and a main control module; The culture carrier module is configured to carry a cell culture carrier container; The imaging module is configured to acquire images of living cells grown in a culture container; The main control module is configured to implement the method described in any one of claims 1 to 5.

8. A living cell dynamic monitoring system according to claim 7, characterized in that: The imaging module includes an LED cold light source, an objective lens module, and a camera module, wherein the objective lens module is located between the observed object and the camera; The LED cold light source comprises an LED light source, a light collecting lens group, a color filter, a field lens group, an aperture grating, a field grating and a filter which are arranged in sequence from top to bottom.

9. A living cell dynamic monitoring system according to claim 7, characterized in that: It also includes a communication module, a storage module, and a display module, which are respectively used for data storage, data storage, and result display.

Images