CN110632069B - Circulating tumor cell detection method, device, equipment and medium - Google Patents
Circulating tumor cell detection method, device, equipment and medium Download PDFInfo
- Publication number
- CN110632069B CN110632069B CN201910768669.8A CN201910768669A CN110632069B CN 110632069 B CN110632069 B CN 110632069B CN 201910768669 A CN201910768669 A CN 201910768669A CN 110632069 B CN110632069 B CN 110632069B
- Authority
- CN
- China
- Prior art keywords
- circulating tumor
- tumor cells
- digital
- cell
- classification
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N1/00—Sampling; Preparing specimens for investigation
- G01N1/28—Preparing specimens for investigation including physical details of (bio-)chemical methods covered elsewhere, e.g. G01N33/50, C12Q
- G01N1/40—Concentrating samples
- G01N1/4077—Concentrating samples by other techniques involving separation of suspended solids
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N1/00—Sampling; Preparing specimens for investigation
- G01N1/28—Preparing specimens for investigation including physical details of (bio-)chemical methods covered elsewhere, e.g. G01N33/50, C12Q
- G01N1/40—Concentrating samples
- G01N1/4077—Concentrating samples by other techniques involving separation of suspended solids
- G01N2001/4088—Concentrating samples by other techniques involving separation of suspended solids filtration
Landscapes
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Pathology (AREA)
- Immunology (AREA)
- General Health & Medical Sciences (AREA)
- Biochemistry (AREA)
- Analytical Chemistry (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Investigating Or Analysing Biological Materials (AREA)
Abstract
The invention discloses a circulating tumor cell detection method, a circulating tumor cell detection device, circulating tumor cell detection equipment and a circulating tumor cell detection medium. The method comprises the following steps: injecting a lymph fluid sample to be detected into the comprehensive molten iron power cell separation iFCS device to obtain a lymph fluid sample after white blood cells are filtered; generating a digital slice of the leukocyte-filtered lymph fluid sample; classifying the digital slices by adopting an image classification identification model to obtain a classification result of classifying the pixel points in the digital slices according to the cell types to which the pixel points belong; and detecting whether the digital slice contains the circulating tumor cells according to the classification result to obtain a detection result. According to the embodiment of the invention, the concentration of the circulating tumor cells can be increased through the iFCS device, and the accuracy and the success rate of the identification of the circulating tumor cells are improved.
Description
Technical Field
The invention belongs to the technical field of cell detection, and particularly relates to a circulating tumor cell detection method, a circulating tumor cell detection device, circulating tumor cell detection equipment and a circulating tumor cell detection medium.
Background
With the rapid development of digital pathology, doctors can directly scan the slide of lymph fluid sample by a scanner to obtain digital sections and display the digital sections on a computer. And then classifying the digital slices through the trained image recognition classification model.
Since only a few circulating tumor cells are transferred to sentinel lymph nodes at the early stage of metastasis of the circulating tumor cells of the breast, a doctor cannot completely ensure that extracted lymph tissues contain the circulating tumor cells in the process of extracting a lymph fluid sample, and even if the extracted lymph tissues contain the circulating tumor cells, the image recognition and classification model has very high recognition accuracy, so that the digital slices need to be divided into the small digital slices for training in the model training process. Therefore, the current detection and identification of the circulating tumor cells have the problem of difficult model training and identification.
Disclosure of Invention
Embodiments of the present invention provide a method, an apparatus, a device, and a medium for detecting circulating tumor cells, which can increase the concentration of circulating tumor cells by an iFCS apparatus, thereby improving the accuracy and success rate of identifying circulating tumor cells.
In a first aspect, the embodiments of the present invention provide a method for detecting circulating tumor cells, the method including:
injecting a lymph fluid sample to be detected into a comprehensive molten iron dynamic cell separation iFCS device to obtain a lymph fluid sample after leukocyte filtration;
generating a digital slice of the leukocyte-filtered lymph fluid sample;
classifying the digital slices by adopting an image classification identification model to obtain a classification result of classifying pixel points in the digital slices according to cell types to which the pixel points belong;
and detecting whether the digital slices contain circulating tumor cells or not according to the classification result to obtain a detection result.
According to one aspect of the embodiment of the invention, the image classification recognition model is a U-Net convolutional neural network model.
According to an aspect of the embodiment of the invention, the classification result is a cell classification probability map; detecting whether the digital slice contains circulating tumor cells according to the classification result to obtain a detection result, wherein the detection result comprises the following steps:
overlapping the cell classification probability map and the digital slice to obtain a heat map;
detecting whether the heat map contains pixel points belonging to circulating tumor cells;
and if so, identifying the region of the pixel point belonging to the circulating tumor cell in the heat map to obtain the identified heat map.
According to an aspect of embodiments of the present invention, the region of the heat map belonging to a pixel point of the circulating tumor cell comprises: and the region of the pixel point with the saturation exceeding the preset saturation threshold value in the heat map.
According to an aspect of the embodiments of the present invention, the detecting whether the digital slice includes circulating tumor cells according to the classification result to obtain a detection result includes:
inputting the cell classification probability map into two continuous full-connection layers;
and inputting the output of the full-junction layer into a softmax layer to obtain a judgment result of whether the digital section contains circulating tumor cells.
According to an aspect of the embodiment of the present invention, after the generating the digital slice of the leukocyte-filtered lymph fluid sample, the method further includes:
segmenting the digital slice by adopting a threshold-based segmentation algorithm to obtain a tissue slice with a background region removed;
the classifying the digital slices by adopting the image classification and identification model specifically comprises the following steps:
and classifying the tissue slices with the background areas removed by adopting the image classification and identification model.
According to an aspect of the embodiment of the present invention, after obtaining the tissue slice with the background region removed, the method further includes:
dividing the tissue section into a plurality of tissue sections with preset sizes;
the step of classifying the digital slices by using the image classification and identification model specifically comprises the following steps:
and classifying the tissue sections with the preset size by adopting the image classification and identification model.
In a second aspect, an embodiment of the present invention provides a circulating tumor cell detection apparatus, including:
the cell filtering module is used for injecting the lymph fluid sample to be detected into the comprehensive molten iron dynamic cell separation iFCS device to obtain the lymph fluid sample after filtering white blood cells;
a section generating module for generating a digital section of the lymph fluid sample after filtering the white blood cells;
the image classification module is used for classifying the digital slices by adopting an image classification identification model to obtain a classification result of classifying the pixel points in the digital slices according to the cell types to which the pixel points belong;
and the detection module is used for detecting whether the digital slices contain circulating tumor cells or not according to the classification result to obtain a detection result.
In a third aspect, an embodiment of the present invention provides a circulating tumor cell detection apparatus, including: a processor and a memory storing computer program instructions;
the processor, when executing the computer program instructions, implements a method of circulating tumor cell detection as described in the first aspect.
In a fourth aspect, embodiments of the present invention provide a computer-readable storage medium having stored thereon computer program instructions, which when executed by a processor, implement the method for detecting circulating tumor cells according to the first aspect.
According to the circulating tumor cell detection method, the device, the equipment and the medium provided by the embodiment of the invention, the comprehensive molten iron dynamic cell separation iFCS device is utilized to separate and concentrate the white blood cells and the circulating tumor cells in the lymph fluid sample to be detected, then the lymph fluid sample after the white blood cells are filtered is obtained, so that the concentration of the circulating tumor cells in the lymph fluid sample after the white blood cells are filtered is increased, and then the digital section of the lymph fluid sample after the white blood cells are filtered is sent to the image classification and identification model for classification and identification to detect whether the circulating tumor cells are contained in the digital section, so that the detection of the circulating tumor cells is completed. Therefore, the embodiment increases the concentration of the circulating tumor cells by integrating the molten iron dynamic cell separation iFCS device, reduces the requirement of the number of slices to be made, further reduces the difficulty degree of model training and identification of detection and identification of the circulating tumor cells, and improves the accuracy and success rate of identification of the circulating tumor cells.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required to be used in the embodiments of the present invention will be briefly described below, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method for detecting circulating tumor cells according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of an iFCS device according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a digital slice provided by one embodiment of the present invention;
FIG. 4 is a schematic diagram of a U-Net convolutional neural network model;
FIG. 5 is a schematic structural diagram of a circulating tumor cell detection apparatus according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of a circulating tumor cell detection apparatus according to an embodiment of the present invention.
Detailed Description
Features and exemplary embodiments of various aspects of the present invention will be described in detail below, and in order to make objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not to be construed as limiting the invention. It will be apparent to one skilled in the art that the present invention may be practiced without some of these specific details. The following description of the embodiments is merely intended to provide a better understanding of the present invention by illustrating examples of the present invention.
Lymph node metastasis is a common tumor metastasis mode, which refers to a phenomenon that infiltrated tumor cells pass through the wall of a lymphatic vessel, are taken to lymph nodes in a confluence area along with lymph fluid after shedding, and grow the same tumor by taking the lymph nodes as the center. Axillary lymph nodes are the main way of metastasis of breast cancer, and whether the axillary lymph nodes are affected or not has important significance on the staging of the tumor, the formulation of a treatment scheme and the prognosis evaluation.
The most common method of assessing lymph node status is sentinel lymph node biopsy, since sentinel lymph nodes are those which are most likely to contain circulating tumor cells, and are typically located and excised by a surgeon and then histopathologically processed and examined by a pathologist. The diagnosis process of a pathologist is tedious, takes long time and mainly depends on the subjective judgment of doctors. However, since only a few circulating tumor cells are transferred into the sentinel lymph node at the initial stage of metastasis of the circulating tumor cells of the breast, a large number of tissue images need to be examined, and some small circulating tumor cells are easily missed.
Therefore, separating circulating tumor cells, such as breast cancer cells, from the lymph fluid presents a great technical challenge, not only because the circulating tumor cells are present in low amounts, but also because circulating tumor cells themselves are very difficult to separate, and a single circulating tumor cell can exhibit many characteristics in a sample containing hundreds of circulating tumor cells. For example, there may be skin cells, muscle cells, and cell sizes that vary greatly.
In order to solve the problems of the prior art, the embodiment of the invention provides a circulating tumor cell detection method, a circulating tumor cell detection device, equipment and a medium. The method for detecting circulating tumor cells provided by the embodiments of the present invention will be described first.
Fig. 1 is a schematic flow chart of a method for detecting circulating tumor cells according to an embodiment of the present invention.
S1: injecting a lymph fluid sample to be detected into the comprehensive molten iron power cell separation iFCS device to obtain a lymph fluid sample after white blood cells are filtered;
by filtering leukocytes, the total amount of cells in the lymph fluid sample can be reduced, making a smaller number of circulating tumor cells more readily identifiable.
S2: generating a digital slice of the leukocyte-filtered lymph fluid sample;
s3: classifying the digital slices by adopting an image classification and identification model to obtain a classification result of classifying the pixel points in the digital slices according to the cell categories to which the pixel points belong;
among them, since the purpose of the present invention is to identify circulating tumor cells, the cell types herein refer to: whether they belong to circulating tumor cells or not.
S4: and classifying the digital slices by adopting an image classification and identification model to obtain a classification result of classifying the pixel points in the digital slices according to the cell categories to which the pixel points belong.
In the embodiment, the comprehensive molten iron dynamic cell separation iFCS device is used for performing concentrated operation after separation on white blood cells and circulating tumor cells in a lymph fluid sample to be detected, and then the lymph fluid sample after white blood cell filtration is obtained, so that the concentration of the circulating tumor cells in the lymph fluid sample after white blood cell filtration is increased, and then the digital section of the lymph fluid sample after white blood cell filtration is sent to an image classification recognition model for classification recognition to detect whether the digital section contains the circulating tumor cells, so that the detection of the circulating tumor cells is completed. Therefore, in the embodiment, the circulating tumor cells and the whole blood sample are separated as much as possible by the iFCS device, the concentration of the circulating tumor cells is increased, the requirement for the number of slices to be made is reduced, the difficulty degree of model training and identification of detection and identification of the circulating tumor cells is further reduced, and the accuracy and success rate of identification of the circulating tumor cells are improved. And subsequently, the time for a doctor to check the digital slices is reduced, and the subjectivity of diagnosing whether the circulating tumor cells exist according to the digital slices is reduced.
In addition, in the image recognition classification model, the training samples are also trained by adopting the process, so that the number of the required slices of the training samples is reduced, and the training efficiency of the image recognition classification model is improved. In addition, due to the optimization of the training samples, the accuracy of the subsequent image recognition classification model recognition is improved.
Wherein, the iFCS device is implemented by a microfluidic chip, and fig. 2 shows a schematic structural diagram of the iFCS device according to an embodiment of the present invention. The working principle is as follows: the iFCS device contains a ferrofluid that can modulate the concentration of magnetic nanoparticles, the magnetic properties of which are modulated between those of white blood cells and circulating tumor cells (e.g., breast cancer cells) (approximately 0) by varying the concentration of magnetic nanoparticles in the ferrofluid. Because the white blood cells have a higher magnetization than the ferrofluid and the circulating tumor cells have a lower magnetization than the ferrofluid, adjusting the magnetic properties of the ferrofluid between the white blood cells and the circulating tumor cells causes the white blood cells and the circulating tumor cells to flow in different directions. Meanwhile, the iFCS device has the functions of clearing leucocytes and enriching circulating tumor cells. Eventually more than 99% of the circulating tumor cells in the lymph fluid sample can be captured.
The specific working process of the iFCS device is as follows: a sample of lymph fluid was injected into the iFCS device inlet and ferrofluid was injected from the other inlet of the iFCS device. Through mutual repulsion of cell magnetization in the ferrofluid, leucocytes and circulating tumor cells in different flow directions are distinguished. Then, the screening process of the device mainly comprises three steps: the first step is to remove large pieces of debris from the lymph fluid sample with a debris filter; the second step is to absorb additional magnetic beads and most circulating tumor cells; the third step is to focus the rest white blood cells to the middle of the channel through a white blood cell filter, push the circulating tumor cells to the side wall, filter the white blood cells, enrich the circulating tumor cells, and collect the filtered white blood cells and the lymph fluid sample after filtering the white blood cells through two outlets respectively. The filtered-off white blood cell lymph fluid sample is then collected in culture.
After the detection result of the circulating tumor cell is obtained, it can be used for detection of cancer, analysis and study of metastatic tumor, and the like. The separation of circulating tumor cells from lymph fluid not only can improve the efficiency of tumor cell recognition, but also can provide a basis for the subsequent enhancement of the understanding, diagnosis and prognosis of metastatic tumors.
Wherein, the digital slice is a full-view digital slice, and the processing procedure of the digital slice is as follows: the quantitative sample solution extracted from the culture medium is fixed on a slide glass, and is subjected to some necessary treatments and stained with hematoxylin and eosin, and the operations can be automatically completed by a digital section system, and after the operations are completed, the operations can be observed by using a microscope and a full-field digital section can be manufactured. The full-field digital section scanning is to acquire a high-resolution digital image by scanning and collecting through a full-automatic microscope or an optical amplification system, and to automatically perform high-precision multi-field seamless splicing and processing on the acquired high-resolution digital image by using a computer to acquire high-quality visual data (namely digital sections) to be applied to various fields of pathology.
Fig. 3 shows a schematic diagram of a digital slice provided by an embodiment of the invention. The full-field digital section (namely, the virtual section) is not a static picture, and is an image containing all lesion information on a glass section, and a digital section system scans the whole glass slide in a full-information and all-round way and quickly, so that the traditional materialized glass slide becomes a new generation of digital pathological section. Full field digital slice scanning and machine learning can be used to self-detect cancer metastasis in lymph nodes with high accuracy.
In an embodiment, the detection result obtained in S4 may be displayed.
In a preferred embodiment, the image classification and identification model is a U-Net convolutional neural network model.
U-Net is a variation of the convolutional neural network, and FIG. 4 shows a schematic structural diagram of the U-Net convolutional neural network model. The whole neural network of the U-Net mainly comprises two parts: a contracting path (contracting path) and an expanding path (expanding path). The search path is mainly used to capture context information (context information) in the picture, and the expansion path is used to accurately locate (localization) a portion that needs to be segmented in the picture. U-Net can train some relatively few-sample data, especially medically relevant data, with data augmentation (data augmentation), so the presence of U-Net is very helpful for deep learning medical imagery for less samples. Therefore, the U-Net convolutional neural network model is adopted to identify and detect the circulating tumor cells, so that the accuracy of identifying the circulating tumor cells can be improved.
In one embodiment, the classification result is a cell classification probability map (segmentation map); the step S4 includes:
overlaying the cell classification probability map and the digital section to obtain a heatmap (heatmap);
detecting whether the heat map contains pixel points belonging to the circulating tumor cells;
and if so, identifying the region of the pixel point belonging to the circulating tumor cell in the heat map to obtain the identified heat map.
In this embodiment, the cell classification probability map includes probabilities that each pixel belongs to each cell classification, and the cell classification probability map includes probabilities that each pixel belongs to two classes (e.g., (0.7, 0.3), because the present invention is divided into two classes, that is, the two classes belong to the circulating tumor cell and that each pixel does not belong to the circulating tumor cell. Through overlapping the cell classification probability map and the digital slice, pixel points with high probability and pixel points with low probability of belonging to the circulating tumor cells can be distinguished in the obtained heat map, so that cells in which regions are likely to belong to the circulating tumor cells can be distinguished, and the parts are identified. The heat map after the identification is used as a detection result, so that a user can intuitively know whether the lymph fluid contains the circulating tumor cells and the quantity of the circulating tumor cells, and the subsequent analysis of the circulating tumor cells and the like according to the heat map after the identification is facilitated. In addition, since the marking operation is difficult and the marking result is not accurate enough in the case of low circulating tumor cell concentration, the embodiment increases the circulating tumor cell concentration in the digital section by the iFCS device based on the above, thereby facilitating the subsequent marking operation and the accuracy of marking.
Further, in one embodiment, the regions of the heat map that belong to pixel points of the circulating tumor cells include: and the region of the pixel point with the saturation exceeding the preset saturation threshold in the heat map.
Saturation reflects the shade of the color. The cell classification probability graph and the digital slice are overlapped, the color of the pixel points with high probability of the circulating tumor cells in the overlapped heat map is darker, and conversely, the color is lighter, so that the regions which probably belong to the circulating tumor cells can be distinguished through the color of the pixel points in the heat map. This embodiment has set up a preset saturation threshold, through this threshold, can be according to the shade of colour with the pixel in the heat map divide into two types, belong to circulating tumor cell promptly and do not belong to circulating tumor cell, make the condition of circulating tumor cell that the user can audio-visually know to contain in the lymph fluid. The invention is not limited to the above-mentioned preset saturation threshold value, and the worker may set the specific value of the preset protection threshold value through experiments or experience values.
In another embodiment, the step S4 includes:
inputting the cell classification probability map into two continuous Fully Connected layers (Fully Connected Layer);
the output of the full junction layer was input to the softmax layer, and the result of determination as to whether or not the digital section contained circulating tumor cells was obtained.
In this embodiment, the fully-connected layer may integrate local information with category distinction in the convolutional layer or the pooling layer, and after the cell classification probability map is input into the two continuous fully-connected layers, the two fully-connected layers perform cell classification according to the cell classification probability map. Subsequently, by activating the softmax function, the classification result of the whole digital slice, that is, whether the digital slice belongs to an image containing circulating tumor cells or an image not containing circulating tumor cells, can be obtained, and the specific expression of the result of the classification can be the probability that the digital slice belongs to the above two image classes, or the class number of the image class to which the digital slice belongs, for example, 1 indicates that the digital slice contains circulating tumor cells, and 2 indicates that the digital slice does not contain circulating tumor cells. Of course, the determination result may be expressed in another manner, and it is only necessary to reflect the image type to which the digital slice belongs. By the method, people who do not know the digital section can know the detection result of the circulating tumor cells in the lymph fluid, so that the circulating tumor cells can be conveniently detected by users with weak specialties.
In one embodiment, after step S2, the method further includes:
segmenting the digital slice by adopting a threshold-based segmentation algorithm (Otsu's method) to obtain a tissue slice with a background region removed;
the step S3 includes: and classifying the tissue sections with the background areas removed by adopting an image classification and identification model.
Since the digital slice is particularly large, if the digital slice is directly subjected to image classification and identification, the amount of calculation is large, and the processing efficiency is low. Therefore, in this embodiment, the digital slice is segmented by the threshold-based segmentation algorithm, and the non-tissue region (i.e., the background region) is removed first, so that useless calculation is reduced, the analysis is performed while focusing on the cell tissue region, and the efficiency of image classification and recognition is improved.
Further, in another embodiment, after obtaining the tissue slice from which the background region is removed, the method further includes: dividing the tissue slice into a plurality of tissue slices with preset sizes;
accordingly, the step S3 is adjusted as follows: and classifying the tissue sections with preset sizes by adopting an image classification and identification model.
In this embodiment, the tissue slices are segmented to obtain tissue slices with smaller sizes, and in this case, the image classification and identification model is enabled to classify each small-sized tissue slice, so that compared with a mode of directly classifying the whole digital slice, the precision is higher, and the identification accuracy of the circulating tumor cells is improved.
In addition, in the process of training the image classification recognition model, in order to improve the accuracy and facilitate selection of the trained model, it is preferable that a segmentation algorithm is also used to segment the tissue region and the background region in the training sample, and the segmented tissue sample is segmented into smaller sample slices.
In other embodiments, the image classification and identification model may also adopt a google net model or an ssd (single shot multi box detector) model, or may also adopt other image classification and identification models.
Fig. 5 is a schematic structural diagram of a circulating tumor cell detection apparatus according to an embodiment of the present invention. The device comprises:
the cell filtering module 1 is used for injecting a lymph fluid sample to be detected into the comprehensive molten iron dynamic cell separation iFCS device to obtain a lymph fluid sample after leukocyte filtration;
a slice generating module 2, configured to generate a digital slice of the lymph fluid sample after filtering leukocytes;
the image classification module 3 is used for classifying the digital slices by adopting an image classification identification model to obtain a classification result of classifying the pixel points in the digital slices according to the cell types to which the pixel points belong;
and the detection module 4 is used for detecting whether the digital slice contains the circulating tumor cells according to the classification result to obtain a detection result.
In the embodiment, the comprehensive molten iron dynamic cell separation iFCS device is used for performing concentrated operation after separation on white blood cells and circulating tumor cells in a lymph fluid sample to be detected, and then the lymph fluid sample after white blood cell filtration is obtained, so that the concentration of the circulating tumor cells in the lymph fluid sample after white blood cell filtration is increased, and then the digital section of the lymph fluid sample after white blood cell filtration is sent to an image classification recognition model for classification recognition to detect whether the digital section contains the circulating tumor cells, so that the detection of the circulating tumor cells is completed. Therefore, in the embodiment, the circulating tumor cells and the whole blood sample are separated as much as possible by the iFCS device, the concentration of the circulating tumor cells is increased, the requirement for the number of slices to be made is reduced, the difficulty degree of model training and identification of detection and identification of the circulating tumor cells is further reduced, and the accuracy and success rate of identification of the circulating tumor cells are improved. And subsequently, the time for a doctor to check the digital slices is reduced, and the subjectivity of diagnosing whether the circulating tumor cells exist according to the digital slices is reduced.
In an embodiment, the system further comprises a display module, configured to display the obtained detection result.
In one embodiment, the image classification recognition model is a U-Net convolutional neural network model. The U-Net convolutional neural network model is adopted to identify and detect the circulating tumor cells, so that the accuracy of identifying the circulating tumor cells can be improved.
In one embodiment, the classification result is a cell classification probability map; the detection module 4 includes:
the overlapping unit is used for overlapping the cell classification probability map and the digital slice to obtain a heat map;
the detection unit is used for detecting whether the heat map contains pixel points belonging to the circulating tumor cells; if yes, triggering an identification unit;
and the identification unit is used for identifying the area of the pixel point belonging to the circulating tumor cell in the heat map to obtain the identified heat map.
In the embodiment, the cell classification probability map and the digital slice are overlapped, so that the pixel points with high probability and the pixel points with low probability of belonging to the circulating tumor cells can be distinguished in the obtained heat map, and therefore, the cells in which regions are likely to belong to the circulating tumor cells can be distinguished, and the parts are identified. The marked heat map is used as a detection result, so that a user can intuitively know whether the lymph fluid contains the circulating tumor cells and the quantity of the circulating tumor cells, and the subsequent analysis of the circulating tumor cells and the like according to the marked heat map is facilitated. In addition, since the marking operation is difficult and the marking result is not accurate enough in the case of low circulating tumor cell concentration, the embodiment increases the circulating tumor cell concentration in the digital section by the iFCS device based on the above, thereby facilitating the subsequent marking operation and the accuracy of marking.
In another embodiment, the detection module 4 comprises:
the full-connection unit is used for inputting the cell classification probability map into the two continuous full-connection layers;
and a determination unit for inputting the output of the all-connected layer to the softmax layer and obtaining a determination result of whether the digital section contains the circulating tumor cells.
In this embodiment, the obtained determination result is whether the digital section belongs to an image containing circulating tumor cells or an image not containing circulating tumor cells, so that a person who does not know the digital section can know the detection result of the circulating tumor cells in the lymph fluid according to the determination result, thereby facilitating the detection of the circulating tumor cells by a user with low expertise.
In one embodiment, the apparatus further comprises:
the background segmentation module is used for segmenting the digital slice by adopting a threshold-based segmentation algorithm to obtain a tissue slice with a background area removed;
correspondingly, the image classification module 3 is specifically configured to: and classifying the tissue slices with the background areas removed by adopting an image classification and identification model.
Since the digital slice is particularly large, if the digital slice is directly subjected to image classification and identification, the amount of calculation is large, and the processing efficiency is low. Therefore, in this embodiment, the digital slice is segmented by the threshold-based segmentation algorithm, and the non-tissue region (i.e., the background region) is removed first, so that useless calculation is reduced, the cell tissue region is more focused on the analysis, and the efficiency of image classification and identification is improved.
In another embodiment, the apparatus further comprises:
the slice dividing module is used for dividing the tissue slices into a plurality of tissue slices with preset sizes after the tissue slices with the background areas removed are obtained;
correspondingly, the image classification module 3 is specifically configured to: and classifying the tissue sections with preset sizes by adopting an image classification and identification model.
In this embodiment, the tissue slices are segmented to obtain tissue slices with smaller sizes, and in this case, the image classification and identification model is enabled to classify each small-sized tissue slice, so that compared with a mode of directly classifying the whole digital slice, the precision is higher, and the identification accuracy of the circulating tumor cells is improved.
Fig. 6 is a schematic diagram illustrating a hardware structure of a circulating tumor cell detection apparatus according to an embodiment of the present invention.
The circulating tumor cell detection apparatus may include a processor 601 and a memory 602 having stored computer program instructions. The processor 601 reads and executes the computer program instructions stored in the memory 602 to implement any one of the above-described methods for detecting circulating tumor cells.
Specifically, the processor 601 may include a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or may be configured as one or more Integrated circuits implementing embodiments of the present invention.
In one example, the circulating tumor cell detection apparatus may further include a communication interface 603 and a bus 610. As shown in fig. 6, the processor 601, the memory 602, and the communication interface 603 are connected via a bus 610 to complete communication therebetween.
The communication interface 603 is mainly used for implementing communication between modules, apparatuses, units and/or devices in the embodiments of the present invention.
In addition, in combination with the method for detecting circulating tumor cells in the above embodiments, the embodiments of the present invention can be implemented by providing a computer-readable storage medium. The computer readable storage medium having stored thereon computer program instructions; the computer program instructions, when executed by a processor, implement any of the methods of circulating tumor cell detection described in the embodiments above.
The functional blocks shown in the above-described structural block diagrams may be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, it may be, for example, an electronic circuit, an Application Specific Integrated Circuit (ASIC), suitable firmware, plug-in, function card, or the like. When implemented in software, the elements of the invention are the programs or code segments used to perform the required tasks. The program or code segments may be stored in a machine-readable medium or transmitted by a data signal carried in a carrier wave over a transmission medium or a communication link. A "machine-readable medium" may include any medium that can store or transfer information. Examples of a machine-readable medium include electronic circuits, semiconductor memory devices, ROM, flash memory, Erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, Radio Frequency (RF) links, and so forth. The code segments may be downloaded via computer networks such as the internet, intranets, etc.
It should be noted that, in this document, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising … …" does not exclude the presence of additional identical elements in the process, method, article, or apparatus that comprises the element.
It should also be noted that the exemplary embodiments mentioned in this patent describe some methods or systems based on a series of steps or devices. However, the present invention is not limited to the order of the above-described steps, that is, the steps may be performed in the order mentioned in the embodiments, may be performed in an order different from the order in the embodiments, or may be performed simultaneously.
As described above, only the specific embodiments of the present invention are provided, and it can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the system, the module and the unit described above may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again. It should be understood that the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive various equivalent modifications or substitutions within the technical scope of the present invention, and these modifications or substitutions should be covered within the scope of the present invention.
Claims (8)
1. A method for detecting circulating tumor cells, comprising:
injecting a lymph fluid sample to be detected into a comprehensive molten iron dynamic cell separation iFCS device to obtain a lymph fluid sample after leukocyte filtration;
generating a digital slice of the leukocyte-filtered lymph fluid sample;
classifying the digital slices by adopting an image classification identification model to obtain a classification result of classifying pixel points in the digital slices according to cell types to which the pixel points belong;
detecting whether the digital slices contain circulating tumor cells or not according to the classification result to obtain a detection result;
the classification result is a cell classification probability map; detecting whether the digital slice contains circulating tumor cells according to the classification result to obtain a detection result, wherein the detection result comprises the following steps:
overlapping the cell classification probability map and the digital slice to obtain a heat map;
detecting whether the heat map contains pixel points belonging to circulating tumor cells;
if yes, identifying the region of the pixel point belonging to the circulating tumor cell in the heat map to obtain the identified heat map;
the regions of the heat map belonging to pixels of circulating tumor cells include: and the region of the pixel point with the saturation exceeding a preset saturation threshold in the heat map.
2. The method of claim 1, wherein the image classification recognition model is a U-Net convolutional neural network model.
3. The method of claim 1, wherein said detecting whether the digital slice contains circulating tumor cells according to the classification result to obtain a detection result comprises:
inputting the cell classification probability map into two continuous full-connection layers;
and inputting the output of the full connecting layer into a softmax layer to obtain a judgment result of whether the digital section contains the circulating tumor cells.
4. The method of any one of claims 1-3, wherein said generating a digital slice of said leukocyte-filtered lymph sample further comprises:
segmenting the digital slice by adopting a threshold-based segmentation algorithm to obtain a tissue slice with a background region removed;
the step of classifying the digital slices by using the image classification and identification model specifically comprises the following steps:
and classifying the tissue slices with the background areas removed by adopting the image classification and identification model.
5. The method of claim 4, wherein after obtaining the tissue slice with the background region removed, further comprising:
dividing the tissue section into a plurality of tissue sections with preset sizes;
the step of classifying the digital slices by using the image classification and identification model specifically comprises the following steps:
and classifying the tissue sections with the preset size by adopting the image classification and identification model.
6. A circulating tumor cell detection apparatus, comprising:
the cell filtering module is used for injecting the lymph fluid sample to be detected into the comprehensive molten iron power cell separation iFCS device to obtain the lymph fluid sample after leukocyte filtration;
a section generating module for generating a digital section of the lymph fluid sample after filtering the white blood cells;
the image classification module is used for classifying the digital slices by adopting an image classification identification model to obtain a classification result of classifying the pixel points in the digital slices according to the cell categories to which the pixel points belong;
the detection module is used for detecting whether the digital slices contain circulating tumor cells or not according to the classification result to obtain a detection result;
the detection module comprises:
the overlapping unit is used for overlapping the cell classification probability map and the digital slice to obtain a heat map;
the detection unit is used for detecting whether the heat map contains pixel points belonging to the circulating tumor cells;
the identification unit is used for identifying the area of the pixel point belonging to the circulating tumor cell in the heat map if the heat map contains the pixel point belonging to the circulating tumor cell to obtain the identified heat map;
the regions of the heat map belonging to pixels of circulating tumor cells include: and the region of the pixel point with the saturation exceeding the preset saturation threshold value in the heat map.
7. A circulating tumor cell detection apparatus, comprising: a processor and a memory storing computer program instructions;
the processor when executing the computer program instructions implements the method of circulating tumor cell detection of any one of claims 1-5.
8. A computer readable storage medium having computer program instructions stored thereon which, when executed by a processor, implement the method of circulating tumor cell detection according to any one of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910768669.8A CN110632069B (en) | 2019-08-20 | 2019-08-20 | Circulating tumor cell detection method, device, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910768669.8A CN110632069B (en) | 2019-08-20 | 2019-08-20 | Circulating tumor cell detection method, device, equipment and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110632069A CN110632069A (en) | 2019-12-31 |
| CN110632069B true CN110632069B (en) | 2022-07-26 |
Family
ID=68970416
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910768669.8A Active CN110632069B (en) | 2019-08-20 | 2019-08-20 | Circulating tumor cell detection method, device, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110632069B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111652095A (en) * | 2020-05-21 | 2020-09-11 | 骏实生物科技(上海)有限公司 | CTC image identification method and system based on artificial intelligence |
| CN111967465A (en) * | 2020-07-07 | 2020-11-20 | 广州金域医学检验中心有限公司 | Method, system, computer device and storage medium for evaluating tumor cell content |
| CN112884725B (en) * | 2021-02-02 | 2022-12-20 | 杭州迪英加科技有限公司 | Correction method for neural network model output result for cell discrimination |
| CN114299044A (en) * | 2021-12-31 | 2022-04-08 | 深圳裕泰抗原科技有限公司 | Method and device for interpreting lymphocytes |
| CN116309543B (en) * | 2023-05-10 | 2023-08-11 | 北京航空航天大学杭州创新研究院 | Image-based circulating tumor cell detection equipment |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP7092503B2 (en) * | 2014-12-30 | 2022-06-28 | ベンタナ メディカル システムズ， インコーポレイテッド | Systems and methods for co-expression analysis |
| CN107287107A (en) * | 2016-03-30 | 2017-10-24 | 无锡纳奥生物医药有限公司 | A kind of circulating tumor cell separation equipment, system and method |
| CN106370488A (en) * | 2016-08-23 | 2017-02-01 | 宁波江丰生物信息技术有限公司 | Cycle tumor cell slice-based diagnosis system and method |
| CN107988042B (en) * | 2017-11-27 | 2021-03-09 | 长春理工大学 | Circulating tumor cell separation device and detection method thereof |
| CN110060245A (en) * | 2019-04-15 | 2019-07-26 | 青岛大学 | A kind of gastric cancer lymph node dyeing pathological image automatic identifying method based on deep neural network |
| CN110120040B (en) * | 2019-05-13 | 2021-09-14 | 广州锟元方青医疗科技有限公司 | Slice image processing method, slice image processing device, computer equipment and storage medium |
-
2019
- 2019-08-20 CN CN201910768669.8A patent/CN110632069B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN110632069A (en) | 2019-12-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110632069B (en) | Circulating tumor cell detection method, device, equipment and medium | |
| CN108961208B (en) | System and method for segmenting and counting aggregated leukocytes | |
| Mohamed et al. | An efficient technique for white blood cells nuclei automatic segmentation | |
| CN111488921B (en) | Intelligent analysis system and method for panoramic digital pathological image | |
| Madhloom et al. | An image processing application for the localization and segmentation of lymphoblast cell using peripheral blood images | |
| Dese et al. | Accurate machine-learning-based classification of leukemia from blood smear images | |
| CN110458835B (en) | Image processing method, device, equipment, system and medium | |
| Mohamed et al. | An enhanced threshold based technique for white blood cells nuclei automatic segmentation | |
| KR101682604B1 (en) | Automated cervical cancer diagnosis system | |
| CN113261012B (en) | Method, device and system for processing image | |
| Dey et al. | An automated system for segmenting platelets from microscopic images of blood cells | |
| Sholeh | White blood cell segmentation for fresh blood smear images | |
| Rachna et al. | Detection of Tuberculosis bacilli using image processing techniques | |
| Kanwal et al. | Quantifying the effect of color processing on blood and damaged tissue detection in whole slide images | |
| CN110060246B (en) | Image processing method, device and storage medium | |
| Madhloom et al. | Computer-aided acute leukemia blast cells segmentation in peripheral blood images | |
| Lal et al. | A robust method for nuclei segmentation of H&E stained histopathology images | |
| CN114387596A (en) | Automatic interpretation system for cytopathology smear | |
| Sulaiman et al. | Semi-automated pseudo colour features extraction technique for cervical cancer's pap smear images | |
| Salman et al. | A machine learning approach to identify prostate cancer areas in complex histological images | |
| Sulaiman et al. | Overlapping cells separation method for cervical cell images | |
| US20220318979A1 (en) | Image Quality Analysis for Artifact Detection in Pathology Slide Images | |
| Deore et al. | Image analysis framework for automatic extraction of the progress of an infection | |
| Danyali et al. | Robust leukocyte segmentation in blood microscopic images based on intuitionistic fuzzy divergence | |
| CN114155399A (en) | Breast pathology whole-section classification method based on multi-feature fusion progressive discrimination |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210716 Address after: Room 103-2, building a, phase II, science and technology building, 1057 Nanhai Avenue, Yanshan community, zhaoshang street, Nanshan District, Shenzhen, Guangdong 518052 Applicant after: Xi Ren Ma Da Zhou (Shenzhen) Medical Technology Co.,Ltd. Address before: 202, east side, 2 / F, building F, Zhizao street, Zhongguancun, 45 Chengfu Road, Haidian District, Beijing Applicant before: Xiren Ma Diyan (Beijing) Technology Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |