Cancer cell inspection method and cancer cell inspection device

Abstract

To provide a cancer cell inspection method and a cancer cell inspection device that can conduct a cancer cell inspection for inspection and diagnosis of lung cancer in a simple and accurate manner.SOLUTION: A cell group to be inspected is captured using a CTC-tip, and the captured cell group to be inspected is fluorescently stained using a prescribed antibody. A blood cell image of the stained cell group to be inspected is photographed, and the feature quantity of the blood cell image is clustered. The feature quantity is to determine a minority cell feature quantity and a majority cell feature quantity, and an abnormality detector is trained by deep learning via Efficient-GAN using blood cell images sampled from healthy individuals and cancer patients. The majority cell feature quantity includes the graph morphologies of luminance histograms of the intensities of a plurality of morphological feature quantities and the occurrence frequencies thereof. The minority cell feature quantity is determined by using a minority cell detector to detect the blood cell image of the cell group to be inspected, clustering the blood cell image of the minority cells by means of a prescribed criterion, and using the probability of the appearance of each classified cluster.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a cancer cell testing method and a cancer cell testing device that utilize machine learning to screen for cancer cells from cell images and perform cancer testing. [Background technology]

[0002] Traditionally, the standard method for detecting lung cancer involves first screening with X-rays, and then those found to be positive undergo a definitive diagnosis via CT scan or biopsy. However, X-rays, the current mainstream screening method, have high specificity but low sensitivity, resulting in approximately 40% of cancer patients going undetected. Early diagnosis is an important factor in lung cancer treatment that directly affects survival rates, so there is a need for more sensitive screening to reduce the number of undetected cases.

[0003] In addition, liquid biopsy-based cancer testing methods use exosomes and circulating tumor cells (CTCs) extracted from bodily fluids such as blood and urine for various diagnostic tests, and research is also being conducted into lung cancer diagnosis due to its minimally invasive characteristics. In lung cancer, CTCs and other biopsies are used as markers for predicting prognosis and screening, but the current situation is that predictions and genetic analysis are limited to a limited number of cells.

[0004] Another cancer testing method proposed is a method for early diagnosis and post-treatment monitoring of breast cancer using multiple cancer gene biomarkers, as shown in Patent Document 1. Specifically, the method includes the steps of (a) extracting mRNA from exosomes isolated from biological samples of normal individuals and test subjects, (b) measuring the mRNA levels of adenine nucleotide translocase 2 and voltage-dependent anion-selective channel 1 genes using the extracted mRNA as a template, and (c) determining breast cancer when the mRNA levels of ANT2 and VDAC1 genes are increased compared to normal individuals.

[0005] Meanwhile, in recent years, machine learning has become increasingly popular as a component technology of AI (Artificial Intelligence). Deep learning, in particular, has been widely applied to fields such as image processing, natural language processing, and speech recognition. Furthermore, there is a type of generative AI called generative adversarial networks (GAN). GANs, which are generative models, can generate non-existent data or transform the features of existing data by learning features from data, and are a type of unsupervised learning method that learns features without providing correct answer data.

[0006] One of the suitable applications of deep learning is anomaly detection in images, etc. Supervised learning in machine learning is difficult to apply when there are countless abnormal patterns, and in such cases, unsupervised learning, which does not use labels and uses only normal examples as training data, is effective, and various methods have been proposed that use GANs as an unsupervised learning method to improve anomaly detection accuracy compared to conventional machine learning methods.

[0007] Therefore, by combining AI-based machine learning with lung cancer testing, it is expected that highly accurate lung cancer testing will be possible using not only the number of CTCs but also characteristics from various other blood cells.

[0008] A digital pathology image processing system disclosed in Patent Document 2 has been proposed as a cancer testing method using AI. Specifically, a digital pathology image of a tissue sample is received and subdivided into multiple patches. The digital pathology image of the tissue sample is a whole-slide scan image of a tumor sample from a patient diagnosed with non-small cell lung cancer (NSCLC), and the digital pathology image or the entire slide image is a hematoxylin and eosin (H&E)-stained image. This method includes, for each patch, identifying image features detected within the patch and using a machine learning model to generate one or more labels corresponding to the identified image features within the patch. The machine learning model can be a deep learning neural network. In one embodiment, the image features include a tissue image, and the one or more labels applied to the patch include cancer regions of adenocarcinoma (ADC) and squamous cell carcinoma (SCC). This method also determines a heterogeneity metric for the tissue sample based on the generated labels. The heterogeneity metric is used to evaluate the degree of heterogeneity of the identified image features and corresponding labels in the tissue sample. [Prior art documents] [Patent documents]

[0009] [Patent Document 1] Patent No. 7187081 [Patent Document 2] Special Publication No. 2023-534448 [Non-patent literature]

[0010] [Non-Patent Document 1] Hyunku Shin, Seunghyun Oh, Soonwoo Hong, Minsung Kang, DaehyeonKang, Yong-gu Ji, Byeong Hyeon Choi, Ka-Won Kang, Hyesun Jeong, Yong Park,Sunghoi Hong, Hyun Koo Kim, and Yeonho Choi “Early-Stage Lung Cancer Diagnosisby Deep Learning-Based Spectroscopic Analysis of Circulating Exosomes” Nano2020 14 (5), 5435-5444 Summary of the Invention [Problem to be solved by the invention]

[0011] The diagnostic method disclosed in Patent Document 1 involves measuring mRNA levels using gene profiling to diagnose breast cancer in order to reduce the number of undetected cases through more sensitive screening, but this requires specialized equipment, involves many steps in the testing and diagnosis, and is expensive to diagnose.

[0012] The method disclosed in the non-patent document diagnoses lung cancer from cells in the blood, but the target is circulating exosomes, and the method for extracting features from circulating exosomes is spectral analysis using surface-enhanced Raman scattering, which costs several tens of thousands of yen per test.

[0013] The AI-based cancer screening method disclosed in Patent Document 2 enables more accurate screening by training a machine learning model (e.g., a deep learning neural network) to identify image features and generate labels corresponding to the identified image features shown in multiple patches from digital pathology images. However, the machine learning model is trained based on pre-labeled or pre-annotated tumor regions in a set of digital pathology images by human experts. For example, tumor regions are manually selected by pathologists, physicians, clinical specialists, lung cancer diagnosis experts, etc. Therefore, there is a need for a system that enables easier and more accurate diagnosis.

[0014] This invention has been made in consideration of the above-mentioned background art, and aims to provide a cancer cell testing method and a cancer cell testing device that can perform cancer cell testing more easily and with higher accuracy in the testing and diagnosis of lung cancer, and can accurately screen cancer cells. [Means for solving the problem]

[0015] This invention uses deep abnormality detection technology to detect small numbers of cells contained in blood cell images in order to obtain information useful for testing. This cancer cell testing method uses machine learning to perform lung cancer testing and other tests using features obtained from the small number of cells and features obtained from other large numbers of cells, thereby obtaining information on cells that appear only in cancer patients and enabling highly accurate predictions.

[0016] (Claim 1) This invention is a cancer cell testing method for testing the possibility of cancer cells using blood cell images taken from collected blood, A group of cells to be examined is captured using a CTC-tip, the captured group of cells to be examined is fluorescently stained with a predetermined antibody, and an image of the stained blood cells of the group of cells to be examined is taken; Thereafter, clustering is performed in machine learning processing on the feature amounts of the blood cell image, and the feature amounts are obtained as a minority cell feature amount and a majority cell feature amount, In advance, we used deep learning to train a GAN anomaly detector using blood cell images sampled from healthy individuals and cancer patients. forming a plurality of brightness histograms including graph forms of brightness histograms of brightness and occurrence frequency of a plurality of morphological features of the blood cell image for the plurality of cell features in the blood cell image, and identifying the plurality of cell features using a predetermined threshold value of the plurality of cell features; detecting minority cells in the blood cell image of the test subject cell group using a minority cell detector, clustering the blood cell image of the minority cells according to a predetermined criterion, determining the minority cell feature amount using the appearance probability of each classified cluster, removing white blood cell images and other noise images from the blood cell image, and identifying the minority cell feature amount for the test subject cell group using a predetermined minority cell feature amount threshold; This is a cancer cell testing method in which the group of test cells that are determined to be abnormal based on the identified majority cell feature and minority cell feature thresholds are deemed to be cancer cells.

[0017] (Claim 2) The antibodies used for the fluorescent staining are three types: Hoechst, which is expressed in the DNA of cell nuclei; CD45, which is expressed in white blood cells; and CK, which is expressed in epithelial cells.

[0018] (Claim 3) The multiple cell features consist of multiple types of average brightness, height of the brightness histogram, width of the brightness histogram, median of the brightness histogram, skewness of the brightness histogram, and kurtosis of the brightness histogram for Hoechst, CD45, and CK in the blood cell image.

[0019] (Claim 4) The majority cell feature threshold is set based on the difference in the occurrence frequency or the brightness value of the brightness histogram for the majority cell feature between a healthy subject and a cancer patient.

[0020] (Claim 5) The removal of the white blood cell images appearing in the blood cell image using the small number of cells detector removes white blood cells appearing singly in the blood cell image, and further removes white blood cells appearing multiple times in the blood cell image.

[0021] (Claim 6) The small number of cell features are clustered, and the trained anomaly detector determines whether a cell is likely to be cancerous based on the differences in the small number of cell features between healthy individuals and cancer patients.

[0022] (Claim 7) In the first step of the removal of white blood cells, a cell count threshold is determined based on the abnormal values and the number of cells in the image, and the cell count is determined by binarizing the image and counting the number of cellular components detected; The cell number threshold is set so that cells that appear alone in the image and have a low abnormal value are considered normal; Next, in order to remove images in which multiple cells appear within images that have been determined to be normal in the first stage, an abnormal value threshold for detecting cells that are specifically emitting light is determined, and the area of the cell at the center of the image is used as a criterion for determining the abnormal value threshold, and the abnormal value and area of each cell are calculated. When the abnormal values are sorted in ascending order, the abnormal value of the cell, which is the average value of the cell area, is used as a second threshold to remove the images of white blood cells, Clustering is performed on the test cell group that has abnormal values above the second threshold, and cancer cells are identified from the clusters obtained by this process using the occurrence probability compared to that of normal individuals as a feature.

[0023] (Claim 8) This invention also provides a cancer cell testing device that tests for the possibility of cancer cells using blood cell images taken from collected blood, a blood cell image capturing device that captures a test subject cell group using a CTC-tip, fluorescently stains the captured test subject cell group with a predetermined antibody, and captures a blood cell image of the stained test subject cell group; clustering is performed on the feature amounts of the blood cell image, the feature amounts comprising a feature amount extraction device for obtaining a few-cell feature amount and a majority-cell feature amount; The feature extraction device has previously trained an anomaly detector using GAN through deep learning using blood cell images sampled from healthy individuals and cancer patients, the feature extraction device comprises a processing device that forms a plurality of brightness histograms for the majority cell features in the blood cell image, the brightness histograms including graph forms of brightness histograms of brightness in a plurality of morphological features of the blood cell image and their occurrence frequencies, identifies the majority cell features using a predetermined majority cell feature threshold, detects minority cells in the blood cell image of the test subject cell group using a minority cell detector, clusters the blood cell image of the minority cells according to a predetermined criterion, obtains the minority cell features using the appearance probability of each classified cluster, removes white blood cell images and other noise images from the blood cell image, and identifies the minority cell features for the test subject cell group using the predetermined minority cell feature threshold, This is a cancer cell testing device that makes it possible to identify whether or not a group of test cells that have been determined to be abnormal based on the majority cell feature threshold and the minority cell feature threshold are cancer cells, using the identified majority cell feature and minority cell feature. [Effects of the Invention]

[0024] The cancer cell testing method using the blood imaging testing system of this invention enables more sensitive testing than the conventional X-ray testing method, and can test with only about 1 ml of blood, making lung cancer testing simple and inexpensive. Furthermore, when used in combination with the conventional method, the cancer cell testing method of this invention has been shown to reduce the number of undetected cases by 74.9%, making it possible to test with extremely high accuracy. [Brief explanation of the drawings]

[0025] [Figure 1] 1 is a flowchart showing a cancer cell testing method according to an embodiment of the present invention. [Figure 2] FIG. 2 is a diagram showing the order of capturing blood cell images in the cancer cell testing method according to one embodiment of the present invention. [Figure 3]1A and 1B are diagrams showing fluorescent stained images and antibodies used in a cancer cell testing method according to one embodiment of the present invention. [Figure 4] FIG. 1 is a diagram showing a phase contrast image without staining used in a cancer cell testing method according to an embodiment of the present invention. [Figure 5] FIG. 1 is a schematic diagram illustrating detection of lung cancer cells (lung cancer) and the like using AI based on minority cell features and majority cell features from a fluorescent stained image and a phase contrast image in a cancer cell testing method according to one embodiment of the present invention. [Figure 6] FIG. 10 is a diagram showing histograms of the fluorescence intensity and occurrence frequency of each fluorescence channel and six types of feature amounts that are morphological feature amounts obtained from the three types of histograms in a cancer cell testing method according to one embodiment of the present invention. [Figure 7] FIG. 10 is a diagram showing the detection of 21 types of cell features, which are morphological features obtained from the histograms of each fluorescence channel, cell regions, and the number of CTCs, in a cancer cell testing method according to one embodiment of the present invention. [Figure 8] 1 is a schematic diagram showing detection of a minority cell feature in a cancer cell testing method according to an embodiment of the present invention. FIG. [Figure 9] FIG. 1 is a conceptual diagram showing the detection of a small number of cells by an AI abnormality detector and the removal of a large number of cells in a cancer cell testing method according to one embodiment of the present invention. [Figure 10] FIG. 1 is a conceptual diagram showing the process of performing two-stage abnormality detection using an AI anomaly detector in a cancer cell testing method according to one embodiment of the present invention. [Figure 11] FIG. 1 is a conceptual diagram showing feature extraction using deep learning by AI in a cancer cell testing method according to one embodiment of the present invention. [Figure 12] 1A shows a cell image of a single white blood cell, which is a majority cell that is removed in the first stage by an abnormality detector in a cancer cell testing method according to one embodiment of the present invention, and FIG. 1B shows a cell image of multiple white blood cells. [Figure 13] 1A and 1B are graphs and conceptual diagrams showing the first stage of threshold processing by an AI anomaly detector in a cancer cell testing method according to one embodiment of the present invention. [Figure 14]10 is a graph showing an example of determining a threshold value for the first stage of threshold processing by an AI anomaly detector in a cancer cell testing method according to one embodiment of the present invention. [Figure 15] 10 is a graph showing an example of determining a threshold value for the second stage of threshold processing by an AI anomaly detector in a cancer cell testing method according to one embodiment of the present invention. [Figure 16] This is an image showing an example of a cluster obtained by performing k-means clustering in the second stage threshold processing by an AI anomaly detector in a cancer cell testing method according to one embodiment of the present invention. [Figure 17] 1 is a graph showing the reduction rate of undetected cancer patients when the cancer cell testing method of the present invention is used in combination with X-ray radiography in an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0026] A cancer cell testing method according to one embodiment of the present invention uses deep learning based on AI (artificial intelligence) to detect cancer cells, and is comprised of four steps: blood sampling (S1), taking an image of blood cells (S2), extracting the image of blood cells (S3), and extracting features from the image of blood cells (S4), as shown in the flowchart of Fig. 1. In this embodiment, the detection of lung cancer cells will be described as an example.

[0027] The blood collection (S1) can be performed using a small amount of blood collected during a health checkup or other such test, and surplus blood from other tests that require blood collection can also be used.

[0028] In the blood cell imaging (S2), as shown in Figure 2, first, screening is performed using a CTC-tip to capture circulating tumor cells (CTCs) to select blood cells significant for lung cancer (S21). Next, fluorescent staining is performed using the three antibodies shown in Table 1 (S22), and then blood cell images are captured using a fluorescence microscope (S23). [Table 1]

[0029] The fluorescent staining using three types of antibodies involves three colors: Hoechst, which is expressed in the DNA of cell nuclei, CD45, which is expressed in white blood cells, and CK, which is expressed in epithelial cells, as shown in Figures 2 and 3. In this embodiment, a total of four channels are used as blood cell images: three-color stained images obtained by staining using these antibodies, and an unstained phase-contrast image as shown in Figure 4.

[0030] In the step of extracting blood cell images (S3), cells are detected and extracted using a region binarization method on the Hoechst channel, which is a fluorescent light that is specifically expressed in the DNA of cell nuclei. Cell extraction is performed by identifying and extracting cells based on the binarized image through computer processing.

[0031] Feature extraction (S4) for lung cancer testing from blood cell images is performed using two methods (S41, S42) as shown in Figures 1 and 5. Feature extraction (S4) is performed by processing within a computer using AI, which is a feature extraction device, through the processing steps described below.

[0032] The first method focuses on the morphological characteristics of histograms of feature quantities for images of a large number of cells in a cell group under examination, and uses 21 types of feature quantities as shown in Table 2. This is calculated using the cell area and fluorescence intensity histograms for each fluorescence channel and the number of CTCs, as shown in Figures 6 and 7. For feature quantities based on the number of CTCs, a CK intensity histogram is used, as shown in Figure 7, without using a CTC detector, to detect cells with CK intensity values 5% or more to the right of the horizontal axis (intensity values) as CTCs, and the number of these is treated as a feature quantity. [Table 2]

[0033] As shown in Figure 6, the multiple cell feature utilizes a brightness histogram of the brightness (horizontal axis) of multiple morphological features of blood cell images and the frequency of occurrence (vertical axis) of that brightness value when the overall amount is set to 1, and uses (1) the height of the brightness histogram, (2) the width (variance) of the brightness histogram, (3) the median (mode) of the brightness histogram, (4) the skewness of the brightness histogram, (5) the kurtosis of the brightness histogram, and (6) the average brightness of the brightness histogram for each fluorescence channel. As shown in Figure 6, each histogram distinguishes multiple cell features by setting a predetermined threshold for the multiple cell features that appear as differences in histogram morphology.

[0034] By distinguishing between the multiple cell features using each histogram, a predetermined multiple cell feature threshold is set based on the morphological differences in the brightness histograms between healthy subjects and cancer patients for the multiple cell features, and healthy subjects and cancer patients are distinguished from each other.

[0035] The second method uses deep learning AI to detect anomalies, extracting and characterizing minority cells. First, 20,000 randomly sampled images from healthy individuals and cancer patients are used to train an anomaly detector using Efficient-GAN (Generative Adversarial Networks), as shown in Figures 8 to 11. To extract minority cells, a minority cell detector is used to detect minority cells, which are then clustered. Minority cell features are created using the probability of occurrence of each cluster. These clusters are divided into, for example, 160, and there are 160 types of minority cell features, the same as the number of clusters. Efficient-GAN uses an anomaly detection model trained to output an image identical to the input image. By training this model, when cells that are abundant in a dataset, such as white blood cells, are input, the error between the input image and the output image is small. However, when other minority cells are input, the error becomes large. Based on the calculated anomaly value A, abnormal images are detected, and the majority cells are removed to detect minority cells. As shown in Figures 9 and 11(a), abnormal images are detected by calculating the root mean square error (RMSE) for the input image Iin of the Encoder (E) and the output image Iout of the Discriminator (D).

[0036] Furthermore, the detection of small numbers of cells is performed in two stages using an AI-based small number of cells detector, as shown in Figure 10. The first stage detects white blood cells that appear alone in the image, and the second stage detects white blood cells that appear multiple times in the image.

[0037] In the first stage of detection using the small number of cells detector, a threshold is set based on the abnormal value and the number of cells in the image, as shown in Figure 12. The number of cells is calculated by binarizing the image and counting the number of detected cells (connected components). In this process, a threshold is set so that cells that appear alone in the image and have a low abnormal value are considered normal, and white blood cells, which make up the majority of blood cells, are mainly removed. Images containing multiple white blood cells or cells that emit specific light are then detected as abnormal cells and used for training the second stage of the anomaly detector.

[0038] In the first stage of processing, the normal majority cell has one cell and the error A between the reconstructed image and the input image is M1 (The horizontal axis of Fig. 14) is the threshold T M1 It is determined as follows. M1 The threshold T M1 As shown in Figure 14, M1 The most frequent value when the histogram was drawn is tripled.

[0039] In the second stage of processing, in order to remove images that contain multiple cells that were deemed normal in the first stage, these are deemed normal and an abnormal value threshold is determined to detect cells that emit specific light. The area of the cell at the center of the image is used as the criterion for determining the threshold for abnormal values. When the abnormal values are sorted in ascending order, the abnormal value where the area of the central cell is equal to or greater than the average value is used as the threshold. This makes it possible to remove images in which multiple cells are sparsely captured and detect the remaining small number of cells.

[0040] The threshold for the majority of cells in the second stage is the abnormal value A calculated by the second stage anomaly detector. M2 The threshold is calculated as shown in Figure 15, where the vertical axis is the area of the cell and the horizontal axis is the abnormal value A of every 1000 blood cell images. M2 The graph is sorted in ascending order of the average values of the cell area (horizontal axis of Figure 15). The A M2 is the threshold T M2 Then, T M2 Cells that are:

[0041] As shown in Figure 11(b), the detected abnormal cell groups are clustered using k-means, and the resulting cluster occurrence probability is used as a feature. By setting features such as a rare cell that appears specifically, as in the example of a cluster (Figure 16), it becomes possible to use the characteristics of cells that appear extremely rarely in a specimen for lung cancer screening. Furthermore, cancer patient screening can be performed by classifying the extracted features using, for example, Random Forest.

[0042] According to the cancer cell testing method using the blood image testing system of the present invention, based on fluorescently stained cell images, a majority cell feature is identified as multiple morphological features of the blood cell image based on a brightness histogram of the blood cell image's brightness and its occurrence frequency. For minority cell features, the blood cell images are clustered according to predetermined criteria, and minority cell features are determined using the occurrence probability of each classified cluster. Images of white blood cells and other noise are removed from the blood cell images, and minority cell features of the test target cell group are identified using a predetermined minority cell feature threshold. This simple process enables more sensitive testing than conventional X-ray testing. For example, testing is possible with only about 1 ml of blood, enabling simple, inexpensive, and highly accurate cancer testing, such as lung cancer testing. Furthermore, the steps of extracting cell images and determining whether or not a cell is cancerous are automatically calculated and executed by a predetermined program that executes the above processes, enabling extremely rapid and accurate screening and diagnosis of cancer cells.

[0043] The multiple cell feature can be selected from multiple types of brightness histogram height, brightness histogram width, brightness histogram median, brightness histogram skewness, brightness histogram kurtosis, and brightness histogram mean brightness for Hoechst, CD45, and CK in blood cell images, enabling more accurate understanding and weekly classification of multiple cell features, and enabling accurate judgments with a simple processing process.

[0044] In addition, the system clusters the features of a small number of cells and uses AI to detect cancer cells, enabling simple and accurate judgments.Furthermore, a small number of cells detector is used to detect small numbers of cells in blood cell images of the cell group being examined, and white blood cell images that appear in the blood cell images are removed.By removing white blood cells that appear alone in the blood cell images and also removing white blood cells that appear multiple times in the blood cell images, cancer cells can be accurately identified from the features of a small number of cells.

[0045] The cancer cell testing method of the present invention can be applied to cancer cells other than lung cancer cells in the above embodiment, and by using fluorescent stained images of cancer cells, it can be used for screening and cancer testing of various cancer cells. [Example]

[0046] Accuracy verification in this study was conducted based on images taken of 26 healthy individuals and 14 cancer patients. The leave-one-out method was used. The accuracy achieved was a sensitivity of 85.7% and a specificity of 73.1%. By combining this with low-sensitivity X-ray testing (59.4%), the undetected rate was reduced by 74.9%, as shown in Figure 17. In other words, while X-ray testing would have missed approximately 40% of cancer patients, the cancer cell testing method of the present invention reduced this 40% miss by 74.9%, reducing the rate to approximately 10%. Sensitivity, undetected rate, reduction in the number of undetected cases, and specificity are expressed by the following formulas. Sensitivity = (number of correctly identified "cancer patients") ÷ (number of "cancer patients" tested) × 100 [%] Undetected rate = 100 - sensitivity [%] Reduction rate of undetected cases = ((Undetected rate with "X-ray examination only") - (Undetected rate with "X-ray examination and blood cell imaging examination combined")) ÷ (Undetected rate with "X-ray examination only") x 100 [%] Specificity = (number of correctly identified "healthy individuals") ÷ (number of "healthy individuals" tested) × 100 [%] [Industrial Applicability]

[0047] The cancer cell testing method using the blood image testing system of this invention can be performed using a small amount of blood, and can easily detect cells suspected of being lung cancer, etc., making it suitable for use in computer-based medical systems. Furthermore, since it can easily test for lung cancer cells, it can also be applied to regular health checkups.

Claims

1. A cancer cell testing method for testing the possibility of cancer cells using blood cell images obtained by photographing collected blood, comprising: A group of cells to be examined is captured using a CTC-tip, the captured group of cells to be examined is fluorescently stained with a predetermined antibody, and an image of the stained blood cells of the group of cells to be examined is taken; Thereafter, clustering is performed in machine learning processing on the feature amounts of the blood cell image, and the feature amounts are obtained as a minority cell feature amount and a majority cell feature amount, In advance, we used deep learning to train a GAN anomaly detector using blood cell images sampled from healthy individuals and cancer patients. forming a plurality of brightness histograms including graph forms of brightness histograms of brightness and occurrence frequency of a plurality of morphological features of the blood cell image for the plurality of cell features in the blood cell image, and identifying the plurality of cell features using a predetermined threshold value of the plurality of cell features; detecting minority cells in the blood cell image of the test subject cell group using a minority cell detector, clustering the blood cell image of the minority cells according to a predetermined criterion, determining the minority cell feature amount using the appearance probability of each classified cluster, removing white blood cell images and other noise images from the blood cell image, and identifying the minority cell feature amount for the test subject cell group using a predetermined minority cell feature amount threshold; A cancer cell testing method in which the test subject cell group that is determined to be abnormal based on the identified majority cell feature and minority cell feature threshold is deemed to be a cancer cell.

2. 2. The cancer cell testing method according to claim 1, wherein the antibodies used for the fluorescent staining are three types: Hoechst, which is expressed in DNA in cell nuclei; CD45, which is expressed in leukocytes; and CK, which is expressed in epithelial cells.

3. 3. The cancer cell testing method according to claim 2, wherein the multiple cell feature comprises a plurality of types of average brightness, height of the brightness histogram, width of the brightness histogram, median of the brightness histogram, skewness of the brightness histogram, and kurtosis of the brightness histogram for the Hoechst, CD45, and CK in the blood cell image.

4. The cancer cell testing method according to claim 3 , wherein the majority cell feature threshold is set based on the difference in the occurrence frequency or the brightness value of the brightness histogram for the majority cell feature between a healthy subject and a cancer patient.

5. The cancer cell testing method according to claim 1, wherein the removal of the white blood cell images appearing in the blood cell image using the small number of cells detector removes white blood cells appearing singly in the blood cell image, and further removes white blood cells appearing multiple times in the blood cell image.

6. The cancer cell testing method according to claim 1, wherein the small number of cell features are clustered, and the trained anomaly detector determines whether a cell is likely to be cancerous based on differences in the small number of cell features between healthy individuals and cancer patients.

7. In the first step of the leukocyte removal, a cell count threshold is determined based on the abnormal values and the number of cells in the image, and the cell count is determined by binarizing the image and counting the number of cellular components detected. The cell number threshold is set so that cells that appear alone in the image and have a low abnormal value are considered normal; Next, in order to remove images in which multiple cells appear within images that have been determined to be normal in the first stage, an abnormal value threshold for detecting cells that are specifically emitting light is determined, and the area of the cells in the image is used as a criterion for determining the abnormal value threshold, and the abnormal value and area of each cell are obtained. When the abnormal values are sorted in ascending order, the abnormal value of the cells, which is the average value of the cell area, is used as a second threshold to remove the images of white blood cells, A cancer cell testing method as described in claim 4, wherein clustering is performed on the test cell group having abnormal values above the second threshold, and the resulting clusters are used to identify cancer cells using the occurrence probability compared to that of normal individuals as a feature.

8. A cancer cell testing device that uses blood cell images taken from collected blood to test for the possibility of cancer cells. a blood cell image capturing device that captures a test subject cell group using a CTC-tip, fluorescently stains the captured test subject cell group with a predetermined antibody, and captures a blood cell image of the stained test subject cell group; clustering is performed on the feature amounts of the blood cell image, the feature amounts comprising a feature amount extraction device for obtaining a few-cell feature amount and a majority-cell feature amount; The feature extraction device has previously trained an anomaly detector using GAN through deep learning using blood cell images sampled from healthy individuals and cancer patients, the feature extraction device comprises a processing device that forms a plurality of brightness histograms for the majority cell features in the blood cell image, the brightness histograms including graph forms of brightness histograms of brightness in a plurality of morphological features of the blood cell image and their occurrence frequencies, identifies the majority cell features using a predetermined majority cell feature threshold, detects minority cells in the blood cell image of the test subject cell group using a minority cell detector, clusters the blood cell image of the minority cells according to a predetermined criterion, obtains the minority cell features using the appearance probability of each classified cluster, removes white blood cell images and other noise images from the blood cell image, and identifies the minority cell features for the test subject cell group using the predetermined minority cell feature threshold, A cancer cell testing device that can identify whether or not a group of test cells that have been determined to be abnormal based on the majority cell feature threshold and the minority cell feature threshold are cancer cells for the identified majority cell feature and minority cell feature.

Images