CN113345576A - Rectal cancer lymph node metastasis diagnosis method based on deep learning multi-modal CT - Google Patents
Rectal cancer lymph node metastasis diagnosis method based on deep learning multi-modal CT Download PDFInfo
- Publication number
- CN113345576A CN113345576A CN202110626711.XA CN202110626711A CN113345576A CN 113345576 A CN113345576 A CN 113345576A CN 202110626711 A CN202110626711 A CN 202110626711A CN 113345576 A CN113345576 A CN 113345576A
- Authority
- CN
- China
- Prior art keywords
- image
- lymph node
- modal
- rectal cancer
- scan
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 64
- 208000007433 Lymphatic Metastasis Diseases 0.000 title claims abstract description 52
- 208000015634 Rectal Neoplasms Diseases 0.000 title claims abstract description 48
- 206010038038 rectal cancer Diseases 0.000 title claims abstract description 48
- 201000001275 rectum cancer Diseases 0.000 title claims abstract description 48
- 238000013135 deep learning Methods 0.000 title claims abstract description 29
- 238000003745 diagnosis Methods 0.000 title claims abstract description 17
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 22
- 238000007781 pre-processing Methods 0.000 claims abstract description 7
- 206010028980 Neoplasm Diseases 0.000 claims description 47
- 239000011159 matrix material Substances 0.000 claims description 27
- 238000011176 pooling Methods 0.000 claims description 22
- 210000001165 lymph node Anatomy 0.000 claims description 21
- 201000011510 cancer Diseases 0.000 claims description 18
- 210000002751 lymph Anatomy 0.000 claims description 18
- 230000004913 activation Effects 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 12
- 206010027476 Metastases Diseases 0.000 claims description 10
- 230000009401 metastasis Effects 0.000 claims description 10
- 210000000664 rectum Anatomy 0.000 claims description 10
- 238000002591 computed tomography Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 description 7
- 238000003384 imaging method Methods 0.000 description 7
- 238000012360 testing method Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 206010009944 Colon cancer Diseases 0.000 description 3
- 208000001333 Colorectal Neoplasms Diseases 0.000 description 3
- 210000000056 organ Anatomy 0.000 description 3
- 238000001574 biopsy Methods 0.000 description 2
- 238000007796 conventional method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000009545 invasion Effects 0.000 description 2
- 230000003902 lesion Effects 0.000 description 2
- 230000001575 pathological effect Effects 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 238000004393 prognosis Methods 0.000 description 2
- 108090000623 proteins and genes Proteins 0.000 description 2
- 230000004083 survival effect Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000009098 adjuvant therapy Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000011337 individualized treatment Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000007170 pathology Effects 0.000 description 1
- 230000002980 postoperative effect Effects 0.000 description 1
- 102000004169 proteins and genes Human genes 0.000 description 1
- 238000013517 stratification Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 210000001519 tissue Anatomy 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000011282 treatment Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/02—Arrangements for diagnosis sequentially in different planes; Stereoscopic radiation diagnosis
- A61B6/03—Computed tomography [CT]
- A61B6/032—Transmission computed tomography [CT]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
- A61B6/5205—Devices using data or image processing specially adapted for radiation diagnosis involving processing of raw data to produce diagnostic data
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
- A61B6/5211—Devices using data or image processing specially adapted for radiation diagnosis involving processing of medical diagnostic data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Public Health (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Pathology (AREA)
- Theoretical Computer Science (AREA)
- High Energy & Nuclear Physics (AREA)
- Optics & Photonics (AREA)
- Veterinary Medicine (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Heart & Thoracic Surgery (AREA)
- Epidemiology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Primary Health Care (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Pulmonology (AREA)
- Apparatus For Radiation Diagnosis (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a rectal cancer lymph node metastasis diagnosis method based on deep learning multi-modal CT, which comprises the following steps: and (3) preprocessing data of the rectal cancer multi-mode CT image, and extracting image features of the cut 3D flat scan CT image and the 3D enhanced CT image with fixed size through a newly constructed Mlenet (multi-mode Lenet convolutional neural network) convolutional neural network respectively. And splicing the feature maps to form a new feature map, inputting the new feature map into a new Mlenet convolutional neural network, and performing two-classification prediction to obtain a two-classification prediction result. The method can effectively extract the characteristics of the multi-modal CT image, and greatly improve the accuracy of lymph node metastasis prediction.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a rectal cancer lymph node metastasis diagnosis method based on deep learning multi-modal CT.
Background
In recent years, lymph node metastasis is an independent risk factor for local recurrence of rectal cancer and distant metastasis, and is also an important basis for evaluating pathological stages, surgical modes and postoperative adjuvant therapy. Whether a patient with rectal cancer has lymph node metastasis has an important influence on the decision of a treatment scheme and the prognosis of the patient, so that the accurate judgment of whether the patient has lymph node metastasis is an important step of treating the rectal cancer. Therefore, the lymph node metastasis characteristic label capable of influencing survival and preoperative risk stratification of the rectal cancer patient is effectively identified, so that the individualized treatment scheme can be formulated, and the prognosis survival period of the locally advanced rectal cancer patient is improved.
The clinical methods for preoperative clinical diagnosis and TNM staging of rectal cancer include pathological, molecular and imaging methods. The spatial and temporal heterogeneity of solid tumors expressed at the gene, protein, cell, microenvironment, tissue and organ level limits the accuracy and representativeness of the results of invasive detection methods such as pathology and molecular, and the biopsy places a great burden on the body of the patient, which limits the molecular detection method based on invasive biopsy; CT is used as a preferred imaging omics examination mode of rectal cancer lymph node metastasis, and has obvious advantages in the evaluation of tumor positions, lymph node metastasis, peripheral organ invasion and the like; however, the method for diagnosing lymph node metastasis by CT usually involves the imaging physician to browse each image layer by layer and identify the shape, boundary and density of lymph node from the images for judgment, and this traditional method is time-consuming and has subjective bias, resulting in reduced accuracy and efficiency; AI medical diagnosis is widely applied to medical image recognition, disease diagnosis and other aspects at home and abroad, and AI can recognize abnormal regions in medical images, so as to provide reference for imaging physicians and improve the detection rate of lesions.
Disclosure of Invention
This section is for the purpose of summarizing some aspects of embodiments of the invention and to briefly introduce some preferred embodiments. In this section, as well as in the abstract and the title of the invention of this application, simplifications or omissions may be made to avoid obscuring the purpose of the section, the abstract and the title, and such simplifications or omissions are not intended to limit the scope of the invention.
The present invention has been made in view of the above-mentioned conventional problems.
Therefore, the technical problem solved by the invention is as follows: the method for diagnosing lymph node metastasis by CT is usually to determine by imaging physicians by browsing each image layer by layer and identifying the shape, boundary and density of lymph nodes from the images, and this traditional method is time-consuming and has subjective bias, resulting in reduced accuracy and efficiency.
In order to solve the technical problems, the invention provides the following technical scheme: carrying out CT scanning on a patient to obtain a rectal cancer lymph node CT image of the patient, and carrying out data preprocessing on the rectal cancer lymph node CT image; carrying out feature extraction on the preprocessed CT image by utilizing a novel multi-mode Lenet convolutional neural network to obtain a new feature map; inputting the extracted feature map into the novel multi-modal Lenet convolutional neural network; and training the convolutional neural network model to obtain final two-dimensional output, obtaining a two-classification prediction result of lymph node metastasis, and completing the metastasis diagnosis of the rectal lymph node.
As a preferable embodiment of the method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT of the present invention, wherein: the rectal cancer lymph node CT image comprises a rectal lymph cancer enhancement CT image and a rectal lymph cancer flat scanning CT image.
As a preferable embodiment of the method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT of the present invention, wherein: the process of preprocessing the rectum lymph cancer enhanced CT image and the rectum lymph cancer flat scanning CT image comprises extracting a tumor area data matrix and cutting a three-dimensional data matrix.
As a preferable embodiment of the method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT of the present invention, wherein: the method comprises the steps of importing a rectum lymph cancer enhanced CT image and rectum lymph cancer flat-scan CT image data acquired by CT scanning into the convolution neural network, converting the data into a three-dimensional data matrix, performing convolution on the three-dimensional data matrix of the flat-scan enhanced CT image and the flat-scan CT image by utilizing a tumor marking matrix which respectively represents a non-tumor area and a tumor area by 0 and 1, extracting a tumor area with a tumor marking matrix of q, and further extracting the tumor area data matrix.
As a preferable embodiment of the method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT of the present invention, wherein: and the step of cutting the three-dimensional data matrix comprises the step of cutting all the extracted tumor region image data in a fixed size a x b to obtain a standard enhanced CT image and standard flat-scan CT image data.
As a preferable embodiment of the method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT of the present invention, wherein: the process of extracting the features of the standard enhanced CT image and the standard flat-scan CT image data comprises the steps of utilizing a first layer of m × m convolution layers to respectively convolve the standard enhanced CT image and the standard flat-scan CT image, reducing the gradients of the standard enhanced CT image and the standard flat-scan CT image through batch regularization, utilizing a Relu activation function to carry out normalization, and mapping the standard flat-scan CT image data and the standard enhanced CT image data to the range of 0 to 1; the image passing through the convolutional layer was maximally pooled through the n × n pooling layers and subjected to one Dropout operation.
As a preferable embodiment of the method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT of the present invention, wherein: the process of feature extraction of the standard enhanced CT image and the standard flat-scan CT image data further comprises the steps of enabling the flat-scan CT image and the enhanced CT image to pass through a second layer of convolution layer respectively, reducing the gradient of the standard enhanced CT image and the gradient of the standard flat-scan CT image by batch regularization, performing normalization by using a Relu activation function, and mapping the standard flat-scan CT image data and the standard enhanced CT image data to the range of 0 to 1; and performing maximum pooling on the image passing through the second convolution layer through the n-x-n pooling layer, and performing a Dropout operation once.
As a preferable embodiment of the method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT of the present invention, wherein: merging the extracted characteristic graphs as new input, and inputting the new input into the constructed new multi-modal Lenet convolutional neural network; and inputting the multi-modal CT image data serving as a characteristic map into the novel multi-modal Lenet convolutional neural network.
As a preferable embodiment of the method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT of the present invention, wherein: training a new convolutional neural network model to obtain final two-dimensional output, wherein the process of obtaining the binary prediction result of lymph node metastasis comprises the steps of enabling the feature diagram to pass through a first layer of m × m convolutional layer, and performing maximum pooling after batch regularization and Relu activation functions are performed; and passing the feature graph which passes through the first convolution layer and the pooling layer through the second convolution layer m x m, and passing through the pooling layer again after passing through the batch regularization and Relu activation function.
As a preferable embodiment of the method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT of the present invention, wherein: the method also comprises the steps that the synthesized feature graph passes through L full-connection layers, and regularization and activation are carried out between every two connection layers; according to the condition that whether the lymph node is metastasized, two classifications are carried out, namely 0 and 1, wherein 1 represents that metastasis is determined, and 0 represents that no metastasis is determined, so that two classification predicted values are obtained.
The invention has the beneficial effects that: based on an AI deep learning theory and framework, rectal cancer CT images and clinical data, the invention provides an Mlenet convolutional neural network for processing enhancement and plain scan CT, and can effectively extract characteristics of multi-modal CT images; providing a colorectal cancer lymph node metastasis intelligent diagnosis model based on a multi-modal CT image; through processing and splicing the multi-modal CT images, the images are input into an Mlenet network, and after training, binary prediction values are output, so that the accuracy of lymph node metastasis prediction is greatly improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise. Wherein:
fig. 1 is a basic flowchart of a method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a novel Mlenet network structure constructed by a rectal cancer lymph node metastasis diagnosis method based on deep learning multi-modal CT according to an embodiment of the present invention;
fig. 3 is another schematic flow chart of a rectal cancer lymph node metastasis diagnosis method based on deep learning multi-modal CT according to an embodiment of the present invention.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, specific embodiments accompanied with figures are described in detail below, and it is apparent that the described embodiments are a part of the embodiments of the present invention, not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making creative efforts based on the embodiments of the present invention, shall fall within the protection scope of the present invention.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways than those specifically described and will be readily apparent to those of ordinary skill in the art without departing from the spirit of the present invention, and therefore the present invention is not limited to the specific embodiments disclosed below.
Furthermore, reference herein to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one implementation of the invention. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments.
The present invention will be described in detail with reference to the drawings, wherein the cross-sectional views illustrating the structure of the device are not enlarged partially in general scale for convenience of illustration, and the drawings are only exemplary and should not be construed as limiting the scope of the present invention. In addition, the three-dimensional dimensions of length, width and depth should be included in the actual fabrication.
Meanwhile, in the description of the present invention, it should be noted that the terms "upper, lower, inner and outer" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of describing the present invention and simplifying the description, but do not indicate or imply that the referred device or element must have a specific orientation, be constructed in a specific orientation and operate, and thus, cannot be construed as limiting the present invention. Furthermore, the terms first, second, or third are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
The terms "mounted, connected and connected" in the present invention are to be understood broadly, unless otherwise explicitly specified or limited, for example: can be fixedly connected, detachably connected or integrally connected; they may be mechanically, electrically, or directly connected, or indirectly connected through intervening media, or may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
Example 1
Referring to fig. 1 to 3, in an embodiment of the present invention, a method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT is provided, including:
s1, carrying out CT scanning on the patient to obtain a rectal cancer lymph node CT image of the patient, and carrying out data preprocessing on the rectal cancer lymph node CT image; it should be noted that, in the following description,
the rectal cancer lymph node CT image comprises a rectal lymph cancer enhancement CT image and a rectal lymph cancer flat scanning CT image.
The process of preprocessing the enhanced CT image of the rectal lymph cancer and the flat-scan CT image of the rectal lymph cancer comprises extracting a tumor area data matrix and cutting a three-dimensional data matrix.
Wherein extracting the tumor region data matrix comprises: importing the rectum lymph cancer enhanced CT image and the rectum lymph cancer flat-scan CT image data acquired by CT scanning into a convolution neural network, converting the data into a three-dimensional data matrix, performing convolution on the three-dimensional data matrix of the flat-scan enhanced CT image and the flat-scan CT image by utilizing a tumor marking matrix which respectively represents a non-tumor area and a tumor area by 0 and 1, extracting an area with the tumor matrix of 1, namely a tumor marking matrix q, and further extracting a tumor area data matrix.
Clipping the three-dimensional data matrix includes: and (3) cutting all the extracted tumor region image data into fixed sizes of 128-26 to obtain standard enhanced CT image data and standard flat-scan CT image data.
S2, utilizing a novel multi-mode Lenet convolutional neural network (Mlenet) to extract the characteristics of the preprocessed CT image to obtain a new characteristic diagram; it should be noted that, in the following description,
the process of feature extraction for standard enhanced CT images and standard flat-scan CT image data includes: respectively convolving the standard enhanced CT image and the standard flat-scan CT image by utilizing the first layer of 3-by-3 convolution layer, reducing the gradient of the standard enhanced CT image and the standard flat-scan CT image by batch regularization, normalizing by utilizing a Relu activation function, and mapping the data of the standard flat-scan CT image and the standard enhanced CT image to the range of 0 to 1; performing maximum pooling on the image passing through the convolution layer through a 2 x 2 pooling layer, and performing Dropout operation again to prevent overfitting;
respectively enabling the flat-scan CT image and the enhanced CT image to pass through a second layer of convolution layer, reducing the gradient of the standard enhanced CT image and the standard flat-scan CT image by batch regularization, performing normalization by using a Relu activation function, and mapping the data of the standard flat-scan CT image and the standard enhanced CT image to the range of 0-1; the image passing through the second convolutional layer was maximally pooled through 2 x 2 pooling layers and subjected to one Dropout operation.
S3, inputting the extracted characteristic diagram into a novel multi-mode Lenet convolutional neural network; it should be noted that, in the following description,
the process comprises the following steps:
splicing the flat-scan CT image and the enhanced CT image which pass through the two convolution layers and the pooling layer back and forth to form a multi-mode CT image with double thickness, namely an image matrix with the size of 48 × 6;
and inputting the multi-modal CT image data serving as a characteristic map into a novel multi-modal Lenet convolutional neural network.
S4, training the convolutional neural network model to obtain the final two-dimensional output, obtaining the two-classification prediction result of lymph node metastasis, and completing the metastasis diagnosis of the rectal lymph node; it should be noted that, in the following description,
training a new convolutional neural network model to obtain final two-dimensional output, wherein the process of obtaining the two-classification prediction result of the lymph node metastasis comprises the following steps: passing the feature graph through a first layer 3 x 3 convolution layer, and performing maximum pooling after passing through batch regularization and Relu activation functions, wherein the size of the pooling layer is 2 x 2;
and (3) passing the feature maps of the first convolution layer and the pooling layer through a second convolution layer 3 x 3, and passing through the pooling layer again after passing through batch regularization and Relu activation functions, wherein the size of the pooling layer is still 2 x 2.
The synthesized feature graph is regularized and activated between every two connection layers through three full connection layers, and a Relu function is used as an activation function; according to the condition that whether the lymph node is metastasized, two classifications are carried out, namely 0 and 1, wherein 1 represents that metastasis is determined, and 0 represents that no metastasis is determined, so that two classification predicted values are obtained.
Based on an AI deep learning theory and framework, rectal cancer CT images and clinical data, the invention provides an Mlenet convolutional neural network for processing enhancement and plain scan CT, and can effectively extract characteristics of multi-modal CT images; providing a colorectal cancer lymph node metastasis intelligent diagnosis model based on a multi-modal CT image; through processing and splicing the multi-modal CT images, the images are input into an Mlenet network, and after training, binary prediction values are output, so that the accuracy of lymph node metastasis prediction is greatly improved.
Example 2
The embodiment is another embodiment of the invention, and in order to verify and explain the technical effects adopted in the method, the embodiment adopts the traditional technical scheme and the method of the invention to carry out comparison test, and compares the test results by means of scientific demonstration to verify the real effects of the method.
The traditional technical scheme is as follows: CT scanning is used as a preferred imaging omics examination mode of rectal cancer lymph node metastasis, and has obvious advantages in the aspects of tumor position, lymph node metastasis, peripheral organ invasion and the like; however, the conventional method for diagnosing lymph node metastasis by CT usually involves the imaging physician to look through each image layer by layer to identify the shape, boundary and density of lymph nodes, which is time-consuming and subject to subjective bias, resulting in reduced accuracy and efficiency.
In order to verify that the method has higher efficiency and higher accuracy compared with the traditional method, the method for distinguishing lymph nodes by browsing images layer by layer and the method are respectively used for carrying out real-time measurement and comparison on diagnosis of lymph node metastasis of colorectal cancer.
And (3) testing environment: in a CT scanning image room, a patient multi-modal CT scanning image test sample is adopted, whether lymph nodes are transferred or not is judged by respectively utilizing manual operation and a diagnosis system of a traditional method, the rectal cancer lymph node transfer intelligent diagnosis system is started to detect the test sample by adopting the method, and the experimental result obtained according to the test data is shown in table 1.
Table 1: the experimental results of the traditional method and the method of the invention are shown in a comparison table.
Detection method | Conventional methods | The method of the invention |
Rate of accuracy | 73.2% | 93.4% |
According to the comparison between the detection accuracy of a clinician and the detection accuracy of the method, the accuracy of the traditional technical scheme is found to reach 73.2 percent in the same time period, and the accuracy of the method can reach 93.4 percent. The accuracy of the two methods is compared, and the accuracy of detecting the lymph node metastasis by the method is improved by 20.2% compared with the accuracy of detecting the lymph node metastasis by the traditional method, so that the lesion detection rate is greatly improved.
It should be noted that the above-mentioned embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, which should be covered by the claims of the present invention.
Claims (10)
1. A rectal cancer lymph node metastasis diagnosis method based on deep learning multi-modal CT is characterized by comprising the following steps:
carrying out CT scanning on a patient to obtain a rectal cancer lymph node CT image of the patient, and carrying out data preprocessing on the rectal cancer lymph node CT image;
carrying out feature extraction on the preprocessed CT image by utilizing a novel multi-mode Lenet convolutional neural network to obtain a new feature map;
inputting the extracted feature map into the novel multi-modal Lenet convolutional neural network;
and training the convolutional neural network model to obtain final two-dimensional output, obtaining a two-classification prediction result of lymph node metastasis, and completing the metastasis diagnosis of the rectal lymph node.
2. The method of claim 1 for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT, wherein: the rectal cancer lymph node CT image comprises a rectal lymph cancer enhancement CT image and a rectal lymph cancer flat scanning CT image.
3. The method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT as claimed in claim 1 or 2, wherein: the process of preprocessing the rectum lymph cancer enhanced CT image and the rectum lymph cancer flat scanning CT image comprises extracting a tumor area data matrix and cutting a three-dimensional data matrix.
4. The method of claim 3 for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT, wherein: the extracting a tumor region data matrix includes,
importing the rectum lymph cancer enhanced CT image and the rectum lymph cancer flat scanning CT image data obtained by CT scanning into the convolution neural network, and converting the data into a three-dimensional data matrix;
and (3) performing convolution on the three-dimensional data matrix of the flat-scan enhanced CT image and the flat-scan CT image by using the tumor marking matrixes which respectively represent the non-tumor area and the tumor area by 0 and 1, extracting the tumor area with the tumor marking matrix of q, and further extracting the tumor area data matrix.
5. The method of claim 3 for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT, wherein: the cropping of the three-dimensional data matrix may include,
and (3) cutting all the extracted tumor region image data in a fixed size a x b to obtain a standard enhanced CT image and standard flat-scan CT image data.
6. The method of claim 5 for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT, wherein: the process of feature extraction for the standard enhanced CT image and standard flat-scan CT image data includes,
respectively convolving the standard enhanced CT image and the standard flat-scan CT image by using a first layer of m × m convolution layers, reducing the gradients of the standard enhanced CT image and the standard flat-scan CT image through batch regularization, normalizing by using a Relu activation function, and mapping the data of the standard flat-scan CT image and the standard enhanced CT image into the range of 0 to 1;
the image passing through the convolutional layer was maximally pooled through the n × n pooling layers and subjected to one Dropout operation.
7. The method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT as claimed in any one of claims 1, 5, and 6, wherein: the process of feature extraction for the standard enhanced CT image and the standard flat-scan CT image data further comprises,
respectively enabling the flat-scan CT image and the enhanced CT image to pass through a second layer of convolutional layer, reducing the gradient of the standard enhanced CT image and the standard flat-scan CT image by batch regularization, performing normalization by using a Relu activation function, and mapping the data of the standard flat-scan CT image and the standard enhanced CT image to the range of 0-1;
and performing maximum pooling on the image passing through the second convolution layer through the n-x-n pooling layer, and performing a Dropout operation once.
8. The method of claim 7 for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT, wherein: and combining the extracted feature maps as new input, wherein the process of inputting the feature maps into the constructed new multi-modal Lenet convolutional neural network comprises the following steps,
splicing the flat-scan CT image and the enhanced CT image which pass through the two convolution layers and the pooling layer back and forth to form a multi-modal CT image with double thickness;
and inputting the multi-modal CT image data serving as a characteristic map into the novel multi-modal Lenet convolutional neural network.
9. The method for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT as claimed in claim 1 or 7, wherein: training a new convolutional neural network model to obtain final two-dimensional output, and obtaining the two-classification prediction result of lymph node metastasis,
enabling the feature graph to pass through a first layer of m × m convolution layers, and performing maximum pooling after batch regularization and Relu activation functions are performed;
and passing the feature graph which passes through the first convolution layer and the pooling layer through the second convolution layer m x m, and passing through the pooling layer again after passing through the batch regularization and Relu activation function.
10. The method of claim 9 for diagnosing lymph node metastasis of rectal cancer based on deep learning multi-modal CT, wherein: also comprises the following steps of (1) preparing,
the synthesized feature graph passes through L full-connection layers, and regularization and activation are carried out between every two connection layers;
according to the condition that whether the lymph node is metastasized, two classifications are carried out, namely 0 and 1, wherein 1 represents that metastasis is determined, and 0 represents that no metastasis is determined, so that two classification predicted values are obtained.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110626711.XA CN113345576A (en) | 2021-06-04 | 2021-06-04 | Rectal cancer lymph node metastasis diagnosis method based on deep learning multi-modal CT |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110626711.XA CN113345576A (en) | 2021-06-04 | 2021-06-04 | Rectal cancer lymph node metastasis diagnosis method based on deep learning multi-modal CT |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113345576A true CN113345576A (en) | 2021-09-03 |
Family
ID=77474429
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110626711.XA Pending CN113345576A (en) | 2021-06-04 | 2021-06-04 | Rectal cancer lymph node metastasis diagnosis method based on deep learning multi-modal CT |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113345576A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114882996A (en) * | 2022-03-17 | 2022-08-09 | 深圳大学 | Hepatocellular carcinoma CK19 and MVI prediction method based on multitask learning |
CN114974567A (en) * | 2022-05-25 | 2022-08-30 | 浙江大学 | Liver tumor interventional response post-evaluation method and system based on antagonistic generation network |
CN115132275A (en) * | 2022-05-25 | 2022-09-30 | 西北工业大学 | Method for predicting EGFR gene mutation state based on end-to-end three-dimensional convolutional neural network |
CN115424724A (en) * | 2022-11-04 | 2022-12-02 | 之江实验室 | Lung cancer lymph node metastasis auxiliary diagnosis system for multi-modal image forest |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108257134A (en) * | 2017-12-21 | 2018-07-06 | 深圳大学 | Nasopharyngeal Carcinoma Lesions automatic division method and system based on deep learning |
CN109544526A (en) * | 2018-11-15 | 2019-03-29 | 首都医科大学附属北京友谊医院 | A kind of atrophic gastritis image identification system, device and method |
CN109949318A (en) * | 2019-03-07 | 2019-06-28 | 西安电子科技大学 | Full convolutional neural networks epileptic focus dividing method based on multi-modal image |
CN110852350A (en) * | 2019-10-21 | 2020-02-28 | 北京航空航天大学 | Pulmonary nodule benign and malignant classification method and system based on multi-scale migration learning |
CN111210909A (en) * | 2020-01-13 | 2020-05-29 | 青岛大学附属医院 | Deep neural network-based rectal cancer T stage automatic diagnosis system and construction method thereof |
CN111275130A (en) * | 2020-02-18 | 2020-06-12 | 上海交通大学 | Deep learning prediction method, system, medium and device based on multiple modes |
CN112258461A (en) * | 2020-10-13 | 2021-01-22 | 江南大学 | Pulmonary nodule detection method based on convolutional neural network |
CN112309576A (en) * | 2020-09-22 | 2021-02-02 | 江南大学 | Colorectal cancer survival period prediction method based on deep learning CT (computed tomography) image omics |
CN112465058A (en) * | 2020-12-07 | 2021-03-09 | 中国计量大学 | Multi-modal medical image classification method under improved GoogLeNet neural network |
CN112489788A (en) * | 2020-11-25 | 2021-03-12 | 武汉大学中南医院 | Multi-modal image analysis method and system for cancer diagnosis |
-
2021
- 2021-06-04 CN CN202110626711.XA patent/CN113345576A/en active Pending
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108257134A (en) * | 2017-12-21 | 2018-07-06 | 深圳大学 | Nasopharyngeal Carcinoma Lesions automatic division method and system based on deep learning |
CN109544526A (en) * | 2018-11-15 | 2019-03-29 | 首都医科大学附属北京友谊医院 | A kind of atrophic gastritis image identification system, device and method |
CN109949318A (en) * | 2019-03-07 | 2019-06-28 | 西安电子科技大学 | Full convolutional neural networks epileptic focus dividing method based on multi-modal image |
CN110852350A (en) * | 2019-10-21 | 2020-02-28 | 北京航空航天大学 | Pulmonary nodule benign and malignant classification method and system based on multi-scale migration learning |
CN111210909A (en) * | 2020-01-13 | 2020-05-29 | 青岛大学附属医院 | Deep neural network-based rectal cancer T stage automatic diagnosis system and construction method thereof |
CN111275130A (en) * | 2020-02-18 | 2020-06-12 | 上海交通大学 | Deep learning prediction method, system, medium and device based on multiple modes |
CN112309576A (en) * | 2020-09-22 | 2021-02-02 | 江南大学 | Colorectal cancer survival period prediction method based on deep learning CT (computed tomography) image omics |
CN112258461A (en) * | 2020-10-13 | 2021-01-22 | 江南大学 | Pulmonary nodule detection method based on convolutional neural network |
CN112489788A (en) * | 2020-11-25 | 2021-03-12 | 武汉大学中南医院 | Multi-modal image analysis method and system for cancer diagnosis |
CN112465058A (en) * | 2020-12-07 | 2021-03-09 | 中国计量大学 | Multi-modal medical image classification method under improved GoogLeNet neural network |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114882996A (en) * | 2022-03-17 | 2022-08-09 | 深圳大学 | Hepatocellular carcinoma CK19 and MVI prediction method based on multitask learning |
CN114974567A (en) * | 2022-05-25 | 2022-08-30 | 浙江大学 | Liver tumor interventional response post-evaluation method and system based on antagonistic generation network |
CN115132275A (en) * | 2022-05-25 | 2022-09-30 | 西北工业大学 | Method for predicting EGFR gene mutation state based on end-to-end three-dimensional convolutional neural network |
CN115132275B (en) * | 2022-05-25 | 2024-02-27 | 西北工业大学 | Method for predicting EGFR gene mutation state based on end-to-end three-dimensional convolutional neural network |
CN115424724A (en) * | 2022-11-04 | 2022-12-02 | 之江实验室 | Lung cancer lymph node metastasis auxiliary diagnosis system for multi-modal image forest |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113345576A (en) | Rectal cancer lymph node metastasis diagnosis method based on deep learning multi-modal CT | |
CN111599464B (en) | Novel multi-mode fusion auxiliary diagnosis method based on rectal cancer image histology | |
CN111553892B (en) | Lung nodule segmentation calculation method, device and system based on deep learning | |
CN110728239B (en) | Gastric cancer enhanced CT image automatic identification system utilizing deep learning | |
CN112364920B (en) | Thyroid cancer pathological image classification method based on deep learning | |
CN112071418B (en) | Gastric cancer peritoneal metastasis prediction system and method based on enhanced CT image histology | |
CN111539918B (en) | Ground glass pulmonary nodule risk layered prediction system based on deep learning | |
CN110276740A (en) | Mammary tumor assists detection model, its system and mammary tumor aided detection method | |
Jiang et al. | Intratumoral analysis of digital breast tomosynthesis for predicting the Ki‐67 level in breast cancer: a multi‐center radiomics study | |
CN115953781B (en) | Mammary gland artificial intelligence analysis system and method based on thermal tomography | |
CN116825372A (en) | Breast cancer patient lymph node metastasis prediction model and construction method thereof | |
CN110738649A (en) | training method of Faster RCNN network for automatic identification of stomach cancer enhanced CT images | |
CN117152054A (en) | Breast cancer patient lymph node metastasis prediction model without incorporating clinical pathological features | |
CN115132275B (en) | Method for predicting EGFR gene mutation state based on end-to-end three-dimensional convolutional neural network | |
CN114998644B (en) | Tumor diagnosis system, construction method thereof, terminal device and storage medium | |
CN116229176A (en) | Medical image classification method and device combining image histology characteristics and clinical characteristics | |
Peddarapu et al. | Liver Tumor Risk Prediction using Ensemble Methods | |
CN115439491A (en) | Segmentation method and device for intestinal injury region of MRI (magnetic resonance imaging) image and grade evaluation system | |
CN114445374A (en) | Image feature processing method and system based on diffusion kurtosis imaging MK image | |
CN112001882A (en) | Lung cancer diagnosis method based on imaging omics small sample learning | |
CN117636064B (en) | Intelligent neuroblastoma classification system based on pathological sections of children | |
CN114155215B (en) | Nasopharyngeal carcinoma recognition and tumor segmentation method and system based on MR image | |
Kawata et al. | Hybrid classification approach of malignant and benign pulmonary nodules based on topological and histogram features | |
CN114663398A (en) | Method and system for detecting postoperative prognosis data of pancreatic cancer patient | |
CN116741390A (en) | Breast cancer patient axillary lymph node metastasis prediction model and construction method thereof |
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 |