CN117694839A - Image-based prediction method and system for recurrence rate of non-myogenic invasive bladder cancer - Google Patents
Image-based prediction method and system for recurrence rate of non-myogenic invasive bladder cancer Download PDFInfo
- Publication number
- CN117694839A CN117694839A CN202410160502.4A CN202410160502A CN117694839A CN 117694839 A CN117694839 A CN 117694839A CN 202410160502 A CN202410160502 A CN 202410160502A CN 117694839 A CN117694839 A CN 117694839A
- Authority
- CN
- China
- Prior art keywords
- tumor
- recurrence rate
- bladder cancer
- patient
- degree
- 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.)
- Granted
Links
- 206010005003 Bladder cancer Diseases 0.000 title claims abstract description 141
- 230000001114 myogenic effect Effects 0.000 title claims abstract description 61
- 238000000034 method Methods 0.000 title claims abstract description 27
- 206010028980 Neoplasm Diseases 0.000 claims abstract description 340
- 210000003205 muscle Anatomy 0.000 claims abstract description 148
- 230000000694 effects Effects 0.000 claims abstract description 31
- 238000003062 neural network model Methods 0.000 claims abstract description 30
- 238000012545 processing Methods 0.000 claims abstract description 30
- 206010064390 Tumour invasion Diseases 0.000 claims abstract description 27
- 230000009400 cancer invasion Effects 0.000 claims abstract description 27
- 230000036541 health Effects 0.000 claims abstract description 27
- 238000009792 diffusion process Methods 0.000 claims abstract description 20
- 238000001356 surgical procedure Methods 0.000 claims abstract description 14
- 238000011282 treatment Methods 0.000 claims description 29
- 210000002700 urine Anatomy 0.000 claims description 26
- 210000001519 tissue Anatomy 0.000 claims description 24
- 230000036770 blood supply Effects 0.000 claims description 23
- 210000004877 mucosa Anatomy 0.000 claims description 18
- 208000007097 Urinary Bladder Neoplasms Diseases 0.000 claims description 16
- 239000000758 substrate Substances 0.000 claims description 16
- 238000013528 artificial neural network Methods 0.000 claims description 13
- 201000005112 urinary bladder cancer Diseases 0.000 claims description 12
- 238000000608 laser ablation Methods 0.000 claims description 4
- 238000003384 imaging method Methods 0.000 claims description 3
- 238000002679 ablation Methods 0.000 claims 2
- 238000013527 convolutional neural network Methods 0.000 description 14
- 210000004027 cell Anatomy 0.000 description 12
- 230000008595 infiltration Effects 0.000 description 9
- 238000001764 infiltration Methods 0.000 description 9
- 201000011510 cancer Diseases 0.000 description 8
- 239000013598 vector Substances 0.000 description 8
- 230000011218 segmentation Effects 0.000 description 7
- 230000002159 abnormal effect Effects 0.000 description 5
- 210000004881 tumor cell Anatomy 0.000 description 5
- 238000000605 extraction Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000035755 proliferation Effects 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 206010027476 Metastases Diseases 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 230000003387 muscular Effects 0.000 description 3
- 238000011176 pooling Methods 0.000 description 3
- 238000000547 structure data Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000002380 cytological effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000012010 growth Effects 0.000 description 2
- 230000009545 invasion Effects 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- 230000003902 lesion Effects 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 230000036210 malignancy Effects 0.000 description 2
- 230000009401 metastasis Effects 0.000 description 2
- 230000000877 morphologic effect Effects 0.000 description 2
- 210000000754 myometrium Anatomy 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 210000000056 organ Anatomy 0.000 description 2
- 230000001225 therapeutic effect Effects 0.000 description 2
- 230000004614 tumor growth Effects 0.000 description 2
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 1
- 230000002146 bilateral effect Effects 0.000 description 1
- 238000001574 biopsy Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 210000004204 blood vessel Anatomy 0.000 description 1
- 230000004663 cell proliferation Effects 0.000 description 1
- 238000005056 compaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000002574 cystoscopy Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000004064 dysfunction Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 230000008821 health effect Effects 0.000 description 1
- 208000006750 hematuria Diseases 0.000 description 1
- 208000028867 ischemia Diseases 0.000 description 1
- 238000002595 magnetic resonance imaging Methods 0.000 description 1
- 230000003211 malignant effect Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000035764 nutrition Effects 0.000 description 1
- 235000016709 nutrition Nutrition 0.000 description 1
- 229910052760 oxygen Inorganic materials 0.000 description 1
- 239000001301 oxygen Substances 0.000 description 1
- 230000010412 perfusion Effects 0.000 description 1
- 230000002980 postoperative effect Effects 0.000 description 1
- 238000011471 prostatectomy Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000035882 stress Effects 0.000 description 1
- 238000011269 treatment regimen Methods 0.000 description 1
- 210000000626 ureter Anatomy 0.000 description 1
- 230000002792 vascular Effects 0.000 description 1
- 230000009790 vascular invasion Effects 0.000 description 1
- 235000013311 vegetables Nutrition 0.000 description 1
Abstract
The invention provides a method and a system for predicting recurrence rate of non-myogenic invasive bladder cancer based on images, which relate to the technical field of non-myogenic invasive bladder cancer, and the method comprises the steps of judging whether a patient is the non-myogenic invasive bladder cancer or not by using a myogenic invasive judgment model; processing two nodes and one side between the two nodes based on the graph neural network model to determine an operation mode, a tumor invasion degree, a tumor diffusion degree, a tumor activity degree and a tumor health influence degree; determining a first recurrence rate of non-myogenic invasive bladder cancer based on the surgical procedure, the degree of tumor invasion, the degree of tumor diffusion, the degree of tumor activity, the degree of tumor health impact; determining a second recurrence rate of the non-myogenic invasive bladder cancer using a second recurrence rate determination model; the recurrence rate is determined based on the first recurrence rate and the second recurrence rate of the non-myogenic invasive bladder cancer. The method can accurately determine the recurrence rate of non-muscle invasive bladder cancer.
Description
Technical Field
The invention relates to the technical field of non-muscle invasive bladder cancer, in particular to an image-based non-muscle invasive bladder cancer recurrence rate prediction method and system.
Background
Non-muscle invasive bladder cancer is a common type of bladder cancer, which refers to a tumor in which cells are localized to only the mucosal or submucosal layer, but not yet invading the muscle layer of the bladder.
Non-myogenic invasive bladder cancer may manifest itself as a lump in the bladder, a vegetable pattern tumor, and the patient may develop painless macroscopic hematuria. Diagnosis usually requires corroboration by cystoscopy and tissue biopsy. The treatment method comprises the steps of surgical excision of tumors, bladder perfusion treatment and the like, and the accurate prediction of the recurrence rate of non-myogenic invasive bladder cancer is of great significance in guiding subsequent treatment and monitoring. At present, the recurrence rate prediction of non-muscle invasive bladder cancer generally depends on experience and clinical examination of doctors, and the method has the problems of strong subjectivity, low accuracy and the like, and is easy to be influenced by subjective factors, so that the prediction result is relatively uncertain.
Therefore, how to accurately determine the recurrence rate of non-myogenic invasive bladder cancer is a current urgent problem.
Disclosure of Invention
The invention mainly solves the technical problem how to accurately determine the recurrence rate of non-muscle invasive bladder cancer.
According to a first aspect, the present invention provides an image-based method for predicting recurrence rate of non-muscle invasive bladder cancer, comprising: acquiring MR images of the bladder of the patient prior to treatment; judging whether the patient is non-myogenic invasive bladder cancer or not by using a myogenic invasive judgment model based on the bladder MR image of the patient before treatment; if the patient is judged to belong to non-muscle invasive bladder cancer, determining tumor information, a bladder wall muscle layer position, a direction and a distance between a tumor and the bladder wall muscle layer by using an image processing model based on a bladder MR image of the patient before treatment, wherein the tumor information comprises a tumor size, a tumor position, a tumor morphology, a basal size of the tumor attached to bladder mucosa, a tumor tissue density, a tumor edge characteristic and a tumor blood supply condition; constructing two nodes and an edge between the two nodes based on the tumor information and the bladder wall muscle layer position, wherein the two nodes comprise a tumor node and a bladder wall muscle layer node, each node of the two nodes comprises a plurality of node characteristics, the node characteristics of the tumor node are tumor size, tumor position, tumor morphology, the size of a substrate of a tumor attached to bladder mucosa, tumor tissue density, tumor edge characteristics and tumor blood supply conditions, the node characteristics of the bladder wall muscle layer node are bladder wall muscle layer position, and the characteristics of the edge between the two nodes comprise the direction and distance between the tumor and the bladder wall muscle layer; processing the two nodes and one side between the two nodes based on a graph neural network model to determine an operation mode, a tumor invasion degree, a tumor diffusion degree, a tumor activity degree and a tumor health influence degree; determining a first recurrence rate of non-myogenic invasive bladder cancer based on the surgical procedure, the degree of tumor invasion, the degree of tumor spread, the degree of tumor activity, the degree of tumor health impact; acquiring a bladder MR image sequence of a plurality of time points after the operation of a patient and a urine cytology examination image sequence of a plurality of time points after the operation of the patient; determining a second recurrence rate of non-myogenic invasive bladder cancer using a second recurrence rate determination model based on the bladder MR image sequence at a plurality of time points after the patient is operated, the urine cytology examination image sequence at a plurality of time points after the patient is operated; determining a recurrence rate of the patient's non-myogenic invasive bladder cancer based on the first recurrence rate of the non-myogenic invasive bladder cancer and the second recurrence rate of the non-myogenic invasive bladder cancer.
Still further, the determining the recurrence rate of the non-myogenic invasive bladder cancer in the patient based on the first recurrence rate of the non-myogenic invasive bladder cancer and the second recurrence rate of the non-myogenic invasive bladder cancer comprises: and respectively giving different weights to the first recurrence rate of the non-muscle layer invasive bladder cancer and the second recurrence rate of the non-muscle layer invasive bladder cancer, and then carrying out weighted summation to obtain the recurrence rate of the non-muscle layer invasive bladder cancer of the patient.
Still further, the method further comprises: and prompting the patient to further check if the difference between the first recurrence rate of the non-muscle layer invasive bladder cancer and the second recurrence rate of the non-muscle layer invasive bladder cancer is greater than a difference threshold.
Still further, the method further comprises: and if the recurrence rate of the non-muscle invasive bladder cancer of the patient is greater than the recurrence rate threshold, reminding the patient to further check.
Furthermore, the input of the graph neural network model is the two nodes and one side between the two nodes, and the output of the graph neural network model is the operation mode, the tumor invasion degree, the tumor diffusion degree, the tumor activity degree and the tumor health influence degree, wherein the operation mode comprises a transurethral bladder tumor electrotome mode and a laser ablation mode.
According to a second aspect, the present invention provides an image-based non-muscle invasive bladder cancer recurrence rate prediction system, comprising: a first acquisition module for acquiring MR images of the bladder of a patient prior to treatment;
the bladder cancer judging module is used for judging whether the patient is non-muscular-invasive bladder cancer or not by using a muscular-layer invasive judging model based on the bladder MR image of the patient before treatment;
the image processing module is used for determining tumor information, a bladder wall muscle layer position, a direction and a distance of a tumor and the bladder wall muscle layer based on a bladder MR image of the patient before treatment if the patient is judged to belong to non-muscle layer invasive bladder cancer, wherein the tumor information comprises tumor size, tumor position, tumor morphology, a substrate size of the tumor attached to bladder mucosa, tumor tissue density, tumor edge characteristics and tumor blood supply conditions;
a building module for building two nodes and one side between two nodes based on the tumor information and the bladder wall muscle layer position, wherein the two nodes comprise a tumor node and a bladder wall muscle layer node, each node of the two nodes comprises a plurality of node characteristics, the node characteristics of the tumor node are tumor size, tumor position, tumor morphology, the size of a substrate of a tumor attached to a bladder mucosa, tumor tissue density, tumor edge characteristics and tumor blood supply condition, the node characteristics of the bladder wall muscle layer node are bladder wall muscle layer position, and the characteristics of one side between the two nodes comprise the direction and distance between the tumor and the bladder wall muscle layer;
the image neural network processing module is used for processing the two nodes and one edge between the two nodes based on the image neural network model to determine an operation mode, a tumor invasion degree, a tumor diffusion degree, a tumor activity degree and a tumor health influence degree;
a first recurrence rate determination module for determining a first recurrence rate of non-myogenic invasive bladder cancer based on the surgical procedure, the degree of tumor invasion, the degree of tumor spread, the degree of tumor activity, the degree of tumor health impact;
the second acquisition module is used for acquiring a bladder MR image sequence of a plurality of time points after the operation of the patient and a urine cytology examination image sequence of a plurality of time points after the operation of the patient;
a second recurrence rate determination module for determining a second recurrence rate of non-myogenic invasive bladder cancer using a second recurrence rate determination model based on the bladder MR image sequence at a plurality of time points after the patient's surgery, the urine cytology examination image sequence at a plurality of time points after the patient's surgery;
and a recurrence rate module for determining a recurrence rate of the non-myogenic invasive bladder cancer in the patient based on the first recurrence rate of the non-myogenic invasive bladder cancer and the second recurrence rate of the non-myogenic invasive bladder cancer.
Still further, the recurrence rate module is further configured to: and respectively giving different weights to the first recurrence rate of the non-muscle layer invasive bladder cancer and the second recurrence rate of the non-muscle layer invasive bladder cancer, and then carrying out weighted summation to obtain the recurrence rate of the non-muscle layer invasive bladder cancer of the patient.
Still further, the system is further configured to: and prompting the patient to further check if the difference between the first recurrence rate of the non-muscle layer invasive bladder cancer and the second recurrence rate of the non-muscle layer invasive bladder cancer is greater than a difference threshold.
Still further, the system is further configured to: and if the recurrence rate of the non-muscle invasive bladder cancer of the patient is greater than the recurrence rate threshold, reminding the patient to further check.
Furthermore, the input of the graph neural network model is the two nodes and one side between the two nodes, and the output of the graph neural network model is the operation mode, the tumor invasion degree, the tumor diffusion degree, the tumor activity degree and the tumor health influence degree, wherein the operation mode comprises a transurethral bladder tumor electrotome mode and a laser ablation mode.
The invention provides a method and a system for predicting recurrence rate of non-muscle invasive bladder cancer based on images, wherein the method comprises the steps of acquiring an MR image of the bladder before treatment of a patient; judging whether the patient is non-myogenic invasive bladder cancer or not by using a myogenic invasive judgment model based on the bladder MR image of the patient before treatment; if the patient is judged to belong to non-muscle invasive bladder cancer, determining tumor information, a bladder wall muscle layer position, a direction and a distance between a tumor and the bladder wall muscle layer by using an image processing model based on a bladder MR image of the patient before treatment, wherein the tumor information comprises a tumor size, a tumor position, a tumor morphology, a basal size of the tumor attached to bladder mucosa, a tumor tissue density, a tumor edge characteristic and a tumor blood supply condition; constructing two nodes and an edge between the two nodes based on the tumor information and the bladder wall muscle layer position, wherein the two nodes comprise a tumor node and a bladder wall muscle layer node, each node of the two nodes comprises a plurality of node characteristics, the node characteristics of the tumor node are tumor size, tumor position, tumor morphology, the size of a substrate of a tumor attached to bladder mucosa, tumor tissue density, tumor edge characteristics and tumor blood supply conditions, the node characteristics of the bladder wall muscle layer node are bladder wall muscle layer position, and the characteristics of the edge between the two nodes comprise the direction and distance between the tumor and the bladder wall muscle layer; processing the two nodes and one side between the two nodes based on a graph neural network model to determine an operation mode, a tumor invasion degree, a tumor diffusion degree, a tumor activity degree and a tumor health influence degree; determining a first recurrence rate of non-myogenic invasive bladder cancer based on the surgical procedure, the degree of tumor invasion, the degree of tumor spread, the degree of tumor activity, the degree of tumor health impact; acquiring a bladder MR image sequence of a plurality of time points after the operation of a patient and a urine cytology examination image sequence of a plurality of time points after the operation of the patient; determining a second recurrence rate of non-myogenic invasive bladder cancer using a second recurrence rate determination model based on the bladder MR image sequence at a plurality of time points after the patient is operated, the urine cytology examination image sequence at a plurality of time points after the patient is operated; determining a recurrence rate of the patient's non-myogenic invasive bladder cancer based on the first recurrence rate of the non-myogenic invasive bladder cancer and the second recurrence rate of the non-myogenic invasive bladder cancer. The method can accurately determine the recurrence rate of non-muscle invasive bladder cancer.
Drawings
FIG. 1 is a flow chart of a method for predicting recurrence rate of non-muscle invasive bladder cancer based on an image according to an embodiment of the present invention;
fig. 2 is a schematic diagram of an image-based prediction system for recurrence rate of non-muscular-layer invasive bladder cancer according to an embodiment of the present invention.
Detailed Description
In an embodiment of the present invention, an image-based prediction method for recurrence rate of non-muscular layer invasive bladder cancer is provided as shown in fig. 1, where the image-based prediction method for recurrence rate of non-muscular layer invasive bladder cancer includes steps S1 to S9:
step S1, acquiring an MR image of the bladder before treatment of the patient.
The bladder MR image is a three-dimensional image of the bladder obtained by Magnetic Resonance Imaging (MRI) techniques. The bladder MR image can provide a high resolution, non-invasive three-dimensional bladder image, and can clearly display information such as tumors, structures and the like in the bladder.
Step S2, judging whether the patient is non-muscular-invasive bladder cancer or not by using a muscular-layer-invasive judgment model based on the MR image of the bladder before the patient is treated;
the muscle layer infiltration judgment model is a deep neural network model, and the deep neural network model comprises a deep neural network (Deep Neural Networks, DNN). The deep neural network may include a plurality of processing layers, each processing layer being composed of a plurality of neurons, each neuron matrixing data. The parameters used by the matrix may be obtained by training. The deep neural network may include convolutional neural networks (Convolutional Neural Networks, CNN), generate countermeasure networks (Generative Adversarial Networks, GAN), and the like. In some embodiments, the initial muscle layer infiltration determination model may be trained by a gradient descent method to obtain a trained muscle layer infiltration determination model. The input of the muscle layer infiltration judging model is a bladder MR image of the patient before treatment, and the output of the muscle layer infiltration judging model is non-muscle layer infiltration bladder cancer or muscle layer infiltration bladder cancer. The trained myometrium infiltration judging model identifies and analyzes the characteristics and the structures in the images by identifying and analyzing the MR images of the bladder of the patient before treatment so as to determine whether the patient is non-myometrium infiltration bladder cancer.
And step S3, if the patient is judged to belong to non-muscle invasive bladder cancer, determining tumor information, bladder wall muscle layer positions, directions and distances of tumors and bladder wall muscle layers by using an image processing model based on bladder MR images of the patient before treatment, wherein the tumor information comprises tumor size, tumor positions, tumor forms, the size of a substrate of the tumor attached to bladder mucosa, tumor tissue density, tumor edge characteristics and tumor blood supply conditions.
Tumor information refers to tumor-related information extracted from bladder MR images, including tumor size, tumor location, tumor morphology, tumor size of the substrate attached to the bladder mucosa, tumor tissue density, tumor margin characteristics, tumor blood supply. These information can reflect the growth characteristics, invasiveness, malignancy, etc. of the tumor.
Tumor size refers to the volume or diameter of a tumor in an MR image of the bladder, typically expressed as a measure of the major and minor axes. Larger tumors are often associated with higher recurrence rates because they may already have more invasion of surrounding tissue.
Tumor location refers to specific location information of a tumor within the bladder. For example, the tumor location may be relative to and distance from the urethral meatus, bilateral ureter openings. Bladder cancer in different locations may have different risks of recurrence, e.g., tumors closer to the surface of the muscle layer of the bladder wall may be more prone to recurrence.
Tumor morphology refers to the shape characteristics of a tumor. The shape characteristics of the tumor can be obtained by analyzing the MR image of the bladder through an image processing model. Certain morphological features may be associated with aggressiveness and risk of recurrence of a tumor, such as irregular shapes, uneven surfaces, etc., may suggest malignancy of the tumor.
The size of the substrate on which the tumor adheres to the bladder mucosa refers to the size of the substrate on which the tumor cells are on the bladder mucosa. The larger the substrate, the more tightly the tumor may be attached to the surrounding tissue and the greater the likelihood of recurrence.
Tumor tissue density includes the density of cell arrangement and the degree of structural compaction within the tumor. High density tumors may indicate active cell proliferation and may increase the risk of recurrence.
Tumor margin features refer to the appearance and structural features of the tumor margin. Irregular, blurred tumor margins may mean that the tumor is more aggressive, increasing the likelihood of recurrence.
Tumor blood supply refers to the blood supply inside a tumor. Tumor blood supply conditions include blood supply enrichment, ischemia, abnormal vascular formation, and vascular invasion. By analyzing the bladder MR image, information about the blood supply of the tumor, such as the abundance of blood supply, the presence or absence of abnormal blood vessels, etc., can be obtained. The abundant blood supply may provide more nutrition and oxygen for tumor growth and metastasis, thereby increasing the risk of recurrence.
The direction of the tumor and the muscle layer of the bladder wall refers to the relative direction or angle between the tumor and the muscle layer of the bladder wall. The distance between the tumor and the muscle layer of the bladder wall refers to the spatial distance between the tumor and the muscle layer of the bladder wall. The closer the tumor is to the muscle layer of the bladder wall, the greater the degree of tumor invasion, and the more likely it is to relapse.
The image processing model is a convolutional neural network model. The convolutional neural network model includes a convolutional neural network, which may be a multi-layer neural network (e.g., including at least two layers). The at least two layers may include at least one of a convolutional layer (CONV), a modified linear unit (ReLU) layer, a pooling layer (POOL), or a fully-connected layer (FC). Convolutional neural networks are capable of extracting useful features from an image and progressively understanding and learning the contextual information of the image. In processing the bladder MR image, a convolutional neural network model can be used to identify and extract tumor information, bladder wall muscle layer location, and direction and distance of the tumor from the bladder wall muscle layer. The input of the image processing model is bladder MR image before the treatment of the patient, and the output of the image processing model is tumor information, bladder wall muscle layer position, direction and distance between the tumor and bladder wall muscle layer.
The convolutional neural network model consists of a plurality of convolutional layers, a pooling layer and a full-connection layer. In the convolution layer, the model applies a series of convolution kernels to extract local features in the image, such as edges, textures, etc. These convolution kernels may be slid over the image to capture features within regions of different sizes and shapes. The pooling layer is used for reducing the dimension of the features, reducing the calculation amount and improving the generalization capability of the model. The fully connected layer is used to map the output of the previous layer to the final classification or regression result.
By training the convolutional neural network model, the convolutional neural network model can be provided with the capability of extracting information of tumors and bladder wall muscle layers from the bladder MR image. During the training process, the model learns the characteristics related to the tumor and the muscle layer of the bladder wall, and uses the characteristics for the tasks of tumor positioning, morphological characteristic extraction, shortest distance measurement and the like. In addition, the CNN model can also realize end-to-end learning, namely directly taking an original image as input and outputting corresponding tumor information and bladder wall muscle layer information without manually designing a feature extraction method. The convolutional neural network model has strong capability in processing the bladder MR image, and can effectively determine tumor information, the position of the bladder wall muscle layer and the direction and distance between the tumor and the bladder wall muscle layer. This helps to more accurately assess the condition of non-myogenic invasive bladder cancer, formulate personalized treatment regimens, and predict recurrence risk.
In some embodiments, the image processing model includes a tumor region segmentation layer, a tumor information determination layer, a bladder wall muscle layer region segmentation layer, a bladder wall muscle layer position determination layer, a distance determination layer. The tumor region segmentation layer, the tumor information determination layer, the bladder wall muscle layer region segmentation layer, the bladder wall muscle layer position determination layer and the distance determination layer all comprise convolutional neural networks. The input of the tumor area dividing layer is a bladder MR image before treatment of a patient, the output of the tumor area dividing layer is an image of a divided tumor area, the input of the tumor information determining layer is an image of a divided tumor area, the output of the tumor information determining layer is tumor information, the input of the bladder wall muscle layer area dividing layer is a bladder MR image before treatment of a patient, the output of the bladder wall muscle layer dividing layer is a divided bladder wall muscle layer image, the input of the bladder wall muscle layer position determining layer is a divided bladder wall muscle layer image, the output of the bladder wall muscle layer position determining layer is a bladder wall muscle layer position, the input of the distance determining layer is an image of a divided tumor area and a divided bladder wall muscle layer image, and the output of the distance determining layer is the direction and distance between a tumor and the bladder wall muscle layer.
By dividing the image processing model into a plurality of layers, the design of the model can be made clearer and easier to manage. Each layer is responsible for handling specific tasks such as segmentation of tumor areas, information extraction, segmentation of bladder wall areas, information extraction, and distance calculation. The design makes the model more modularized, and is convenient to maintain and update. Each hierarchy may be focused on specific tasks, making the model more specialized and efficient in handling these tasks. For example, the tumor region segmentation layer focuses on segmenting a tumor region from an image, and the tumor information determination layer focuses on extracting useful information from the segmented tumor region. Such a design may allow each level to achieve optimal processing.
And S3, constructing two nodes and an edge between the two nodes based on the tumor information and the bladder wall muscle layer position, wherein the two nodes comprise a tumor node and a bladder wall muscle layer node, each node of the two nodes comprises a plurality of node characteristics, the node characteristics of the tumor node are tumor size, tumor position, tumor morphology, the size of a substrate of a tumor attached to a bladder mucosa, tumor tissue density, tumor edge characteristics and tumor blood supply condition, the node characteristics of the bladder wall muscle layer node are bladder wall muscle layer position, and the characteristics of the edge between the two nodes comprise the direction and distance between the tumor and the bladder wall muscle layer.
In the graphical structural data, nodes represent entities or concepts such as tumors and bladder wall muscle layers. Each node contains a set of features that describe the attributes or states of the entity.
Edges in the graph structure data, edges represent a relationship between two nodes.
The tumor node is a node constructed by tumor information, and the node characteristics of the tumor node are tumor size, tumor position, tumor morphology, the size of a substrate of a tumor attached to a bladder mucosa, tumor tissue density, tumor edge characteristics and tumor blood supply.
The bladder wall muscle layer node is a node constructed from bladder wall muscle layers, the node of the bladder wall muscle layer node being characterized by a bladder wall muscle layer location.
Features of one edge between the tumor node and the bladder wall muscle layer node include the direction and distance of the tumor from the bladder wall muscle layer. The direction of the tumor and the muscle layer of the bladder wall refers to the relative direction or angle between the tumor and the muscle layer of the bladder wall. The distance between the tumor and the muscle layer of the bladder wall refers to the spatial distance between the tumor and the muscle layer of the bladder wall.
And S4, processing the two nodes and one edge between the two nodes based on the graph neural network model to determine an operation mode, a tumor invasion degree, a tumor diffusion degree, a tumor activity degree and a tumor health influence degree.
The graphic neural network model comprises a graphic neural network (Graph Neural Network, GNN) and a full-connection layer, wherein the graphic neural network is a neural network directly acting on graphic structure data, and the graphic structure data is a data structure consisting of nodes and edges.
The input of the graphic neural network model is one side between the two nodes, and the output of the graphic neural network model is an operation mode, a tumor invasion degree, a tumor diffusion degree, a tumor activity degree and a tumor health influence degree, wherein the operation mode comprises a partial bladder excision operation, a full bladder excision operation, a laparoscopic operation and an auxiliary operation. The auxiliary surgery includes lymphadenotomy and prostatectomy.
The degree of tumor invasion indicates the degree of invasion of surrounding tissues by cancer cells. The degree of tumor spread refers to the spread of cancer cells in the body. The greater the extent of tumor invasion and spread, the more tumor cells have spread to surrounding tissues and organs. This means that even when a tumor is resected, there may still be minimal tumor residues or metastases, increasing the risk of tumor recurrence.
The tumor activity level is used for evaluating the proliferation and diffusion capacity of cancer cells, and the higher the tumor activity level is, the stronger the proliferation and diffusion capacity of the cancer cells is, and the activity level of the tumor reflects the proliferation capacity of the tumor cells. Tumor cells with high activity tend to have more malignant characteristics, have stronger proliferation capacity and have higher therapeutic resistance. This can lead to the potential for tumor cells that are not completely cleared during treatment, thereby increasing the risk of tumor recurrence.
The degree of tumor health effect is used to evaluate the extent of effect of lesions on the health of a patient. The extent of the effect of a tumor on human health is also related to tumor recurrence. For example, tumors may cause stress, damage or dysfunction to surrounding tissues and organs, thereby affecting the quality of life and therapeutic efficacy of the patient. If the health of the tumor is affected to a greater extent, this may lead to poorer overall patient performance and reduced resistance, thereby increasing the risk of tumor recurrence.
By constructing the graph structure, the relationship between nodes (e.g., tumor and bladder wall muscle layer) and edges (e.g., direction and distance of tumor from bladder wall muscle layer) can be clearly represented. This helps the model better understand and process the data. Considering the positional relationship between nodes helps the graph neural network model capture the spatial distribution characteristics between tumors. Recurrence and spread of non-myogenic invasive bladder cancer is generally closely related to the location of the tumor and surrounding tissue. By constructing the graph structure and analyzing the position relation among the nodes, the comprehensiveness and the accuracy of the graph neural network model prediction can be improved. The characteristics of the tumor nodes and the bladder wall muscle layer nodes are analyzed and processed through the graph neural network model, so that the degree of threat of the tumor to the life of the patient can be comprehensively estimated.
Step S5, determining a first recurrence rate of non-myogenic invasive bladder cancer based on the surgical mode, the tumor invasion degree, the tumor diffusion degree, the tumor activity degree and the tumor health influence degree.
The first recurrence rate of non-myogenic invasive bladder cancer is the preoperatively predicted recurrence rate.
In some embodiments, the surgical mode, the tumor invasion degree, the tumor diffusion degree, the tumor activity degree and the tumor health influence degree can be constructed as a vector to be matched, and the recurrence rate corresponding to the reference vector with the distance smaller than the threshold value is determined as the first recurrence rate of the non-myogenic invasive bladder cancer by calculating the distance between the vector to be matched and each reference vector in the database. The database is pre-constructed, the database comprises reference vectors and recurrence rates corresponding to the reference vectors, the reference vectors are constructed based on operation modes, tumor invasion degrees, tumor diffusion degrees, tumor activity degrees and tumor health influence degrees in historical data, and the recurrence rates corresponding to the reference vectors are determined as recurrence rates in the historical data.
In some embodiments, the first recurrence rate of non-myogenic invasive bladder cancer may be determined by querying a predetermined table of recurrence rates. The recurrence rate preset table comprises an operation mode, a tumor invasion degree, a tumor diffusion degree, a tumor activity degree, a tumor health influence degree and a corresponding recurrence rate, and the recurrence rate preset table can be artificially constructed based on historical data.
Step S6, acquiring a bladder MR image sequence of a plurality of time points after the operation of the patient and a urine cytology examination image sequence of a plurality of time points after the operation of the patient.
The plurality of time points after the operation of the patient may be a plurality of time points photographed once a day after the operation is completed. The multiple time points after the operation of the patient may also be multiple time points photographed every 12 hours after the completion of the operation.
The sequence of bladder MR images at a plurality of time points after the patient operation refers to the bladder MR images at a plurality of time points obtained by performing a magnetic resonance scan over a period of time after the patient operation. MR images of the bladder taken at various points in time after the patient's operation can indicate conditions within the bladder, including whether a new tumor or lesion has occurred. If new tumor growth or metastasis is found in the MR image, this may mean recurrence of the tumor. In addition, the growth speed and characteristic change of the tumor can be displayed by comparing MR images at different time points, so that the activity degree of the tumor and the response condition to treatment can be estimated.
The sequence of urine cytological examination images at a plurality of time points after the operation of the patient refers to urine cytological examination images at a plurality of time points obtained by microscopic examination of urine samples during a period of time after the operation of the patient. Urine cytology is a method for determining whether bladder cancer recurs by examining cancer cells, shed cells, and other abnormal cells in urine. Urine cytology examination is carried out at a plurality of time points after operation, and whether abnormal cell change exists in urine can be timely found. If an increase in cancer cells or abnormal cells is found in the urine cytologic examination at multiple time points, this may be indicative of recurrence of bladder cancer.
Step S7, determining a second recurrence rate of non-myogenic invasive bladder cancer based on the bladder MR image sequence at a plurality of time points after the operation of the patient and the urine cytology examination image sequence at a plurality of time points after the operation of the patient by using a second recurrence rate determination model.
The second recurrence rate of non-myogenic invasive bladder cancer is the postoperative predicted recurrence rate.
The second recurrence rate determination model is a long-short term neural network model. The Long-Short Term neural network model includes a Long-Short Term neural network (LSTM). The long-term and short-term neural network model can process sequence data with any length, capture sequence information and output results based on the association relationship of front data and rear data in the sequence. The bladder MR image sequences at a plurality of time points and the urine cytology examination image sequences at a plurality of time points in a continuous time period are processed through the long-short-term neural network model, so that the characteristics of the association relationship between the bladder MR image sequences and the urine cytology examination image sequences at all time points can be output and comprehensively considered, and the output characteristics are more accurate and comprehensive. The input of the second recurrence rate determination model comprises a bladder MR image sequence of a plurality of time points after the operation of the patient and a urine cytology examination image sequence of a plurality of time points after the operation of the patient, and the output of the second recurrence rate determination model is a second recurrence rate of non-myogenic invasive bladder cancer. A long and short term neural network model is used to learn the time dependence between these images. In the training process, the long-term and short-term neural network model can automatically extract and learn related characteristics and time dependence, so as to predict whether the patient can relapse bladder cancer in the future.
By combining the results of the bladder MR image sequence and the urine cytology examination image sequence, the tumor condition of the patient can be more comprehensively evaluated, thereby helping to determine the recurrence rate of bladder cancer tumors.
Step S8, determining the recurrence rate of the non-muscle layer invasive bladder cancer of the patient based on the first recurrence rate of the non-muscle layer invasive bladder cancer and the second recurrence rate of the non-muscle layer invasive bladder cancer.
In some embodiments, the first recurrence rate of the non-muscle invasive bladder cancer and the second recurrence rate of the non-muscle invasive bladder cancer may be given different weights, and then weighted and summed to obtain the recurrence rate of the non-muscle invasive bladder cancer in the patient.
In some embodiments, the method further comprises: and prompting the patient to further check if the difference between the first recurrence rate of the non-muscle layer invasive bladder cancer and the second recurrence rate of the non-muscle layer invasive bladder cancer is greater than a difference threshold. The difference threshold may be manually set in advance.
In some embodiments, the method further comprises: and if the recurrence rate of the non-muscle invasive bladder cancer of the patient is greater than the recurrence rate threshold, reminding the patient to further check. The recurrence rate threshold may be manually set in advance.
Based on the same inventive concept, fig. 2 is a schematic diagram of an image-based non-muscular-layer invasive bladder cancer recurrence rate prediction system according to an embodiment of the present invention, where the image-based non-muscular-layer invasive bladder cancer recurrence rate prediction system includes:
a first acquisition module 21 for acquiring MR images of the bladder prior to treatment of the patient;
a bladder cancer determination module 22, configured to determine whether the patient is non-myogenic invasive bladder cancer using a myogenic invasive determination model based on the MR image of the bladder before the patient is treated;
the image processing module 23 is configured to determine, based on an MR image of the bladder before treatment of the patient, tumor information including a tumor size, a tumor position, a tumor morphology, a size of a substrate to which the tumor adheres to a mucosa of the bladder, a tumor tissue density, a tumor edge feature, and a tumor blood supply condition, using an image processing model if the patient is determined to be a non-muscular invasive bladder cancer;
a construction module 24 for constructing two nodes and two nodes between the two nodes based on the tumor information and the bladder wall muscle layer position, wherein the two nodes comprise a tumor node and a bladder wall muscle layer node, each of the two nodes comprises a plurality of node features, the node features of the tumor node are tumor size, tumor position, tumor morphology, size of a substrate of tumor attached to a bladder mucosa, tumor tissue density, tumor edge feature, tumor blood supply, the node features of the bladder wall muscle layer node are bladder wall muscle layer position, and the features of one edge between the two nodes comprise direction and distance of the tumor from the bladder wall muscle layer;
the graph neural network processing module 25 is configured to determine a surgical mode, a tumor invasion degree, a tumor diffusion degree, a tumor activity degree, and a tumor health influence degree by processing the two nodes and one edge between the two nodes based on the graph neural network model;
a first recurrence rate determination module 26 for determining a first recurrence rate of non-myogenic invasive bladder cancer based on the surgical procedure, the degree of tumor invasion, the degree of tumor spread, the degree of tumor activity, the degree of tumor health impact;
a second acquisition module 27 for acquiring a bladder MR image sequence at a plurality of time points after a patient operation, a urine cytology examination image sequence at a plurality of time points after a patient operation;
a second recurrence rate determination module 28 for determining a second recurrence rate of non-myogenic invasive bladder cancer using a second recurrence rate determination model based on the bladder MR image sequence at a plurality of time points after the patient's surgery, the urine cytology examination image sequence at a plurality of time points after the patient's surgery;
a recurrence rate module 29 for determining a recurrence rate of the patient's non-myogenic invasive bladder cancer based on the first recurrence rate of the non-myogenic invasive bladder cancer and the second recurrence rate of the non-myogenic invasive bladder cancer.
Claims (10)
1. An image-based method for predicting recurrence rate of non-muscle invasive bladder cancer, comprising:
acquiring MR images of the bladder of the patient prior to treatment;
judging whether the patient is non-myogenic invasive bladder cancer or not by using a myogenic invasive judgment model based on the bladder MR image of the patient before treatment;
if the patient is judged to belong to non-muscle invasive bladder cancer, determining tumor information, a bladder wall muscle layer position, a direction and a distance between a tumor and the bladder wall muscle layer by using an image processing model based on a bladder MR image of the patient before treatment, wherein the tumor information comprises a tumor size, a tumor position, a tumor morphology, a basal size of the tumor attached to bladder mucosa, a tumor tissue density, a tumor edge characteristic and a tumor blood supply condition;
constructing two nodes and an edge between the two nodes based on the tumor information and the bladder wall muscle layer position, wherein the two nodes comprise a tumor node and a bladder wall muscle layer node, each node of the two nodes comprises a plurality of node characteristics, the node characteristics of the tumor node are tumor size, tumor position, tumor morphology, the size of a substrate of a tumor attached to bladder mucosa, tumor tissue density, tumor edge characteristics and tumor blood supply conditions, the node characteristics of the bladder wall muscle layer node are bladder wall muscle layer position, and the characteristics of the edge between the two nodes comprise the direction and distance between the tumor and the bladder wall muscle layer;
processing the two nodes and one side between the two nodes based on a graph neural network model to determine an operation mode, a tumor invasion degree, a tumor diffusion degree, a tumor activity degree and a tumor health influence degree;
determining a first recurrence rate of non-myogenic invasive bladder cancer based on the surgical procedure, the degree of tumor invasion, the degree of tumor spread, the degree of tumor activity, the degree of tumor health impact;
acquiring a bladder MR image sequence of a plurality of time points after the operation of a patient and a urine cytology examination image sequence of a plurality of time points after the operation of the patient;
determining a second recurrence rate of non-myogenic invasive bladder cancer using a second recurrence rate determination model based on the bladder MR image sequence at a plurality of time points after the patient is operated, the urine cytology examination image sequence at a plurality of time points after the patient is operated;
determining a recurrence rate of the patient's non-myogenic invasive bladder cancer based on the first recurrence rate of the non-myogenic invasive bladder cancer and the second recurrence rate of the non-myogenic invasive bladder cancer.
2. The image-based method of predicting the recurrence rate of non-muscle-layer invasive bladder cancer in a patient according to claim 1, wherein the determining the recurrence rate of non-muscle-layer invasive bladder cancer in the patient based on the first recurrence rate of non-muscle-layer invasive bladder cancer and the second recurrence rate of non-muscle-layer invasive bladder cancer comprises: and respectively giving different weights to the first recurrence rate of the non-muscle layer invasive bladder cancer and the second recurrence rate of the non-muscle layer invasive bladder cancer, and then carrying out weighted summation to obtain the recurrence rate of the non-muscle layer invasive bladder cancer of the patient.
3. The method of predicting the recurrence rate of image-based non-muscle-layer invasive bladder cancer of claim 1, further comprising: and prompting the patient to further check if the difference between the first recurrence rate of the non-muscle layer invasive bladder cancer and the second recurrence rate of the non-muscle layer invasive bladder cancer is greater than a difference threshold.
4. The method of predicting the recurrence rate of image-based non-muscle-layer invasive bladder cancer of claim 1, further comprising: and if the recurrence rate of the non-muscle invasive bladder cancer of the patient is greater than the recurrence rate threshold, reminding the patient to further check.
5. The method for predicting the recurrence rate of non-muscle invasive bladder cancer based on the image according to claim 1, wherein the input of the graphic neural network model is a side between the two nodes, and the output of the graphic neural network model is a surgical mode, a tumor invasion degree, a tumor diffusion degree, a tumor activity degree and a tumor health influence degree, and the surgical mode comprises a transurethral bladder tumor electro-ablation mode and a laser ablation mode.
6. An image-based non-muscular-layer invasive bladder cancer recurrence rate prediction system, comprising:
a first acquisition module for acquiring MR images of the bladder of a patient prior to treatment;
the bladder cancer judging module is used for judging whether the patient is non-muscular-invasive bladder cancer or not by using a muscular-layer invasive judging model based on the bladder MR image of the patient before treatment;
the image processing module is used for determining tumor information, a bladder wall muscle layer position, a direction and a distance of a tumor and the bladder wall muscle layer based on a bladder MR image of the patient before treatment if the patient is judged to belong to non-muscle layer invasive bladder cancer, wherein the tumor information comprises tumor size, tumor position, tumor morphology, a substrate size of the tumor attached to bladder mucosa, tumor tissue density, tumor edge characteristics and tumor blood supply conditions;
a building module for building two nodes and one side between two nodes based on the tumor information and the bladder wall muscle layer position, wherein the two nodes comprise a tumor node and a bladder wall muscle layer node, each node of the two nodes comprises a plurality of node characteristics, the node characteristics of the tumor node are tumor size, tumor position, tumor morphology, the size of a substrate of a tumor attached to a bladder mucosa, tumor tissue density, tumor edge characteristics and tumor blood supply condition, the node characteristics of the bladder wall muscle layer node are bladder wall muscle layer position, and the characteristics of one side between the two nodes comprise the direction and distance between the tumor and the bladder wall muscle layer;
the image neural network processing module is used for processing the two nodes and one edge between the two nodes based on the image neural network model to determine an operation mode, a tumor invasion degree, a tumor diffusion degree, a tumor activity degree and a tumor health influence degree;
a first recurrence rate determination module for determining a first recurrence rate of non-myogenic invasive bladder cancer based on the surgical procedure, the degree of tumor invasion, the degree of tumor spread, the degree of tumor activity, the degree of tumor health impact;
the second acquisition module is used for acquiring a bladder MR image sequence of a plurality of time points after the operation of the patient and a urine cytology examination image sequence of a plurality of time points after the operation of the patient;
a second recurrence rate determination module for determining a second recurrence rate of non-myogenic invasive bladder cancer using a second recurrence rate determination model based on the bladder MR image sequence at a plurality of time points after the patient's surgery, the urine cytology examination image sequence at a plurality of time points after the patient's surgery;
and a recurrence rate module for determining a recurrence rate of the non-myogenic invasive bladder cancer in the patient based on the first recurrence rate of the non-myogenic invasive bladder cancer and the second recurrence rate of the non-myogenic invasive bladder cancer.
7. The image-based non-muscle-layer invasive bladder cancer recurrence rate prediction system according to claim 6, wherein the recurrence rate module is further configured to: and respectively giving different weights to the first recurrence rate of the non-muscle layer invasive bladder cancer and the second recurrence rate of the non-muscle layer invasive bladder cancer, and then carrying out weighted summation to obtain the recurrence rate of the non-muscle layer invasive bladder cancer of the patient.
8. The image-based non-muscle-layer invasive bladder cancer recurrence rate prediction system according to claim 6, wherein the system is further configured to: and prompting the patient to further check if the difference between the first recurrence rate of the non-muscle layer invasive bladder cancer and the second recurrence rate of the non-muscle layer invasive bladder cancer is greater than a difference threshold.
9. The image-based non-muscle-layer invasive bladder cancer recurrence rate prediction system according to claim 6, wherein the system is further configured to: and if the recurrence rate of the non-muscle invasive bladder cancer of the patient is greater than the recurrence rate threshold, reminding the patient to further check.
10. The image-based non-muscle-layer invasive bladder cancer recurrence rate prediction system according to claim 6, wherein the input of the graph neural network model is a side between the two nodes, and the output of the graph neural network model is a surgical mode, a tumor invasion degree, a tumor diffusion degree, a tumor activity degree, and a tumor health influence degree, and the surgical mode includes a transurethral bladder tumor electro-ablation mode and a laser ablation mode.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410160502.4A CN117694839B (en) | 2024-02-05 | 2024-02-05 | Image-based prediction method and system for recurrence rate of non-myogenic invasive bladder cancer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410160502.4A CN117694839B (en) | 2024-02-05 | 2024-02-05 | Image-based prediction method and system for recurrence rate of non-myogenic invasive bladder cancer |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117694839A true CN117694839A (en) | 2024-03-15 |
| CN117694839B CN117694839B (en) | 2024-04-16 |
Family
ID=90162805
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410160502.4A Active CN117694839B (en) | 2024-02-05 | 2024-02-05 | Image-based prediction method and system for recurrence rate of non-myogenic invasive bladder cancer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117694839B (en) |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050262031A1 (en) * | 2003-07-21 | 2005-11-24 | Olivier Saidi | Systems and methods for treating, diagnosing and predicting the occurrence of a medical condition |
| WO2014025892A1 (en) * | 2012-08-10 | 2014-02-13 | Memorial Sloan-Kettering Cancer Center | Predicting bladder cancer responsiveness to bcg |
| US20140363373A1 (en) * | 2011-01-25 | 2014-12-11 | Oncofluor, Inc. | Method for combined imaging and treating organs and tissues |
| WO2016191567A1 (en) * | 2015-05-26 | 2016-12-01 | Memorial Sloan-Kettering Cancer Center | System, method and computer-accessible medium for texture analysis of hepatopancreatobiliary diseases |
| US20180285531A1 (en) * | 2017-03-29 | 2018-10-04 | Imaging Endpoints II LLC | Predicting Breast Cancer Responsiveness To Hormone Treatment Using Quantitative Textural Analysis |
| RU2688310C1 (en) * | 2017-12-05 | 2019-05-21 | Федеральное государственное бюджетное научное учреждение "Томский национальный исследовательский медицинский центр Российской академии наук" ("Томский НИМЦ") | Method for detecting perivesical fibre invasion in bladder cancer using dynamic magnetic resonance cystography |
| US20190251687A1 (en) * | 2018-02-09 | 2019-08-15 | Case Western Reserve University | Predicting prostate cancer biochemical recurrence using combined nuclear nf-kb/p65 localization and gland morphology |
| WO2019201186A1 (en) * | 2018-04-16 | 2019-10-24 | 图灵人工智能研究院（南京）有限公司 | Apparatus and method for identifying and evaluating tumor progression |
| CN113850788A (en) * | 2021-09-28 | 2021-12-28 | 中国医学科学院北京协和医院 | System for judging bladder cancer muscle layer infiltration state and application thereof |
| WO2022015000A1 (en) * | 2020-07-13 | 2022-01-20 | 가톨릭대학교 산학협력단 | Cancer progression/relapse prediction system and cancer progression/relapse prediction method using multiple images |
| CN114068003A (en) * | 2020-08-07 | 2022-02-18 | 四川医枢科技股份有限公司 | Clinical decision, teaching and scientific research auxiliary support system and method for urinary system tumor |
| CN114240839A (en) * | 2021-11-17 | 2022-03-25 | 东莞市人民医院 | Bladder tumor muscle layer invasion prediction method based on deep learning and related device |
| WO2022197044A1 (en) * | 2021-03-15 | 2022-09-22 | 주식회사 딥바이오 | Bladder lesion diagnosis method using neural network, and system thereof |
| CN115661107A (en) * | 2022-11-07 | 2023-01-31 | 中国医学科学院北京协和医院 | Image analysis method, system and equipment based on bladder cancer risk stratification |
| WO2023086970A2 (en) * | 2021-11-12 | 2023-05-19 | Arna Genomics Us Incorporated | Methods and compositions for analyzing nucleic acids and nucleases in a body fluid as indicators of disease |
-
2024
- 2024-02-05 CN CN202410160502.4A patent/CN117694839B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050262031A1 (en) * | 2003-07-21 | 2005-11-24 | Olivier Saidi | Systems and methods for treating, diagnosing and predicting the occurrence of a medical condition |
| US20140363373A1 (en) * | 2011-01-25 | 2014-12-11 | Oncofluor, Inc. | Method for combined imaging and treating organs and tissues |
| WO2014025892A1 (en) * | 2012-08-10 | 2014-02-13 | Memorial Sloan-Kettering Cancer Center | Predicting bladder cancer responsiveness to bcg |
| WO2016191567A1 (en) * | 2015-05-26 | 2016-12-01 | Memorial Sloan-Kettering Cancer Center | System, method and computer-accessible medium for texture analysis of hepatopancreatobiliary diseases |
| US20180285531A1 (en) * | 2017-03-29 | 2018-10-04 | Imaging Endpoints II LLC | Predicting Breast Cancer Responsiveness To Hormone Treatment Using Quantitative Textural Analysis |
| RU2688310C1 (en) * | 2017-12-05 | 2019-05-21 | Федеральное государственное бюджетное научное учреждение "Томский национальный исследовательский медицинский центр Российской академии наук" ("Томский НИМЦ") | Method for detecting perivesical fibre invasion in bladder cancer using dynamic magnetic resonance cystography |
| US20190251687A1 (en) * | 2018-02-09 | 2019-08-15 | Case Western Reserve University | Predicting prostate cancer biochemical recurrence using combined nuclear nf-kb/p65 localization and gland morphology |
| WO2019201186A1 (en) * | 2018-04-16 | 2019-10-24 | 图灵人工智能研究院（南京）有限公司 | Apparatus and method for identifying and evaluating tumor progression |
| WO2022015000A1 (en) * | 2020-07-13 | 2022-01-20 | 가톨릭대학교 산학협력단 | Cancer progression/relapse prediction system and cancer progression/relapse prediction method using multiple images |
| CN114068003A (en) * | 2020-08-07 | 2022-02-18 | 四川医枢科技股份有限公司 | Clinical decision, teaching and scientific research auxiliary support system and method for urinary system tumor |
| WO2022197044A1 (en) * | 2021-03-15 | 2022-09-22 | 주식회사 딥바이오 | Bladder lesion diagnosis method using neural network, and system thereof |
| CN113850788A (en) * | 2021-09-28 | 2021-12-28 | 中国医学科学院北京协和医院 | System for judging bladder cancer muscle layer infiltration state and application thereof |
| WO2023086970A2 (en) * | 2021-11-12 | 2023-05-19 | Arna Genomics Us Incorporated | Methods and compositions for analyzing nucleic acids and nucleases in a body fluid as indicators of disease |
| CN114240839A (en) * | 2021-11-17 | 2022-03-25 | 东莞市人民医院 | Bladder tumor muscle layer invasion prediction method based on deep learning and related device |
| CN115661107A (en) * | 2022-11-07 | 2023-01-31 | 中国医学科学院北京协和医院 | Image analysis method, system and equipment based on bladder cancer risk stratification |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117694839B (en) | 2024-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111539930B (en) | Dynamic ultrasonic breast nodule real-time segmentation and identification method based on deep learning | |
| CN111798462B (en) | Automatic delineation method of nasopharyngeal carcinoma radiotherapy target area based on CT image | |
| CN111640120B (en) | Pancreas CT automatic segmentation method based on significance dense connection expansion convolution network | |
| CN110782474B (en) | Deep learning-based method for predicting morphological change of liver tumor after ablation | |
| Lin et al. | Computer-aided kidney segmentation on abdominal CT images | |
| Banerjee et al. | Single seed delineation of brain tumor using multi-thresholding | |
| CN112396621B (en) | High-resolution microscopic endoscope image nucleus segmentation method based on deep learning | |
| Nikan et al. | PWD-3DNet: a deep learning-based fully-automated segmentation of multiple structures on temporal bone CT scans | |
| CN111353998A (en) | Tumor diagnosis and treatment prediction model and device based on artificial intelligence | |
| Eltoukhy et al. | Classification of multiclass histopathological breast images using residual deep learning | |
| CN113299372B (en) | Photoacoustic pathological image processing method, storage medium and terminal device | |
| CN117694839B (en) | Image-based prediction method and system for recurrence rate of non-myogenic invasive bladder cancer | |
| Lin et al. | Hyperparameter Optimization of Deep Learning Networks for Classification of Breast Histopathology Images. | |
| Rajesh et al. | Lung cancer diagnosis and treatment using AI and Mobile applications | |
| CN116128895A (en) | Medical image segmentation method, apparatus and computer readable storage medium | |
| CN114511564B (en) | Image analysis method for breast cancer residual tumor load based on DCE-MRI | |
| CN113476033B (en) | Deep neural network-based automatic generation method for prostatic hyperplasia target area | |
| CN115439491A (en) | Segmentation method and device for intestinal injury region of MRI (magnetic resonance imaging) image and grade evaluation system | |
| CN115222675A (en) | Hysteromyoma automatic typing method and device based on deep learning | |
| Shao et al. | Semantic segmentation method of 3D liver image based on contextual attention model | |
| CN114519722A (en) | Carotid artery extraction method based on convolutional neural network | |
| Rodriguez et al. | Blood vessel segmentation via neural network in histological images | |
| Li | Fully automatic segmentation of kidney and tumor based on cascaded U-Nets | |
| Sengupta et al. | Optimized segmentation of white patches in skin lesion images | |
| KR102311472B1 (en) | Method for predicting regions of normal tissue and device for predicting regions of normal tissue using the same |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |