Prostate cancer prediction system based on multiple modes and construction method

Abstract

The invention discloses a multi-mode-based prostate cancer prediction system and a construction method, and relates to the technical field of prostate cancer prediction.The construction method comprises the steps that prostate index data are collected through a data collection module, and then a data processing module calculates a comprehensive prostate index coefficient according to the prostate index data; then specific value judgment is carried out by setting prostate index uniformity parameters, whether the prostate is abnormal or not is determined, and if yes, the image processing module collects a prostate CT image and outputs the prostate CT image to obtain a prostate image recognition result; and integrating the comprehensive prostate index coefficient and the prostate image recognition result through a prediction module to obtain a prostate integrated index result, inputting the prostate integrated index result into a prostate cancer prediction model, and outputting to obtain a prostate cancer prediction result.

Description

Technical Field

[0001] The present invention relates to the technical field of prostate cancer prediction, specifically a multi-modal based prostate cancer prediction system and its construction method. Background Art

[0002] Regarding the etiology of prostate cancer, it is related to genetics, environment, diet, and age. Those with a family history of prostate cancer have a relatively high incidence rate and tend to be younger in age. Prostate cancer is prone to occur in elderly men over 65 years old, those with an unhealthy lifestyle, and those with a direct relative who has had prostate cancer. Factors such as diet and obesity are likely to induce it.

[0003] Currently, it is a great challenge for clinicians to determine whether patients with a prostate-specific antigen (PSA) level between 4 and 10 have prostate cancer or prostate hyperplasia through prostate imaging, and often biopsies need to be performed to confirm. However, biopsies often increase the detection time or cause pain to patients. Summary of the Invention

[0004] To solve the deficiencies mentioned in the above background art, the purpose of the present invention is to provide a multi-modal based prostate cancer prediction system and its construction method, which can quickly predict prostate problems of patients.

[0005] In the first aspect, the purpose of the present invention can be achieved through the following technical solutions: A multi-modal based prostate cancer prediction system, including:

[0006] A data acquisition module, a data processing module, an image processing module, a prediction module, and a server;

[0007] The data acquisition module is used to collect prostate index data; and send the prostate index data to the data processing module for processing, where the prostate index data includes prostate-specific antigen and digital prostate index;

[0008] The data processing module is used to preprocess and label the prostate index data, calculate a comprehensive index using the labeled prostate index data to obtain a comprehensive prostate index coefficient, set a prostate index uniformity parameter, calculate the ratio of the comprehensive prostate index coefficient to the prostate index uniformity parameter, set a ratio threshold, compare the ratio with the ratio threshold, and determine whether there is an abnormality in prostate function according to the comparison result. If there is an abnormality, send an image processing signal to the image processing module; if there is no abnormality, send a normal signal to the server;

[0009] The image processing module is used to obtain prostate CT images, label the ROI regions for image segmentation of the prostate CT images to obtain the prostate segmentation regions, input the labeled prostate segmentation regions into a pre-established prostate image recognition model, and output the prostate image recognition results; send the prostate image recognition results to the prediction module;

[0010] The prediction module is used to obtain the comprehensive prostate index coefficient in the data processing module, integrate the prostate image recognition results with the comprehensive prostate index coefficient to obtain the prostate integration index results, input the prostate integration index results into a pre-established prostate cancer prediction model, and output the prostate cancer prediction results; according to the prostate cancer prediction results, send a prediction ratio signal to the server;

[0011] The server is used to remind the corresponding patient that there is no abnormal problem with the prostate after receiving the normal signal, and remind the corresponding patient of the predicted possibility ratio of prostate cancer occurrence after receiving the prediction ratio signal.

[0012] Combined with the first aspect, in some implementation manners of the first aspect, the system further includes: the calculation process of the data processing module:

[0013] Label the prostate-specific antigen as Ki and the digital prostate index as Xi; where i is the acquisition times label of the data acquisition module, and i = 1, 2, 3,..., n, and n is the total number of acquisition times of the data acquisition module;

[0014] The calculation process of the comprehensive prostate index coefficient is as follows:

[0015] Use the formula Calculate the comprehensive prostate index coefficient Zhi, where X0 is the standard prostate-specific antigen coefficient, K0 is the standard digital prostate index coefficient, a1 is the preset prostate-specific antigen correlation coefficient, and a2 is the preset digital prostate index correlation coefficient.

[0016] Combined with the first aspect, in some implementation manners of the first aspect, the system further includes: the analysis process of the data processing module:

[0017] Set the maximum prostate index Zhmax and the minimum prostate index Zhmin, use Zhmax - Zhmin as the numerator of the formula for calculating the prostate index uniformity parameter, use Zhmax + Zhmin as the denominator of the formula for calculating the prostate index uniformity parameter, and use the ratio of the numerator to the denominator of the formula as the prostate index uniformity parameter, which is Zh0. The formula is as follows:

[0018]

[0019] Calculate the ratio Bli of the comprehensive prostate index coefficient Zhi to the prostate index uniformity parameter Zh0, and set the ratio threshold Bl0;

[0020] If Bli > Bl0, send an image processing signal to the image processing module;

[0021] If Bli ≤ Bl0, send a normal signal to the server.

[0022] Combined with the first aspect, in some implementation manners of the first aspect, the system further includes: the calculation process of data integration of the prediction module:

[0023]

[0024] In the formula, Zbi is the result of the prostate integration index, Si is the result of prostate image recognition, and both α and β are preset proportionality coefficients.

[0025] Combined with the first aspect, in some implementation manners of the first aspect, the system further includes: the training process of the prostate cancer prediction model in the prediction module:

[0026] Acquire and obtain standard prostate index and standard prostate image data through the data acquisition unit in the prediction module, where the prostate index includes standard prostate specific antigen data and standard digital prostate index data; the prostate image data is the standard ROI segmentation area of the prostate;

[0027] Combine the standard prostate index and the standard prostate image data to generate a standard prostate data set, and input the standard prostate data set into the artificial intelligence model for training to output a prostate cancer prediction model.

[0028] In a second aspect, in order to achieve the above object, the present invention discloses a construction method of a prostate cancer prediction system based on multi-modal, and the method includes the following steps:

[0029] Acquire prostate index data, preprocess and label the prostate index data to obtain the labeled prostate index data, and use the labeled prostate index data to calculate the comprehensive index to obtain the comprehensive prostate index coefficient;

[0030] Set the prostate index uniformity parameter, calculate the ratio of the comprehensive prostate index coefficient to the prostate index uniformity parameter, set the ratio threshold, compare the ratio with the ratio threshold, and determine whether there is an abnormality in the prostate function according to the comparison result;

[0031] If the ratio is not greater than the ratio threshold, the prostate function is normal. If the ratio is greater than the ratio threshold, obtain the prostate CT image, segment the prostate CT image to obtain the prostate segmentation region, input the prostate segmentation region into the pre-established prostate image recognition model, and output the prostate image recognition result;

[0032] Integrate the comprehensive prostate index coefficient and the prostate image recognition result to obtain the prostate integration index result, input the prostate integration index result into the pre-established prostate cancer prediction model, and output the prostate cancer prediction result.

[0033] In another aspect of the present invention, in order to achieve the above object, a terminal device is disclosed, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The memory stores a computer program capable of running on the processor. When the processor loads and executes the computer program, the above-mentioned multi-modal based prostate cancer prediction system is adopted.

[0034] In another aspect of the present invention, in order to achieve the above object, a computer-readable storage medium is disclosed. The computer-readable storage medium stores a computer program. When the computer program is loaded and executed by the processor, the above-mentioned multi-modal based prostate cancer prediction system is adopted.

[0035] Advantages of the present invention:

[0036] The present invention collects prostate index data through the data collection module, then the data processing module calculates the comprehensive prostate index coefficient using the prostate index data, and then determines the ratio by setting the prostate index uniformity parameter to determine whether the prostate is abnormal. If abnormal, the image processing module collects the prostate CT image and outputs the prostate image recognition result. Then, the prediction module integrates the comprehensive prostate index coefficient and the prostate image recognition result to obtain the prostate integration index result, and then inputs it into the prostate cancer prediction model to output the prostate cancer prediction result. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts;

[0038] Figure 1 is a schematic structural diagram of the system of the present invention;

[0039] Figure 2 is a schematic flowchart of the method of the present invention. Detailed implementation manners

[0040] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0041] Embodiment 1:

[0042] Next, relevant terms related to the embodiments of the present application will be introduced:

[0043] Convolutional neural network: A convolutional neural network (CNN) is a class of feedforward neural networks (Feedforward Neural Networks) that contain convolutional calculations and have a deep structure, and is one of the representative algorithms of deep learning. A convolutional neural network has the ability of representation learning and can perform shift-invariant classification on input information according to its hierarchical structure, so it is also called "Shift-Invariant Artificial Neural Networks (SIANN)".

[0044] As Figure 1 shown, the prostate cancer prediction system based on multi-modal includes:

[0045] A data acquisition module, a data processing module, an image processing module, a prediction module and a server;

[0046] The data acquisition module is used to acquire prostate index data; and send the prostate index data to the data processing module for processing, where the prostate index data includes prostate specific antigen and digital prostate index;

[0047] It should be further noted that in the specific implementation process, prostate specific antigen (PSA): Prostate cancer cells will produce PSA, and an increase in the PSA level in the blood can be an indicator of prostate cancer; digital prostate index (PSAD): This index combines the PSA level with the prostate volume and can more accurately evaluate whether a patient is likely to have prostate cancer;

[0048] After receiving the prostate index data sent by the data acquisition module, the data processing module performs data processing. Specifically, the processing process of the data processing module includes the following steps:

[0049] Preprocess the prostate index data to obtain the preprocessed prostate index data;

[0050] In this embodiment, the process of preprocessing the prostate index data includes the data cleaning process of format standardization, abnormal data clearing, error correction, and duplicate data clearing;

[0051] Use the preprocessed prostate index data for annotation. Among them, prostate-specific antigen is marked as Ki, and digital prostate index is marked as Xi; where i is the acquisition times label of the data acquisition module, and i = 1, 2, 3,..., n, and n is the total number of acquisition times of the data acquisition module;

[0052] Use the annotated prostate index data to calculate the comprehensive index, and obtain the comprehensive prostate index coefficient. Specifically, the calculation process of the comprehensive prostate index coefficient is as follows:

[0053] Use the formula Calculate the comprehensive prostate index coefficient Zhi. In the formula, X0 is the standard prostate-specific antigen coefficient, K0 is the standard digital prostate index coefficient, a1 is the preset prostate-specific antigen correlation coefficient, and a2 is the preset digital prostate index correlation coefficient;

[0054] Set the maximum value Zhmax and the minimum value Zhmin of the prostate index. Take Zhmax - Zhmin as the numerator of the formula for calculating the prostate index uniformity parameter, take Zhmax + Zhmin as the denominator of the formula for calculating the prostate index uniformity parameter, and take the ratio of the numerator to the denominator of the formula as the prostate index uniformity parameter, that is, Zh0. The formula is as follows:

[0055]

[0056] Calculate the ratio of the comprehensive prostate index coefficient Zhi to the prostate index uniformity parameter Zh0, that is, Bli. Set the ratio threshold Bl0, compare the ratio Bli with the ratio threshold Bl0, and determine whether there is an abnormality in the prostate function according to the comparison result:

[0057] If Bli > Bl0, it is determined that there is an abnormality in the prostate function, and the data processing module sends an image processing signal to the image processing module;

[0058] If Bli ≤ Bl0, it is determined that there is no abnormality in the prostate function, and the data processing module sends a normal signal to the server;

[0059] After receiving the image processing signal sent by the data processing module, the image processing module performs image processing and analysis. Specifically, the processing and analysis process of the image processing module includes the following steps:

[0060] Obtain the prostate CT image, perform ROI region annotation for image segmentation on the prostate CT image to obtain the prostate segmentation region; among them, in this embodiment, U-net segmentation model is used for image segmentation;

[0061] Input the labeled prostate segmentation region into the pre-established prostate image recognition model, and output the prostate image recognition result; and send the prostate image recognition result to the prediction module;

[0062] After receiving the prostate image recognition result sent by the image processing module, the prediction module obtains the comprehensive prostate index coefficient Zhi in the data processing module; integrates the prostate image recognition result with the comprehensive prostate index coefficient Zhi to obtain the prostate integration index result;

[0063] Furthermore, the calculation process of data integration by the prediction module:

[0064]

[0065] In the formula, Zbi is the prostate integration index result, Si is the prostate image recognition result, and both α and β are preset proportionality coefficients;

[0066] Input the prostate integration index result into the pre-established prostate cancer prediction model, and output the prostate cancer prediction result; according to the prostate cancer prediction result, send a prediction ratio signal to the server;

[0067] It should be noted that the prostate cancer prediction model is trained based on an artificial intelligence model;

[0068] Specifically, the following further elaborates the solution of the present invention through embodiments: The process of training the prostate cancer prediction model based on the artificial intelligence model is as follows:

[0069] Collect and obtain standard prostate indicators and standard prostate image data through the data acquisition unit in the prediction module. Among them, the prostate indicators include standard prostate specific antigen data and standard digital prostate index data; the prostate image data is the standard ROI segmentation region of the prostate;

[0070] Combine standard prostate indicators and standard prostate image data to generate a standard prostate dataset, and input the standard prostate dataset into an artificial intelligence model for training to output a prostate cancer prediction model. Among them, the artificial intelligence model includes a deep convolutional neural network model and an RBF neural network model.

[0071] After receiving the normal signal sent by the data processing module, the server reminds the corresponding patient that there is no abnormality in the prostate. After receiving the prediction ratio signal sent by the prediction module, the server reminds the corresponding patient of the predicted probability ratio of prostate cancer occurrence.

[0072] Embodiment 2: Second aspect, as Figure 2 shown, a construction method of a multi-modal prostate cancer prediction system, the method includes the following steps:

[0073] S101: Obtain prostate indicator data, preprocess and label the prostate indicator data to obtain the labeled prostate indicator data, and use the labeled prostate indicator data to calculate comprehensive indicator to obtain a comprehensive prostate indicator coefficient;

[0074] S102: Set the prostate indicator uniformity parameter, calculate the ratio of the comprehensive prostate indicator coefficient to the prostate indicator uniformity parameter, set a ratio threshold, compare the ratio with the ratio threshold, and determine whether there is an abnormality in the prostate function according to the comparison result;

[0075] S103: If the ratio is not greater than the ratio threshold, the prostate function is normal. If the ratio is greater than the ratio threshold, obtain the prostate CT image, segment the prostate CT image to obtain the prostate segmentation region, and input the prostate segmentation region into the pre-established prostate image recognition model to output the prostate image recognition result;

[0076] S104: Integrate the comprehensive prostate indicator coefficient and the prostate image recognition result to obtain the prostate integration indicator result, and input the prostate integration indicator result into the pre-established prostate cancer prediction model to output the prostate cancer prediction result.

[0077] Based on the same inventive concept, the present invention further provides a computer device, which includes: one or more processors, and a memory for storing one or more computer programs; the program includes program instructions, and the processor is configured to execute the program instructions stored in the memory. The processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and is used to implement one or more instructions. Specifically, it is used to load and execute one or more instructions in the computer storage medium to implement the above method.

[0078] It should be further noted that, based on the same inventive concept, the present invention further provides a computer storage medium, on which a computer program is stored, and the computer program, when run by a processor, executes the above method. The storage medium may adopt any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may, for example, but is not limited to, an electrical, magnetic, optical, electrical, magnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the computer-readable storage medium include: an electrical connection having one or more wires, a portable computer disk, a hard disk, a Random Access Memory (RAM), a Read Only Memory (ROM), an Erasable Programmable Read Only Memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, the computer-readable storage medium may be any tangible medium that contains or stores a program, and the program may be used by or combined with an instruction execution system, apparatus, or device.

[0079] The above formulas are all calculated by removing the dimension and taking their numerical values. The formulas are obtained by collecting a large amount of data for software simulation to obtain a formula that is closest to the actual situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained through a large amount of data simulation.

[0080] In the description of this specification, the descriptions referring to the terms "one embodiment", "example", "specific example", etc. mean that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0081] The above shows and describes the basic principles, main features and advantages of the present disclosure. Those skilled in the art should understand that the present disclosure is not limited by the above embodiments, and what is described in the above embodiments and the specification only illustrates the principles of the present disclosure. Without departing from the spirit and scope of the present disclosure, the present disclosure will have various changes and improvements, and these changes and improvements all fall within the scope of the present disclosure claimed.

Claims

1. A multimodal prostate cancer prediction system, characterized in that: include: Data acquisition module, data processing module, image processing module, prediction module and server; The data acquisition module is used to collect prostate index data; and send the prostate index data to the data processing module for processing, wherein the prostate index data includes prostate specific antigen and digital prostate index; The data processing module is used to pre-process and annotate the prostate index data, calculate the comprehensive index using the annotated prostate index data, obtain the comprehensive prostate index coefficient, set the prostate index uniformity parameter, calculate the ratio of the comprehensive prostate index coefficient and the prostate index uniformity parameter, set the ratio threshold, compare the ratio with the ratio threshold, and determine whether the prostate function is abnormal based on the comparison result. If there is an abnormality, an image processing signal is sent to the image processing module, and if there is no abnormality, a normal signal is sent to the server; The image processing module is used to obtain a prostate CT image, annotate the ROI region of the prostate CT image for image segmentation, obtain a prostate segmentation region, input the annotated prostate segmentation region into a pre-established prostate image recognition model, and output a prostate image recognition result; and send the prostate image recognition result to the prediction module; The prediction module is used to obtain the comprehensive prostate index coefficient in the data processing module, perform data integration on the prostate image recognition result and the comprehensive prostate index coefficient to obtain the prostate integrated index result, input the prostate integrated index result into the pre-established prostate cancer prediction model, and output the prostate cancer prediction result; according to the prostate cancer prediction result, send a prediction ratio signal to the server; The server is used to remind the corresponding patient that there is no abnormality in the prostate after receiving the normal signal, and to remind the corresponding patient of the predicted probability ratio of prostate cancer after receiving the predicted ratio signal.

2. The multimodal prostate cancer prediction system according to claim 1, characterized in that: The calculation process of the data processing module: The prostate specific antigen is marked as Ki, and the digital prostate index is marked as Xi; wherein i is the number of acquisition times of the data acquisition module, and i=1, 2, 3, ..., n, and n is the total number of acquisition times of the data acquisition module; The calculation process of the comprehensive prostate index coefficient is as follows: Using the formula The comprehensive prostate index coefficient Zhi is calculated, where X0 is the standard prostate-specific antigen coefficient, K0 is the standard digital prostate index coefficient, a1 is the preset prostate-specific antigen correlation coefficient, and a2 is the preset digital prostate index correlation coefficient.

3. The multimodal prostate cancer prediction system according to claim 2, characterized in that: The analysis process of the data processing module: Set the maximum value Zhmax and the minimum value Zhmin of the prostate index, use Zhmax-Zhmin as the numerator of the formula for calculating the uniformity parameter of the prostate index, use Zhmax+Zhmin as the denominator of the formula for calculating the uniformity parameter of the prostate index, and use the ratio of the numerator to the denominator of the formula as the uniformity parameter of the prostate index, which is Zh0. The formula is as follows: Calculate the ratio Bli of the comprehensive prostate index coefficient Zhi and the prostate index uniformity parameter Zh0, and set the ratio threshold Bl0; If Bli>Bl0, the image processing signal is sent to the image processing module; If Bli≤Bl0, a normal signal is sent to the server.

4. The multimodal prostate cancer prediction system according to claim 1, characterized in that: The calculation process of data integration of the prediction module: Where Zbi is the prostate integrated index result, Si is the prostate image recognition result, and α and β are both preset proportional coefficients.

5. The multimodal prostate cancer prediction system according to claim 4, characterized in that: The training process of the prostate cancer prediction model in the prediction module: The data acquisition unit in the prediction module acquires standard prostate indexes and standard prostate image data, wherein the prostate indexes include standard prostate specific antigen data and standard digital prostate index data; the prostate image data is a standard ROI segmentation area of ​​the prostate; Standard prostate indicators and standard prostate image data are combined to generate a standard prostate data set, which is input into an artificial intelligence model for training, and the output is a prostate cancer prediction model.

6. A method for constructing a multimodal prostate cancer prediction system, characterized in that: The method comprises the following steps: Acquire prostate index data, pre-process and label the prostate index data to obtain labeled prostate index data, and use the labeled prostate index data to calculate comprehensive indicators to obtain comprehensive prostate index coefficients; Setting a prostate index uniformity parameter, calculating a ratio of a comprehensive prostate index coefficient to a prostate index uniformity parameter, setting a ratio threshold, comparing the ratio with the ratio threshold, and determining whether there is an abnormality in prostate function based on the comparison result; If the ratio is not greater than the ratio threshold, the prostate function is normal; if the ratio is greater than the ratio threshold, a prostate CT image is acquired, the prostate CT image is segmented to obtain a prostate segmentation region, the prostate segmentation region is input into a pre-established prostate image recognition model, and a prostate image recognition result is output; The comprehensive prostate index coefficient and the prostate image recognition result are integrated to obtain the prostate integrated index result, and the prostate integrated index result is input into a pre-established prostate cancer prediction model to output the prostate cancer prediction result.

7. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: The memory stores a computer program that can be run on a processor, and when the processor loads and executes the computer program, the multi-modality-based prostate cancer prediction system described in any one of claims 1 to 5 is adopted.

8. A computer-readable storage medium having a computer program stored therein, characterized in that: When the computer program is loaded and executed by a processor, the multimodality-based prostate cancer prediction system according to any one of claims 1 to 5 is adopted.

Images