Intelligent tumor diagnosis and treatment auxiliary system and method based on large language model

Abstract

The present invention relates to an intelligent tumor diagnosis and treatment auxiliary system and method based on a large language model, which creates a tumor computer simulation system and performs three-dimensional simulation of the tumor in the first tissue based on CT image data to obtain a first three-dimensional model, calculates its tumor volume, and selects a first treatment method when the tumor volume is greater than a preset volume threshold, otherwise the first three-dimensional model is cut in the computer simulation system and the tumor volume in each model is calculated respectively; CT image data of the tumor of the first tissue at another time point is obtained, and the tumor of the tissue is simulated in three dimensions through the computer simulation system, and the second three-dimensional model is cut after obtaining, and the tumor volume and tumor growth rate in each model after cutting are calculated respectively, and the tumor growth rate in each model is compared with the preset growth rate threshold, and the treatment method is adjusted according to the growth rate. The selection of treatment methods is assisted to improve the decision-making efficiency of treatment methods.

Description

Technical Field

[0001] The present invention belongs to the technical field of server heat dissipation, and in particular to an intelligent tumor diagnosis and treatment auxiliary system and method based on a large language model. Background Art

[0002] A tumor is a neoplasm formed by the proliferation of local tissue cells under the action of various tumorigenic factors, and it often manifests as a local abnormal mass. Tumor cells grow vigorously and out of control, and are relatively autonomous. They can still stably propagate after the initial factors that cause neoplastic proliferation disappear. A tumor is a disease that affects the entire body, and especially for patients with mid- to late-stage tumors, most patients may develop tumors in multiple organs of the body. Therefore, the diagnosis and treatment process of tumor patients will inevitably involve multiple disciplines. Because of this, during the treatment of existing tumor patients, doctors from different departments will give different treatment suggestions. For example, some doctors in some departments recommend treatment through surgery, while doctors in other departments recommend treatment through radiotherapy and chemotherapy, or doctors in other departments will give other treatment suggestions.

[0003] Therefore, in the prior art, there is no effective and unified standard for selecting treatment methods, which may waste a lot of time in the treatment process due to the selection of treatment methods. In addition, due to the different suggestions of doctors in different departments, the patients themselves are unable to make a decision. On the one hand, the selection of treatment methods may lead to delays in the disease. On the other hand, the final selected treatment method may not be the best, leading to worsening of the disease.

[0004] Therefore, how to provide a tumor diagnosis and treatment auxiliary system and method to assist tumor patients or doctors in choosing treatment methods during the treatment process, improve the efficiency of treatment decision-making, and avoid disease delays. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent tumor diagnosis and treatment assistance system and method based on a large language model, so as to assist tumor patients or doctors in selecting treatment methods during the treatment process, improve the decision-making efficiency of treatment methods, and avoid disease delays.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0007] In the first aspect, an intelligent tumor diagnosis and treatment auxiliary method based on a large language model is provided, comprising the following steps:

[0008] S1: creating a tumor computer simulation system, and obtaining first CT image data of a tumor of a first tissue at different positions at time point T0, inputting the first CT image data into the computer simulation system, and performing a three-dimensional simulation of the tumor in the first tissue through the computer simulation system to obtain a first three-dimensional model, and calculating the tumor volume in the first three-dimensional model, and determining whether the volume of the tumor is greater than a preset volume threshold, if so, preliminarily selecting a first treatment method, if not, executing step S2;

[0009] S2: cutting the first three-dimensional model into eight pieces using three planes perpendicular to each other in a computer simulation system, and calculating the tumor volume in each piece of the model after cutting;

[0010] S3: acquiring second CT image data of a tumor of the first tissue at different positions at time point T1, and inputting the second CT image data into the computer simulation system, performing three-dimensional simulation of the tumor of the tissue through the computer simulation system to obtain a second three-dimensional model;

[0011] S4: cutting the second three-dimensional model in a computer simulation system, cutting the second three-dimensional model into eight pieces using the three planes perpendicular to each other in step S2, and calculating the tumor volume in each piece of the model after cutting;

[0012] S5: Calculate the tumor growth rate in each block model according to the corresponding relationship between the eight block models in step S2 and the eight block models in step S4, and compare the tumor growth rate in each block model with the preset growth rate threshold to determine whether there is a block model whose tumor growth rate is greater than the preset growth rate threshold, if not, adjust the first treatment method to the second treatment method, if yes, maintain the first treatment method;

[0013] S6: For the second tissue with the tumor, execute steps S1-S5 in sequence, select a treatment method for the second tissue with the tumor based on steps S1-S5, and select a treatment method for the remaining tissues with tumors.

[0014] Preferably, the first treatment method is surgical resection treatment, and the second treatment method is radiotherapy, chemotherapy or immunotherapy.

[0015] Preferably, the calculation formula for the tumor volume V0 in a certain block of the model in step S1 is:

[0016] V0=abc*π;

[0017] Among them, a is the longest horizontal diameter of the tumor in the model, b is the longest vertical diameter of the tumor in the model, and c is the height of the tumor in the model.

[0018] Preferably, the method for calculating the tumor volume in a certain block of the model in step S1 is to divide the tumor volume in a certain block of the model into micro-areas in the three-dimensional model, where the micro-areas are cubes of specified sizes, and the number of cubes after the micro-area division is counted, and the number of cubes is multiplied by the volume of the cube to obtain the tumor volume of the model, and the side length of the cube for micro-area division is less than or equal to 10um.

[0019] Preferably, a tumor prediction method is also included, comprising the following process:

[0020] Historical tumor data is obtained and input into the created tumor development prediction model for model training. The tumor data in each tissue of the current tumor patient at each time point is obtained and input into the tumor development prediction model. The tumor development prediction model outputs corresponding tumor development prediction data, including the direction and rate of tumor spread.

[0021] Preferably, the eight models obtained by cutting the first three-dimensional model into eight pieces using three planes that are perpendicular to each other respectively represent eight directions of the first three-dimensional model, and the tumor growth rate in each model calculated based on the corresponding relationship between the eight models in step S2 and the eight models in step S4 respectively represents the tumor growth rate of the first tissue in eight directions.

[0022] Preferably, the formula for calculating the tumor growth rate in a certain model in step S5 is:

[0023] R=[V0*log(V1 / V0)] / [(T1-T0)*log2];

[0024] Among them, R is the tumor growth rate of a certain model, V1 is the tumor volume of a certain model at time point T1, V0 is the tumor volume of a certain model at time point T0, and T1 and T0 are two time points respectively.

[0025] In a second aspect, an intelligent tumor diagnosis and treatment assistance system based on a large language model is provided, which is used to implement any one of the intelligent tumor diagnosis and treatment assistance methods based on a large language model, including a computer simulation model creation module, a computer simulation model, a data acquisition module, a model cutting module, a calculation module, a data analysis module, and a treatment method selection module; the computer simulation model creation module is connected to the computer simulation model, the computer simulation model is connected to the model cutting module, the model cutting module is connected to the calculation module, the calculation module is connected to the data analysis module, and the data analysis module is connected to the treatment method selection module;

[0026] The computer simulation model creation module is used to create the computer simulation model;

[0027] The computer simulation model is used to perform three-dimensional simulation of tumors in tissues based on CT image data and obtain a three-dimensional model of the tumor;

[0028] The data acquisition module is used to acquire CT image data of the tumor in the tissue and transmit it to the computer simulation model;

[0029] The model cutting module is used to cut the three-dimensional tumor model obtained by three-dimensional simulation of the computer simulation model;

[0030] The calculation module is used to calculate the tumor volume in the three-dimensional model, and calculate the tumor growth rate in each model according to the volume of the corresponding model at different time points;

[0031] The data analysis module is used to perform data analysis on the tumor growth rate in each model obtained based on the tumor volume in the three-dimensional model obtained by the calculation module and the volume of the corresponding model at different time points, and match the corresponding treatment method.

[0032] Preferably, it also includes a tumor prediction model, which is used to receive tumor data in various tissues of the current tumor patient at various time points and output corresponding tumor development prediction data, wherein the tumor development prediction data includes the diffusion direction and diffusion rate of the tumor.

[0033] The beneficial effects of the present invention include:

[0034] The intelligent tumor diagnosis and treatment auxiliary system and method based on a large language model provided by the present invention create a tumor computer simulation system and perform three-dimensional simulation of the tumor in the first tissue based on CT image data to obtain a first three-dimensional model, calculate its tumor volume, and when the tumor volume is greater than a preset volume threshold, select a first treatment method, otherwise the first three-dimensional model is cut in the computer simulation system and the tumor volume in each model is calculated respectively; CT image data of the tumor of the first tissue at another time point is obtained, and the tumor of the tissue is three-dimensionally simulated through the computer simulation system, and the second three-dimensional model is cut after being obtained, and the tumor volume and tumor growth rate in each model after cutting are calculated respectively, and the tumor growth rate in each model is compared with the preset growth rate threshold, and the treatment method is adjusted according to the growth rate.

[0035] First, a three-dimensional simulation of the tumor in each tissue of the tumor patient is performed through the created computer simulation model to obtain a three-dimensional model of the tumor, a three-dimensional visualization of the tumor in each tissue is performed, and the volume of the tumor is calculated based on the three-dimensional model. Then, a preliminary selection of the treatment method is made based on the volume of the corresponding tissue, and then the three-dimensional model of the tumor is cut in various directions, and the growth rate of the model tumor in different blocks is calculated to obtain the growth rate of the tumor in each direction in the tissue, and the initial treatment method is adjusted or maintained based on the growth rate. The final treatment method is selected based on the specific data of the tumor, which not only improves the decision-making efficiency of the treatment method and avoids delays in the disease, but also improves the rationality of the treatment method.

[0036] Secondly, by cutting the three-dimensional model of the tumor in various directions and calculating the tumor growth rates of different blocks of the model, the growth rates of the tumor in the tissue in various directions are obtained, which provides a certain data basis for grasping the direction of tumor spread and provides direction for subsequent treatment.

[0037] Secondly, the tumor development prediction model is trained based on historical tumor data to obtain the tumor data in each tissue of the current tumor patient at each time point and input the data into the tumor development prediction model. The tumor development prediction model outputs corresponding tumor development prediction data, including the direction and rate of tumor spread, to further provide a data basis and treatment direction for the pre-selection of treatment plans.

[0038] Finally, when a cancer patient has tumors in multiple tissues, the tumors in different tissues are treated separately according to the process of steps S1-S5, and the treatment method for the tumor in each tissue is obtained, so that the tumor in each tissue can be treated in a targeted manner according to its specific situation, thereby promoting the treatment of the tumor in each tissue. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flowchart of the intelligent tumor diagnosis and treatment auxiliary method based on a large language model of the present invention.

[0040] Figure 2 Schematic diagram of the principle of the tumor development prediction model of the present invention.

[0041] Figure 3 Schematic diagram of the architecture of the intelligent tumor diagnosis and treatment assistance system based on a large language model of the present invention. DETAILED DESCRIPTION

[0042] The following is combined with Figure 1~Figure 3 The present invention is further described in detail:

[0043] Example 1

[0044] See attached Figure 1 As shown, the intelligent tumor diagnosis and treatment auxiliary method based on the large language model includes the following steps:

[0045] S1: Create a tumor computer simulation system, and obtain first CT image data of different positions of a tumor of a first tissue at time point T0, input the first CT image data into the computer simulation system, and perform three-dimensional simulation of the tumor in the first tissue through the computer simulation system to obtain a first three-dimensional model, and calculate the tumor volume in the first three-dimensional model, and determine whether the volume of the tumor is greater than a preset volume threshold. If so, preliminarily select a first treatment method, and if not, execute step S2. By performing a three-dimensional visualization of the tumor in the tissue through the computer simulation system, patients and doctors can intuitively understand the patient's tumor condition.

[0046] S2: Cut the first three-dimensional model into eight pieces using three planes perpendicular to each other in the computer simulation system, and calculate the tumor volume in each piece of the model after cutting. The reason for cutting the three-dimensional tumor model into eight pieces using three planes perpendicular to each other is that tumor data in various directions can be obtained, and further tumor data in various directions can be obtained.

[0047] S3: Obtain the second CT image data of the tumor of the first tissue at different positions at the T1 time point, and input the second CT image data into the computer simulation system, and perform three-dimensional simulation of the tumor of the tissue through the computer simulation system to obtain a second three-dimensional model. The T1 time point is a certain time away from the T0 time point, so the tumor data at the T1 time point is obtained and compared with the tumor data at the T0 time point to obtain the change of the tumor, and then based on the subsequent cutting and analyzing the corresponding cutting model for data analysis, the tumor development in all directions of the tumor can be known to provide a rigorous data basis for the subsequent treatment direction selection.

[0048] S4: Cut the second three-dimensional model in the computer simulation system, and cut the second three-dimensional model into eight pieces using the three two-by-two perpendicular planes in step S2, and calculate the tumor volume in each piece of the model after cutting. The volume change is obtained based on the tumor volume of each piece of the model after cutting, and a data basis is provided for the subsequent calculation of the total growth rate of the tumor and the tumor growth data in each direction.

[0049] S5: Calculate the tumor growth rate in each block of the eight models in step S2 and the eight models in step S4 according to the corresponding relationship, and compare the tumor growth rate in each block of the model with the preset growth rate threshold to determine whether there is a block of the model whose tumor growth rate is greater than the preset growth rate threshold. If not, adjust the first treatment method to the second treatment method. If so, maintain the first treatment method. Make a preliminary selection of the treatment method based on the volume of the corresponding tissue, then cut the three-dimensional model of the tumor in various directions, and calculate the tumor growth rate of the models of different blocks to obtain the growth rate of the tumor in the tissue in various directions, and adjust the preliminary treatment method based on the growth rate or maintain the preliminary treatment method. The final treatment method is selected based on the specific data of the tumor, which not only improves the decision-making efficiency of the treatment method and avoids delays in the disease, but also improves the rationality of the treatment method.

[0050] S6: For the second tissue with a tumor, execute steps S1-S5 in sequence, select a treatment method for the second tissue with a tumor based on steps S1-S5, and select a treatment method for the remaining tissues with tumors. If a tumor patient has tumors in multiple tissues, the tumors in different tissues are treated separately according to the process of steps S1-S5, and the treatment method for the tumor in each tissue is obtained, so as to achieve targeted treatment of the tumor in each tissue according to its specific situation. The first treatment method is a surgical resection treatment method, and the second treatment method is a radiotherapy and chemotherapy treatment method or an immunotherapy treatment method.

[0051] In this embodiment, the calculation formula for the tumor volume V0 in a certain block of the model in step S1 is:

[0052] V0=abc*π;

[0053] Among them, a is the longest horizontal diameter of the tumor in the model, b is the longest vertical diameter of the tumor in the model, and c is the height of the tumor in the model. The calculation formula of the tumor can be applied to tumors of any shape and is universal, which improves the adaptability and flexibility of volume calculation of tumors of various tissues.

[0054] In this embodiment, the tumor computer simulation system in step S1 performs three-dimensional simulation of the tumor in the tissue based on the generative adversarial network. The specific process is: first, the CT image data of the tumor is formatted and saved separately as an image file in PNG format, then the tumor lesion area is uniformly marked as white, and the original tumor CT image is combined to form a paired image, and the pix2pix architecture of the generative adversarial network is used to simulate and generate a three-dimensional model of the tumor lesion. The generative adversarial network includes a generator and a discriminator. The generator receives the CT image after the tumor lesion area is marked, and generates artificial samples close to the distribution of real data through the processing of a multi-layer neural network. The goal of the generator is to deceive the discriminator by generating realistic data. The discriminator distinguishes whether the input data is real or generated by the generator. The discriminator receives real samples and generated samples and tries to distinguish them. Its parameters are updated by back propagation to improve the accuracy of the discrimination. The discriminator strives to distinguish between real data and generated data to determine the authenticity of the input data. The generator attempts to generate realistic data to deceive the discriminator, while the discriminator strives to distinguish between real data and generated data. By continuously adjusting parameters, the generator and discriminator gradually improve their performance. Eventually, the generator can generate high-quality realistic images, while the discriminator can accurately judge the authenticity of the images, and finally generate a three-dimensional model of tumor lesions.

[0055] Example 2

[0056] On the basis of Example 1, the method for calculating the tumor volume in a certain block of the model in step S1 is to divide the tumor volume in a certain block of the model into micro-areas in the three-dimensional model, wherein the micro-areas are cubes of a specified size, and the number of cubes after micro-area division is counted, and the number of cubes is multiplied by the volume of the cube to obtain the tumor volume of the model, and the side length of the cube for micro-area division is less than or equal to 10um. The reason for micro-area division of the tumor is that for tumors of various tissues, in most cases, their shapes are irregular, so for the irregular characteristics of the tumor, the tumor can be divided into several tiny areas, which are cubes, and the smaller the size of the cube, the more accurate the tumor volume obtained based on the micro-area division will be. In this embodiment, the side length of the cube for micro-area division is less than or equal to 10um, and the size of the micro-area is selected to be so small because it is necessary to ensure the accuracy of the tumor volume obtained by micro-area division and improve the accuracy of the volume result.

[0057] Example 3

[0058] On the basis of Example 1 or Example 2, a tumor prediction method is also included, and the tumor prediction method includes the following process:

[0059] The historical tumor data is obtained and input into the created tumor development prediction model for model training. The tumor data in each tissue of the current tumor patient at each time point is obtained and input into the tumor development prediction model. The tumor development prediction model outputs corresponding tumor development prediction data, including the direction and rate of tumor spread. The tumor development prediction model is used to train the model based on the historical tumor data. The tumor data in each tissue of the current tumor patient at each time point is obtained and input into the tumor development prediction model. The tumor development prediction model outputs corresponding tumor development prediction data, including the direction and rate of tumor spread, to further provide a data basis and treatment direction for the pre-selection of treatment plans. For the principle of the tumor development prediction model, see Figure 2 As shown, the tumor development prediction model extracts features from data through multiple feature extraction layers to predict the tumor development.

[0060] In this embodiment, the amount of historical tumor data obtained must be sufficient to ensure that the subsequent training enables the model to have the required accuracy. The historical tumor data is preprocessed, including removing noise data and invalid data, supplementing missing data, and then converting the data to obtain a format that the model can understand or recognize. Then input the tumor development prediction model to train it. After the training is completed, another part of the preprocessed historical tumor data is input for testing to test whether the trained tumor development prediction model has the required prediction ability. After the test, a small amount of historical tumor data is input again to adjust the behavior of the tumor development prediction model. This process uses the stochastic gradient descent algorithm to update the parameters of the model to adapt to the prediction of tumor development. The predicted content includes the direction of tumor growth, growth rate, etc.

[0061] Example 4

[0062] On the basis of Example 1 or Example 2 or Example 3, the eight models obtained by cutting the first three-dimensional model into eight pieces by three planes perpendicular to each other respectively represent the eight directions of the first three-dimensional model, and the tumor growth rate in each model calculated according to the corresponding relationship between the eight models in step S2 and the eight models in step S4 respectively represents the tumor growth rate of the first tissue in the eight directions. The formula for calculating the tumor growth rate in a certain model in step S5 is:

[0063] R=[V0*log(V1 / V0)] / [(T1-T0)*log2];

[0064] Among them, R is the tumor growth rate of a certain model, V1 is the tumor volume of a certain model at the T1 time point, V0 is the tumor volume of a certain model at the T0 time point, and T1 and T0 are two time points respectively. Based on the above formula, the overall growth rate of the tumor can be calculated by obtaining the tumor volume at at least two time points, or the growth rate of each model can be calculated by obtaining the tumor volume of each segmented model at at least two time points to obtain the tumor growth rate in each direction. Then, through the tumor growth rate in each direction, the direction of tumor spread can be known, and then the tissue in that direction can be diagnosed and treated in advance, so as to take measures in advance to avoid lesions in the tissue in that direction.

[0065] During the entire process of diagnosis and treatment of cancer patients, doctors from various disciplines or departments can also jointly analyze the aforementioned measured data, jointly intervene in the comprehensive diagnosis and treatment of cancer patients, and promote the stabilization or improvement of the cancer patients' condition.

[0066] See also Figure 3 , an intelligent tumor diagnosis and treatment assistance system based on a large language model, is used to implement any one of the intelligent tumor diagnosis and treatment assistance methods based on a large language model, including a computer simulation model creation module, a computer simulation model, a data acquisition module, a model cutting module, a calculation module, a data analysis module, and a treatment method selection module; the computer simulation model creation module is connected to the computer simulation model, the computer simulation model is connected to the model cutting module, the model cutting module is connected to the calculation module, the calculation module is connected to the data analysis module, and the data analysis module is connected to the treatment method selection module.

[0067] A computer simulation model creation module is used to create a computer simulation model; the computer simulation model is used to perform three-dimensional simulation of tumors in tissues based on CT image data and obtain a three-dimensional model of the tumor; a data acquisition module is used to obtain CT image data of tumors in tissues and transmit them to the computer simulation model; a model cutting module is used to cut the three-dimensional model of the tumor obtained by three-dimensional simulation of the computer simulation model; a calculation module is used to calculate the tumor volume in the three-dimensional model and calculate the tumor growth rate in each model according to the volume of the corresponding model at different time points; a data analysis module is used to perform data analysis on the tumor growth rate in each model obtained based on the tumor volume in the three-dimensional model obtained by the calculation module and the volume of the corresponding model at different time points, and match the corresponding treatment method.

[0068] The intelligent tumor diagnosis and treatment assistance system based on a large language model also includes a tumor prediction model, which is used to receive tumor data in various tissues of the current tumor patient at various time points and output corresponding tumor development prediction data, which includes the tumor's diffusion direction and diffusion rate.

[0069] In summary, the intelligent tumor diagnosis and treatment auxiliary system and method based on a large language model provided by the present invention creates a tumor computer simulation system and performs three-dimensional simulation of the tumor in the first tissue based on CT image data to obtain a first three-dimensional model, calculates its tumor volume, and when the tumor volume is greater than a preset volume threshold, selects a first treatment method, otherwise the first three-dimensional model is cut in the computer simulation system and the tumor volume in each model is calculated respectively; CT image data of the tumor of the first tissue at another time point is obtained, and the tumor of the tissue is three-dimensionally simulated through the computer simulation system, and the second three-dimensional model is cut after being obtained, and the tumor volume and tumor growth rate in each model after cutting are calculated respectively, and the tumor growth rate in each model is compared with the preset growth rate threshold, and the treatment method is adjusted according to the growth rate.

[0070] Through the computer simulation model created, the tumors in various tissues of the tumor patient are simulated in three dimensions to obtain the three-dimensional model of the tumor, and the tumors in each tissue are visualized in three dimensions, and the tumor volume is calculated based on the three-dimensional model. Then, the treatment method is initially selected based on the volume of the corresponding tissue, and then the three-dimensional model of the tumor is cut in various directions, and the growth rate of the model tumors in different blocks is calculated to obtain the growth rate of the tumor in various directions in the tissue, and the initial treatment method is adjusted or maintained based on the growth rate. The final treatment method is selected based on the specific data of the tumor, which not only improves the decision-making efficiency of the treatment method, avoids delays in the disease, but also improves the rationality of the treatment method. The three-dimensional model of the tumor is cut in various directions, and the growth rate of the model tumors in different blocks is calculated to obtain the growth rate of the tumor in various directions in the tissue, which provides a certain data basis for grasping the direction of tumor spread and provides direction for subsequent treatment. The tumor development prediction model is trained based on historical tumor data, and the tumor data in each tissue of the current tumor patient at each time point is obtained and input into the tumor development prediction model. The tumor development prediction model outputs corresponding tumor development prediction data, including the direction and rate of tumor spread, further providing a data basis and treatment direction for the pre-selection of treatment plans.

Claims

1. An intelligent tumor diagnosis and treatment auxiliary method based on a large language model, characterized in that: The following steps are involved: S1: creating a tumor computer simulation system, and obtaining first CT image data of a tumor of a first tissue at different positions at time point T0, inputting the first CT image data into the computer simulation system, and performing a three-dimensional simulation of the tumor in the first tissue through the computer simulation system to obtain a first three-dimensional model, and calculating the tumor volume in the first three-dimensional model, and determining whether the volume of the tumor is greater than a preset volume threshold, if so, preliminarily selecting a first treatment method, if not, executing step S2; S2: cutting the first three-dimensional model into eight pieces using three planes perpendicular to each other in a computer simulation system, and calculating the tumor volume in each piece of the model after cutting; S3: acquiring second CT image data of a tumor of the first tissue at different positions at time point T1, and inputting the second CT image data into the computer simulation system, performing three-dimensional simulation of the tumor of the tissue through the computer simulation system to obtain a second three-dimensional model; S4: cutting the second three-dimensional model in a computer simulation system, cutting the second three-dimensional model into eight pieces using the three planes perpendicular to each other in step S2, and calculating the tumor volume in each piece of the model after cutting; S5: Calculate the tumor growth rate in each block model according to the corresponding relationship between the eight block models in step S2 and the eight block models in step S4, and compare the tumor growth rate in each block model with the preset growth rate threshold to determine whether there is a block model whose tumor growth rate is greater than the preset growth rate threshold, if not, adjust the first treatment method to the second treatment method, if yes, maintain the first treatment method; S6: for the second tissue with the tumor, sequentially executing steps S1-S5, selecting a treatment method for the second tissue with the tumor based on steps S1-S5, and selecting a treatment method for the remaining tissues with the tumor; The calculation formula of the tumor volume V0 in a certain block of the model in step S1 is: V0=abc*π; Among them, a is the longest horizontal diameter of the tumor in the model, b is the longest vertical diameter of the tumor in the model, and c is the height of the tumor in the model; The method for calculating the tumor volume in a certain block of the model in step S1 is to divide the tumor volume in a certain block of the model into micro-areas in the three-dimensional model, wherein the micro-areas are cubes of a specified size, and the number of cubes after the micro-area division is counted, and the number of cubes is multiplied by the volume of the cube to obtain the tumor volume of the model, wherein the side length of the cube for the micro-area division is less than or equal to 10 um; The formula for calculating the tumor growth rate in a certain block of the model in step S5 is: R=[V0*log(V1 / V0)] / [(T1-T0)*log2]; Among them, R is the tumor growth rate of a certain model, V1 is the tumor volume of a certain model at the T1 time point, V0 is the tumor volume of a certain model at the T0 time point, and T1 and T0 are two time points respectively; Also included are tumor prediction methods, including the following processes: Historical tumor data is obtained and input into the created tumor development prediction model for model training. The tumor data in each tissue of the current tumor patient at each time point is obtained and input into the tumor development prediction model. The tumor development prediction model outputs corresponding tumor development prediction data, including the direction and rate of tumor spread.

2. The intelligent tumor diagnosis and treatment auxiliary method based on a large language model according to claim 1 is characterized in that: The first treatment method is surgical resection treatment, and the second treatment method is radiotherapy, chemotherapy or immunotherapy.

3. The intelligent tumor diagnosis and treatment auxiliary method based on a large language model according to claim 1 is characterized in that: The eight models obtained by cutting the first three-dimensional model into eight pieces using three planes that are perpendicular to each other respectively represent eight directions of the first three-dimensional model. The tumor growth rates in each model calculated based on the corresponding relationship between the eight models in step S2 and the eight models in step S4 respectively represent the tumor growth rates of the first tissue in eight directions.

4. An intelligent tumor diagnosis and treatment assistance system based on a large language model, used to implement an intelligent tumor diagnosis and treatment assistance method based on a large language model as described in any one of claims 1 to 3, characterized in that: It includes a computer simulation model creation module, a computer simulation model, a data acquisition module, a model cutting module, a calculation module, a data analysis module, and a treatment method selection module; The computer simulation model creation module is connected to the computer simulation model, the computer simulation model is connected to the model cutting module, the model cutting module is connected to the calculation module, the calculation module is connected to the data analysis module, and the data analysis module is connected to the treatment method selection module; The computer simulation model creation module is used to create the computer simulation model; The computer simulation model is used to perform three-dimensional simulation of tumors in tissues based on CT image data and obtain a three-dimensional model of the tumor; The data acquisition module is used to acquire CT image data of the tumor in the tissue and transmit it to the computer simulation model; The model cutting module is used to cut the three-dimensional tumor model obtained by three-dimensional simulation of the computer simulation model; The calculation module is used to calculate the tumor volume in the three-dimensional model, and calculate the tumor growth rate in each model according to the volume of the corresponding model at different time points; The data analysis module is used to perform data analysis on the tumor growth rate in each model obtained based on the tumor volume in the three-dimensional model obtained by the calculation module and the volume of the corresponding model at different time points, and match the corresponding treatment method.

5. The intelligent tumor diagnosis and treatment auxiliary system based on a large language model according to claim 4 is characterized in that: It also includes a tumor prediction model, which is used to receive tumor data in various tissues of the current tumor patient at various time points and output corresponding tumor development prediction data, wherein the tumor development prediction data includes the diffusion direction and diffusion rate of the tumor.

Images