Quality control method based on electron radiation field image system and maximum tissue motif ratio

Abstract

The invention belongs to the technical field of radiotherapy quality control applying an electron radiation field image system, and discloses a quality control method based on the electron radiation field image system and the maximum tissue motif ratio. Improving the DRR image by combining the maximum tissue motif ratio (TMR) of radiotherapy equipment to obtain a TMR weighted DRR image; the TMR weighted DRR image and a flux graph generated by the radiotherapy plan are used as data to be input into a training improved U-net network, and then a prediction EPID graph is obtained; and analyzing and predicting the EPID graph based on the gamma index and an EPID graph actually generated by the radiotherapy equipment, and carrying out quality control on the radiotherapy equipment according to the gamma index. According to the method, the three-dimensional CT data are converted into the two-dimensional data by utilizing the maximum tissue motif ratio data acquired during equipment operation testing, the data need to be additionally collected, the accuracy can be relatively high after gamma analysis by utilizing an improved U-net network, and ranking errors in verification can be relatively easily found and distinguished.

Description

Technical Field

[0001] The present invention relates to the technical field of radiotherapy quality control using an electronic portal imaging system, and in particular to a quality control method based on an electronic portal imaging system and a maximum tissue phantom ratio. Background Art

[0002] EPID (Electronic Portal Imaging Device), or electronic portal imaging system, is an important imaging device in the field of modern radiotherapy. It uses a flat-panel detector to receive and record images of X-rays after they pass through the patient, thereby realizing the monitoring of the radiotherapy process and position verification. By applying appropriate quality control methods, EPID can be used to detect multiple performance parameters of radiotherapy equipment, such as output dose, flatness and symmetry of the radiation field, consistency of the light field and the radiation field, and the positioning accuracy of the multi-leaf grating. These parameters are crucial to ensure the accuracy and stability of the operation of radiotherapy equipment. By regularly using EPID for equipment quality control, problems with the equipment can be discovered and corrected in a timely manner, thereby ensuring the smooth progress of the radiotherapy process. EPID also has the function of real-time monitoring of radiation dose. During radiotherapy, EPID can monitor the changes in radiation dose in real time, thereby helping doctors to promptly discover and correct errors in the radiotherapy process. This helps to improve the accuracy and safety of radiotherapy and reduce the side effects and complications caused by radiotherapy.

[0003] The core of EPID quality control is to evaluate the accuracy and consistency of the radiotherapy process by comparing the difference between the actual EPID image and the expected or reference image. There are two existing methods for accurately predicting the transmission EPID response image based on the treatment plan data. One is the simulation method, which establishes a physical model between the accelerator, patient and EPID based on the principles of radiation physics. The model should be able to simulate the transmission, attenuation and scattering process of the radiation in the patient's body, as well as the response of the EPID to the radiation; the second is to use a machine learning algorithm to use the treatment plan data and the corresponding EPID response image to train the grid model so that it can learn the mapping relationship between the data.

[0004] The response of EPID is linear with the beam dose, but the changes in human tissue density, organ movement and volume increase the difficulty of prediction in actual work. At the same time, the image quality of transmission EPID is affected by factors such as ray energy, scattering, background noise, etc., resulting in low accuracy of simulation algorithms. The training data of existing machine learning methods usually uses treatment plan data and corresponding EPID response data, and a large amount of training data is required to establish a prediction model. However, in actual applications, it is difficult to obtain enough training data to support model training. Summary of the invention

[0005] The present invention proposes a quality control method based on an electronic portal imaging system and a maximum tissue phantom ratio, which aims to accurately predict the electronic portal imaging system (EPID) image to evaluate the performance of radiotherapy equipment and the execution effect of the radiotherapy plan. The method combines the maximum tissue phantom ratio (TMR) parameter of the radiotherapy equipment, improves the generation of digitally reconstructed radiographic images (DRR), and obtains a TMR-weighted DRR image. The TMR-weighted DRR image can more accurately simulate the dose distribution during radiotherapy. By generating a TMR-weighted DRR image based on a CT image, the flux data generated by the radiotherapy plan and the TMR-weighted DRR image are merged into a two-channel data as the input of an improved U-net network. In the training stage, a training data set containing a large number of actual radiotherapy cases is constructed, each case including two-channel input data and corresponding EPID image data. By training the improved U-net network, it can learn the dose distribution characteristics during radiotherapy and become a prediction model for accurately predicting EPID images. In the prediction stage, the trained improved U-net network is used to predict the new radiotherapy plan and generate a predicted EPID image. Then, the gamma index analysis method is used to compare the predicted EPID image with the EPID image actually generated by the radiotherapy equipment to evaluate the difference between the two. By comparing the results, it can be determined whether the performance of the radiotherapy equipment meets the requirements and whether the radiotherapy plan has been accurately implemented.

[0006] The technical solution of the present invention is as follows: a quality control method based on an electronic portal imaging system and a maximum tissue phantom ratio, combining the maximum tissue phantom ratio TMR of the radiotherapy equipment to improve the DRR image and obtain a TMR-weighted DRR image; the TMR-weighted DRR image and the flux map generated by the radiotherapy plan are merged into a two-channel data as the input of the improved U-net network; the improved U-net network is trained, and a predicted EPID map is obtained through the trained improved U-net network; based on the gamma index analysis, the predicted EPID map and the actual EPID map generated by the radiotherapy equipment are compared, and the quality control of the radiotherapy equipment is performed according to the gamma index.

[0007] When the relative positions of the radiation source and the detector of the radiotherapy equipment are fixed, the depth of the detector in the water is changed to obtain a series of measurement values ​​that change with depth. The ratio of the series of measurement values ​​that change with depth to the maximum measurement value is TMR; combined with TMR, a TMR-weighted DRR image is obtained.

[0008] The steps of obtaining the TMR-weighted DRR image are as follows:

[0009] Step 1: According to the CT value and electron density correspondence table, the CT value in the CT image is converted into relative electron density to generate a three-dimensional relative electron density volume data;

[0010] Step 2: According to the beam isocenter position of the radiotherapy plan, adjust the volume data of the three-dimensional relative electron density obtained in the first step so that its volume center is located at the beam isocenter position;

[0011] Step 3: Data format conversion and adjustment, so that the volume data size of the three-dimensional relative electron density in the second step is 128x128x128, and the voxel size is 3mm; at the same time, the flux map of each radiation field of the radiotherapy plan is generated according to the DCM data of the radiotherapy plan; the data size of the flux map after conversion is 128x128x1, and the pixel size is 3mm;

[0012] Step 4: Select the TMR parameters of the radiotherapy equipment to be quality controlled; select the TMR parameters under the corresponding radiation field of 10x10; when the radiation field of the radiotherapy plan is 0 degrees, the top layer of the beam direction of the three-dimensional relative electron density volume data in the third step is 0mm deep, and the top three-dimensional relative electron density volume data is multiplied by the corresponding TMR parameter value at 0mm to become the first layer of weighted volume data, and the following layers are analogous to obtain weighted volume data; accumulate the weighted volume data of each layer along the beam direction to obtain a TMR-weighted DRR image;

[0013] When the radiation field in the radiotherapy plan is not 0 degrees, the three-dimensional relative electron density volume data of the third step is rotated according to the radiation field angle. The first layer of the rotated three-dimensional relative electron density volume data is multiplied by the TMR parameter value corresponding to the thickness of the first layer, and the subsequent layers are deduced in the same way to obtain TMR-weighted volume data. The weighted volume data of each layer are accumulated to obtain the TMR-weighted DRR image.

[0014] The improved U-net network adopts a res U-net ++ network; the res U-net ++ network includes: a backbone block, four residual blocks, two receptive field blocks and two attention blocks;

[0015] The input data are the flux map of the field and the TMR-weighted DRR map of the corresponding field, and the output data label is the predicted EPID map of the corresponding beam after passing through the phantom at different angles;

[0016] The input data first passes through the backbone block, the filter size is 16, and the output of the backbone block is c1;

[0017] c1 passes through the first residual block, the filter size is 32, and the output of the first residual block is c2;

[0018] c2 passes through the second residual block with a filter size of 64, and the output of the second residual block is c3;

[0019] c3 passes through the first receptive field block, the filter size is 128, and the output result of the first receptive field block is b1;

[0020] b1 and c2 are used as the input of the first attention block. The output is d1a after upsampling with a convolution kernel of (2,2). After d1a is connected with c2, it passes through a third residual block with a filter size of 64, and the output is d1.

[0021] d1 and c1 pass through the second attention block, and the output result is d2a after upsampling with a convolution kernel of (2,2). After d2a is connected with c1, it passes through the fourth residual block with a filter size of 32, and the output result is d2;

[0022] d2 passes through a second receptive field block with a filter size of 16, and then passes through a convolution layer with a convolution kernel size of (1,1), an output channel of 1, and uses 'ReLu' activation. The activation result is used as the final output result.

[0023] The input data in the backbone block passes through two branches respectively, the first branch is a convolutional layer, a batch normalization layer, a ReLu activation layer, and a convolutional layer in sequence, and the second branch is a convolutional layer and a batch normalization layer in sequence; the outputs of the two branches are added together as the output of the backbone block;

[0024] The input data in the residual block passes through two branches respectively. The first branch is a batch normalization layer, a ReLu activation layer, a convolution layer with a convolution kernel of (3,3), a batch normalization layer, a ReLu activation layer, and a convolution layer with a convolution kernel of (3,3). The second branch is a convolution layer with a convolution kernel of (1,1) and a batch normalization layer. The outputs of the two branches are added together to form the output of the residual block.

[0025] In the receptive field block, according to the dose deposition characteristics of beam physics in tissue, the input data are respectively added after passing through four convolution layers with convolution kernels of (3,3) and batch normalization, and then passing through a convolution layer with a convolution kernel of (1,1) as the output of the receptive field block; among the four convolution layers, the first three convolution layers are hollow convolutions, and the expansion rates are 2, 3 and 5 respectively;

[0026] The data input of the attention block is divided into two parameters, which are named g and x in the attention block respectively; g passes through two batch normalization layers, a ReLu activation layer, a convolution layer with a convolution kernel of (3,3), and then passes through a maximum pooling layer, and the result is marked as g_pool; the result of x passing through two batch normalization layers, a ReLu activation layer, and a convolution layer with a convolution kernel of (3,3) is added to g_pool, and the result is marked as gc_sum; gc_sum passes through a batch normalization layer, a ReLu activation layer, and a convolution layer with a convolution kernel of (3,3) and then multiplied by x, and the result is used as the output of the attention block; the input of the first attention block is b1 and c2, b1 corresponds to g in the attention block, and c2 corresponds to x in the attention block; the input of the second attention block is d1 and c1, d1 corresponds to g in the attention block, and c1 corresponds to x in the attention block.

[0027] The improved U-net network uses the following beam sequence and measurement method to obtain training data, all beams illuminate the phantom, and the EPID system is used to synchronously obtain the actual EPID map:

[0028] Square field: including 1x1, 2x2, 3x3, 4x4, 5x5, 7x7, 10x10, 12x12, 15x15, 18x18 fields, the center of the field is the center of the EPID map; the first 5 fields, i.e. 1x1, 2x2, 3x3, 4x4, 5x5, are set to the center (5,0), (-5,0), (0,5), (0, -5), (5,5), (-5,5), (-5, -5), (5, -5); each field is 50MU, a total of 50 fields;

[0029] Strip field: 1x2, 1x5, 1x7, 1x10, 2x5, 2x7, 2x10, the beam isocenters of each field are set at (5, 0), (0, 0), (-5, 0) respectively; 2x1, 5x1, 7x1, 10x1, 5x2, 7x2, 10x2, the beam isocenters of each field are set at (0, 5), (0, 0), (0, -5) respectively; 50MU per field, 42 fields in total;

[0030] Intensity modulated radiation field: An intensity modulated radiation field with a radiation field size of 24x24 cm was selected and the radiotherapy plan was transplanted onto the quality control phantom, for a total of 40 cases and 360 beams.

[0031] Based on the gamma index analysis, when the gamma index value of the predicted EPID graph and the actual EPID graph is greater than 93%, the operation status of the radiotherapy equipment is normal; when the gamma index value is between 90-93%, the equipment operation status is manually confirmed; when the gamma index is less than 90%, the operation status of the radiotherapy equipment is abnormal, the radiotherapy equipment is stopped and the relevant operation parameters of the radiotherapy equipment are checked.

[0032] Beneficial effects of the present invention: Compared with the prior art, the method of the present invention has low difficulty in implementation, fast running speed, high accuracy, and can quickly and accurately identify positioning errors. This study uses the maximum tissue phantom ratio data collected during the equipment operation test to convert three-dimensional CT data into two-dimensional data. The data itself does not need to be collected additionally, and the conversion process is very fast under existing computer technology. An improved U-net network is used. The improved U-net network has a high running speed after being established. After gamma analysis, it can be found that the accuracy is high, and the ranking errors in verification can be easily discovered and identified. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 The flowchart is a quality control method based on the electronic portal imaging system and the maximum tissue phantom ratio;

[0034] Figure 2 This is the predicted EPID image with 5% dose deviation introduced;

[0035] Figure 3 Gamma analysis diagram of the actual EPID image and the predicted EPID image with 5% dose deviation;

[0036] Figure 4 Dose deviation diagram between the predicted EPID image and the actual EPID image with a blade deviation of 0.5-2mm introduced;

[0037] Figure 5 Gamma analysis diagram of actual EPID image and predicted EPID image of blade position deviation;

[0038] Figure 6 This is the backbone network structure diagram of the improved U-net network;

[0039] Figure 7 This is the residual block network structure diagram of the improved U-net network;

[0040] Figure 8 This is the attention block network structure diagram of the improved U-net network;

[0041] Fig. 9 This is the network structure diagram of the receptive field block of the improved U-net network;

[0042] Fig.10 This is the network structure diagram of the improved U-net network. DETAILED DESCRIPTION

[0043] In order to accurately predict the EPID image based on the radiotherapy plan, the present invention designs a quality control method based on the electronic portal imaging system and the maximum tissue phantom ratio. The maximum tissue phantom ratio (TMR) of the device is combined with the TMR parameters based on the CT image to generate a TMR-weighted DRR image. Based on the improved U-net network, the flux map generated by the radiotherapy plan and the TMR-weighted DRR image data are merged into a two-channel data as input, and the image data of the EPID data is used as output to construct a training data set. The trained improved U-net network performs prediction, and the difference between the predicted EPID map and the actual EPID map is analyzed based on the gamma index, thereby achieving quality control of the radiotherapy equipment.

[0044] A quality control method based on an electronic portal imaging system and a maximum tissue phantom ratio, combining the maximum tissue phantom ratio TMR of a radiotherapy device to improve a DRR image, and obtaining a TMR-weighted DRR image; merging the TMR-weighted DRR image and the flux map generated by the radiotherapy plan into a two-channel data as the input of an improved U-net network; training the improved U-net network, and obtaining a predicted EPID map through the trained improved U-net network; based on the gamma index analysis, the predicted EPID map and the actual EPID map generated by the radiotherapy device are compared, and the quality control of the radiotherapy device is performed according to the gamma index.

[0045] When the relative positions of the radiation source and the detector of the radiotherapy equipment are fixed, the depth of the detector in the water is changed to obtain a series of measurement values ​​that change with depth. The ratio of the series of measurement values ​​that change with depth to the maximum measurement value is TMR; combined with TMR, a TMR-weighted DRR image is obtained.

[0046] The steps of obtaining the TMR-weighted DRR image are as follows:

[0047] Step 1: According to the CT value and electron density correspondence table, the CT value in the CT image is converted into relative electron density to generate a three-dimensional relative electron density volume data;

[0048] Step 2: According to the beam isocenter position in the radiotherapy plan file, adjust the volume data of the three-dimensional relative electron density obtained in the first step so that its volume center is located at the beam isocenter;

[0049] Step 3: Data format conversion and adjustment, so that the volume data size of the three-dimensional relative electron density in the second step is 128x128 x 128, and the voxel size is 3mm; at the same time, the flux map of each radiation field of the radiotherapy plan is generated according to the DCM data of the radiotherapy plan; the data size of the flux map after conversion is 128x128x1, and the pixel size is 3mm;

[0050] Step 4: Select the TMR parameters of the current radiotherapy equipment to be quality controlled; select the TMR parameters under the corresponding radiation field of 10x10. When the radiation field of the radiotherapy plan is 0 degrees, the top layer of the beam direction of the three-dimensional relative electron density volume data in the third step is 0mm deep. The top three-dimensional relative electron density volume data is multiplied by the corresponding TMR parameter value at 0mm to become the first layer of weighted volume data, and the following layers are analogous to obtain weighted volume data; accumulate the weighted volume data of each layer along the beam direction to obtain a TMR-weighted DRR image;

[0051] When the radiation field in the radiotherapy plan is not 0 degrees, the three-dimensional relative electron density volume data of the third step is rotated according to the radiation field angle. The first layer of the rotated three-dimensional relative electron density volume data is multiplied by the TMR parameter value corresponding to the thickness of the first layer, and the subsequent layers are deduced in the same way to obtain TMR-weighted volume data. The weighted volume data of each layer are accumulated to obtain the TMR-weighted DRR image.

[0052] The improved U-net network adopts a res U-net ++ network; the res U-net ++ network includes: a backbone block, four residual blocks, two receptive field blocks and two attention blocks;

[0053] The input data is the flux map under the field and the TMR-weighted DRR map of the corresponding field, and the output data label is the predicted EPID image after the corresponding beam passes through the phantom at different angles;

[0054] Set the filter size to (16,32,64,128);

[0055] The input data first passes through the backbone block, the filter size is 16, and the output of the backbone block is c1;

[0056] c1 passes through the first residual block, the filter size is 32, and the output result of the first residual block is c2;

[0057] c2 passes through the second residual block with a filter size of 64, and the output of the second residual block is c3;

[0058] c3 passes through the first receptive field block, the filter size is 128, and the output result of the first receptive field block is b1;

[0059] b1 and c2 are used as the input of the first attention block. The output is d1a after upsampling with a convolution kernel of (2,2). After d1a is connected with c2, it passes through a third residual block with a filter size of 64, and the output is d1.

[0060] d1 and c1 pass through the second attention block, and the output result is d2a after upsampling with a convolution kernel of (2,2). After d2a is connected with c1, it passes through the fourth residual block with a filter size of 32, and the output result is d2;

[0061] d2 passes through a second receptive field block with a filter size of 16, and then passes through a convolution layer with a convolution kernel size of (1,1), an output channel of 1, and uses 'ReLu' activation. The activation result is used as the final output result.

[0062] The input data in the backbone block passes through two branches respectively, the first branch is a convolutional layer, a batch normalization layer, a ReLu activation layer, and a convolutional layer in sequence, and the second branch is a convolutional layer and a batch normalization layer in sequence; the outputs of the two branches are added together as the output of the backbone block;

[0063] The input data in the residual block passes through two branches respectively. The first branch is a batch normalization layer, a ReLu activation layer, a convolution layer with a convolution kernel of (3,3), a batch normalization layer, a ReLu activation layer, and a convolution layer with a convolution kernel of (3,3). The second branch is a convolution layer with a convolution kernel of (1,1) and a batch normalization layer. The outputs of the two branches are added together to form the output of the residual block.

[0064] In the receptive field block, according to the dose deposition characteristics of beam physics in tissue, the input data is added after passing through four convolution layers with a convolution kernel of (3,3) and batch normalization, and then passing through a convolution layer with a convolution kernel of (1,1) as the output of the receptive field block; among the four convolution layers, the first three convolution layers are hollow convolutions, and the expansion rates are 2, 3 and 5 respectively;

[0065] The data input of the attention block is two parameters, so they are named g and x in the attention block respectively; g passes through two batch normalization layers, ReLu activation layers, and convolution layers with a convolution kernel of (3,3) and then passes through a maximum pooling layer, and the result is marked as g_pool; x passes through two batch normalization layers, ReLu activation layers, and convolution layers with a convolution kernel of (3,3) and the result is added to g_pool, and the result is marked as gc_sum; gc_sum passes through a batch normalization layer, a ReLu activation layer, and a convolution layer with a convolution kernel of (3,3) and then multiplied by x, and the result is used as the output of the attention block. The inputs of the first attention block are b1 and c2, b1 corresponds to g in the attention block, and c2 corresponds to x in the attention block. The inputs of the second attention block are d1 and c1, d1 corresponds to g in the attention block, and c1 corresponds to x in the attention block.

[0066] The improved U-net network uses the following beam sequence and measurement method to obtain the data required for model training. All beams irradiate the quality control phantom, and the actual EPID map is synchronously obtained using the EPID system:

[0067] Square field: including 1x1, 2x2, 3x3, 4x4, 5x5, 7x7, 10x10, 12x12, 15x15, 18x18 fields, the center of the field is the center of the EPID map; the first 5 fields, i.e. 1x1, 2x2, 3x3, 4x4, 5x5, are set to the center (5,0), (-5,0), (0,5), (0, -5), (5,5), (-5,5), (-5, -5), (5, -5); each field is 50MU, a total of 50 fields;

[0068] Strip field: 1x2, 1x5, 1x7, 1x10, 2x5, 2x7, 2x10, the beam isocenters of each field are respectively (5, 0), (0, 0), (-5, 0); 2x1, 5x1, 7x1, 10x1, 5x2, 7x2, 10x2, the beam isocenters of each field are respectively set at (0, 5), (0, 0), (0, -5); 50MU per field, a total of 42 fields;

[0069] Intensity modulated radiation field: An intensity modulated radiation field with a radiation field size of 24x24 cm was selected and the radiotherapy plan was transplanted onto the quality control phantom, for a total of 40 cases and 360 beams.

[0070] Based on the gamma index analysis, when the gamma index value of the predicted EPID graph and the actual EPID graph is greater than 93%, the radiotherapy equipment is operating normally. When the gamma index is between 90-93%, the equipment operating status is manually confirmed. When the gamma index is less than 90%, the radiotherapy equipment is operating abnormally. The radiotherapy equipment is stopped and the relevant operating parameters of the radiotherapy equipment are checked.

[0071] Furthermore, taking a specific embodiment as an example, its TMR parameters are obtained through Table 1, and the 10x10 TMR parameters are converted from the PDD (percent depth dose) when the SSD (source skin distance) = 100 cm, and the depth is 310 mm.

[0072] Table 1 Dose distribution with depth

[0073]

[0074] The corresponding table of CT value and electron density is as follows:

[0075] Table 2 CT value and relative electron density used in current measurement

[0076]

[0077] The specific steps of this embodiment are as follows:

[0078] Compared with the conventional DRR projection method, the DRR projection with fused TMR parameters is adopted to obtain the TMR-weighted DRR image.

[0079] The conventional DRR projection method is generally: the 3D CT volume data is generated through a mathematical simulation algorithm to generate a 2D image, which is called a DRR image. The basic method of generating a DRR image is the ray projection method. Define a point as a virtual X-ray source, and the X-ray passes through the object to be measured from the virtual ray source and is projected to the detector plane perpendicular to the central axis of the ray. The intersection of the ray and the plane determines the position of the DRR midpoint; in this process, the intersection of the point source and each CT slice in the CT volume data can be calculated to obtain the corresponding CT value. After a ray is projected, the CT value obtained by accumulating the entire path is the DRR image pixel value of the corresponding point on the detector; repeat the above steps. When all the rays are projected, the CT value of each pixel block of the entire DRR image will be obtained. Mapping its accumulated value to the pixel gray value will obtain a DRR image.

[0080] In this embodiment, when the radiation field of the radiotherapy plan is 0 degrees, the top layer of the beam direction of the three-dimensional relative electron density volume data is 0 mm deep, and the top layer of the three-dimensional relative electron density volume data is multiplied by the corresponding TMR parameter value at 0 mm to become the first layer of weighted volume data, and the following layers are deduced in the same way to obtain weighted volume data, as shown in Formula 1, where the 3x3 matrix represents the relative electron density matrix of the current layer, and 0.95 represents the TMR parameter value at the corresponding depth of the layer obtained from the TMR table; the weighted volume data of each layer are accumulated along the beam direction to obtain a TMR-weighted DRR image;

[0081] (1)

[0082] When the radiation field in the radiotherapy plan is not 0 degrees, the three-dimensional relative electron density volume data is rotated according to the radiation field angle. The first layer of the rotated three-dimensional relative electron density volume data is multiplied by the TMR parameter value corresponding to the thickness of the first layer, and the subsequent layers are deduced in the same way to obtain TMR-weighted volume data. The weighted volume data of each layer are accumulated to obtain the TMR-weighted DRR image.

[0083] The radiotherapy plan file includes: radiotherapy plan file RT Plan, radiotherapy outline file RT Structure, radiotherapy dose file RT Dose and radiotherapy plan CT file. All files are in DICOM format with the file suffix .DCM. The RT Plan file contains the radiation field information of the radiotherapy plan, such as the source-skin distance, the position of the beam isocenter, the intensity modulated radiation field size, and the position information of all multi-leaf collimator leaves.

[0084] Furthermore, the gamma index is a statistical method for comparing the difference between the calculated dose distribution and the actual measured dose distribution. It comprehensively considers the distance between the calculated point and the measured point (Distance to Agreement, DTA) and the degree of difference between the dose values ​​(Dose Difference, DD). The present invention defines a gamma index of 2%, and if the 2mm average value exceeds 93%, it is considered that the radiotherapy plan data can be used.

[0085] To artificially add errors to the radiotherapy plan, in this example, it is assumed that the output dose is adjusted by 5%. Figure 2 This is the predicted EPID image with 5% dose deviation introduced; Figure 3 The gamma index diagram of the predicted EPID response and the actual EPID image has a gamma pass rate of 70%, which is much lower than the threshold of 90% that requires human intervention in the commonly used clinical plan evaluation. This example illustrates that the present invention can effectively detect the output dose deviation of the radiotherapy equipment.

[0086] To artificially add errors to the treatment plan, in this embodiment, it is assumed that the blade positions have a deviation of 0.5-2mm. Figure 4 is the dose deviation map, i.e., the dose difference between the corresponding positions of the predicted EPID response and the actual EPID image; Figure 5 The gamma index diagram of the predicted EPID response and the actual EPID image has a gamma pass rate of 53%, which is much lower than the threshold of 90% that requires human intervention in the commonly used clinical plan evaluation. This example illustrates that the present invention can effectively detect the position deviation of the multi-leaf collimator of the radiotherapy equipment.

[0087] In summary, the present invention does not simply use the traditional DRR projection algorithm, but combines the TMR parameters of the radiotherapy equipment for optimization. The TMR parameters reflect the dose response characteristics of the radiotherapy equipment under different tissue densities. By incorporating them into the DRR image generation method, the dose distribution during radiotherapy can be simulated more accurately, thereby improving the accuracy of dose prediction.

[0088] The present invention innovatively introduces deep learning technology, especially the U-net network, to learn the dose distribution characteristics during radiotherapy from the flux data generated by CT images and radiotherapy plans. By constructing a training data set containing a large number of actual radiotherapy cases and training the improved U-net network, it can accurately predict EPID images. This data-driven method can capture complex dose distribution patterns and improve the accuracy and robustness of predictions.

[0089] The present invention combines the flux data generated by the radiotherapy plan with the TMR-weighted DRR image data into a two-channel data set as the input of the deep learning model. This strategy makes full use of the information of the radiotherapy plan and the anatomical structure of the phantom, which helps the model better learn the dose distribution characteristics during radiotherapy.

[0090] In the prediction stage, the present invention uses the gamma index analysis method to compare the predicted EPID image with the actual EPID image to quantify the difference between the two. The gamma index is a widely used dose verification tool that can comprehensively consider dose differences and position differences to provide a comprehensive dose distribution assessment. By comparing the results, it is possible to accurately determine whether the performance of the radiotherapy equipment meets the requirements and whether the radiotherapy plan has been accurately implemented.

Claims

1. A quality control method based on an electronic portal imaging system and a maximum tissue phantom ratio, characterized in that: Improved DRR images based on the maximum tissue phantom ratio TMR of the radiotherapy equipment are used to obtain TMR-weighted DRR images; the TMR-weighted DRR images and the flux maps generated by the radiotherapy plan are merged into a two-channel data as the input of the improved U-net network; the improved U-net network is trained, and the predicted EPID map is obtained through the trained improved U-net network; the predicted EPID map is compared with the EPID map actually generated by the radiotherapy equipment based on the gamma index, and the quality control of the radiotherapy equipment is performed according to the gamma index.

2. The quality control method based on the electronic portal imaging system and the maximum tissue phantom ratio according to claim 1, characterized in that: When the relative positions of the radiation source and the detector of the radiotherapy equipment are fixed, the depth of the detector in the water is changed to obtain a series of measurement values ​​that change with depth. The ratio of the series of measurement values ​​that change with depth to the maximum measurement value is TMR; combined with TMR, a TMR-weighted DRR image is obtained.

3. The quality control method based on the electronic portal imaging system and the maximum tissue phantom ratio according to claim 2, characterized in that: The steps of obtaining the TMR-weighted DRR image are as follows: Step 1: According to the CT value and electron density correspondence table, the CT value in the CT image is converted into relative electron density to generate a three-dimensional relative electron density volume data; Step 2: According to the beam isocenter position of the radiotherapy plan, adjust the volume data of the three-dimensional relative electron density obtained in the first step so that its volume center is located at the beam isocenter position; Step 3: Data format conversion and adjustment, so that the volume data size of the three-dimensional relative electron density in the second step is 128x128x128, and the voxel size is 3mm; at the same time, the flux map of each radiation field of the radiotherapy plan is generated according to the DCM data of the radiotherapy plan; the data size of the flux map after conversion is 128x128x1, and the pixel size is 3mm; Step 4: Select the TMR parameters of the radiotherapy equipment to be quality controlled; select the TMR parameters under the corresponding radiation field of 10x10; when the radiation field of the radiotherapy plan is 0 degrees, the top layer of the beam direction of the three-dimensional relative electron density volume data in the third step is 0mm deep, and the top layer of the three-dimensional relative electron density volume data is multiplied by the corresponding TMR parameter value at 0mm to become the first layer of weighted volume data, and the following layers are deduced in the same way to obtain weighted volume data; The weighted volume data of each layer are accumulated along the beam direction to obtain the TMR-weighted DRR image; When the radiation field in the radiotherapy plan is not 0 degrees, the three-dimensional relative electron density volume data of the third step is rotated according to the radiation field angle. The first layer of the rotated three-dimensional relative electron density volume data is multiplied by the TMR parameter value corresponding to the thickness of the first layer, and the subsequent layers are deduced in the same way to obtain TMR-weighted volume data. The weighted volume data of each layer are accumulated to obtain the TMR-weighted DRR image.

4. The quality control method based on the electronic portal imaging system and the maximum tissue phantom ratio according to claim 1, characterized in that: The improved U-net network adopts res U-net++ network; The res U-net++ network includes: a backbone block, four residual blocks, two receptive field blocks and two attention blocks; The input data are the flux map of the field and the TMR-weighted DRR map of the corresponding field, and the output data label is the predicted EPID map of the corresponding beam after passing through the phantom at different angles; The input data first passes through the backbone block, the filter size is 16, and the output of the backbone block is c1; c1 passes through the first residual block, the filter size is 32, and the output of the first residual block is c2; c2 passes through the second residual block with a filter size of 64, and the output of the second residual block is c3; c3 passes through the first receptive field block, the filter size is 128, and the output result of the first receptive field block is b1; b1 and c2 are used as the input of the first attention block. The output is d1a after upsampling with a convolution kernel of (2,2). After d1a is connected with c2, it passes through a third residual block with a filter size of 64, and the output is d1. d1 and c1 pass through the second attention block, and the output result is d2a after upsampling with a convolution kernel of (2,2). After d2a is connected with c1, it passes through the fourth residual block with a filter size of 32, and the output result is d2; d2 passes through a second receptive field block with a filter size of 16, and then passes through a convolution layer with a convolution kernel size of (1,1), an output channel of 1, and uses 'ReLu' activation. The activation result is used as the final output result.

5. The quality control method based on the electronic portal imaging system and the maximum tissue phantom ratio according to claim 4, characterized in that: The input data in the backbone block passes through two branches respectively, the first branch is a convolutional layer, a batch normalization layer, a ReLu activation layer, and a convolutional layer in sequence, and the second branch is a convolutional layer and a batch normalization layer in sequence; the outputs of the two branches are added together as the output of the backbone block; The input data in the residual block passes through two branches respectively. The first branch is a batch normalization layer, a ReLu activation layer, a convolution layer with a convolution kernel of (3,3), a batch normalization layer, a ReLu activation layer, and a convolution layer with a convolution kernel of (3,3). The second branch is a convolution layer with a convolution kernel of (1,1) and a batch normalization layer. The outputs of the two branches are added together to form the output of the residual block. In the receptive field block, according to the dose deposition characteristics of beam physics in tissue, the input data are respectively added after passing through four convolution layers with convolution kernels of (3,3) and batch normalization, and then passing through a convolution layer with a convolution kernel of (1,1) as the output of the receptive field block; among the four convolution layers, the first three convolution layers are hollow convolutions, and the expansion rates are 2, 3 and 5 respectively; The data input of the attention block is divided into two parameters, which are named g and x in the attention block respectively; g passes through two batch normalization layers, a ReLu activation layer, a convolution layer with a convolution kernel of (3,3), and then passes through a maximum pooling layer, and the result is marked as g_pool; the result of x passing through two batch normalization layers, a ReLu activation layer, and a convolution layer with a convolution kernel of (3,3) is added to g_pool, and the result is marked as gc_sum; gc_sum passes through a batch normalization layer, a ReLu activation layer, and a convolution layer with a convolution kernel of (3,3) and then multiplied by x, and the result is used as the output of the attention block; the input of the first attention block is b1 and c2, b1 corresponds to g in the attention block, and c2 corresponds to x in the attention block; the input of the second attention block is d1 and c1, d1 corresponds to g in the attention block, and c1 corresponds to x in the attention block.

6. The quality control method based on the electronic portal imaging system and the maximum tissue phantom ratio according to claim 1, characterized in that: The improved U-net network uses the following beam sequence and measurement method to obtain training data, all beams illuminate the phantom, and the EPID system is used to synchronously obtain the actual EPID map: Square field: including 1x1, 2x2, 3x3, 4x4, 5x5, 7x7, 10x10, 12x12, 15x15, 18x18 fields, the center of the field is the center of the EPID map; the first 5 fields, i.e. 1x1, 2x2, 3x3, 4x4, 5x5, are set to the center (5,0), (-5,0), (0,5), (0, -5), (5,5), (-5,5), (-5, -5), (5, -5); each field is 50MU, a total of 50 fields; Strip field: 1x2, 1x5, 1x7, 1x10, 2x5, 2x7, 2x10, the beam isocenters of each field are set at (5, 0), (0, 0), (-5, 0) respectively; 2x1, 5x1, 7x1, 10x1, 5x2, 7x2, 10x2, the beam isocenters of each field are set at (0, 5), (0, 0), (0, -5) respectively; 50MU per field, 42 fields in total; Intensity modulated radiation field: An intensity modulated radiation field with a radiation field size of 24x24 cm was selected and the radiotherapy plan was transplanted onto the quality control phantom, for a total of 40 cases and 360 beams.

7. The quality control method based on the electronic portal imaging system and the maximum tissue phantom ratio according to claim 1, characterized in that: Based on the gamma index analysis, when the gamma index value of the predicted EPID graph and the actual EPID graph is greater than 93%, the operation status of the radiotherapy equipment is normal; when the gamma index value is between 90-93%, the equipment operation status is manually confirmed; when the gamma index is less than 90%, the operation status of the radiotherapy equipment is abnormal, the radiotherapy equipment is stopped and the relevant operation parameters of the radiotherapy equipment are checked.

Images