Breast cancer automatic radiotherapy plan design method and system based on OVH

Abstract

The invention relates to the technical field of medical image processing and radiotherapy, in particular to an OVH-based breast cancer automatic radiotherapy plan design method and system. Constructing a historical database, and integrating the image data of the patient, the delineation structure of the target region and the organ at risk and the dose information; calculating an OVH index, and quantifying a spatial geometrical relationship between the breast cancer tumor target region and the key endangered organ; and extracting a target area external expansion distance Lx corresponding to the specific volume x of the organ in the OVH curve and a dose value Dx corresponding to the volume point in the dose-volume histogram. And constructing a function model between Lx and Dx through linear regression analysis. And for a new patient, the system calculates the Lx value of the organ of interest of the new patient, inputs the value into the linear model, and predicts the corresponding radiotherapy dose Dx. According to the method provided by the invention, intelligent aided design of the breast cancer radiotherapy plan can be realized, and a scientific and acceptable dose prediction basis is provided.

Description

Technical Field

[0001] The present invention relates to the technical field of medical image processing and radiotherapy, and in particular to a method and system for designing an automatic radiotherapy plan for breast cancer based on Overlap Volume Histogram (OVH). Background Art

[0002] As an important part of the comprehensive treatment of breast cancer, radiotherapy (RT) is mainly used to reduce the risk of local recurrence in patients after mastectomy and prolong their survival. It is also an important palliative treatment for patients with locally advanced and metastatic breast cancer.

[0003] Currently, the mainstream radiotherapy technology is intensity-modulated radiotherapy (IMRT). IMRT achieves high-dose irradiation of the tumor target while protecting adjacent normal organs by driving the dynamic changes in the multileaf collimator and the beam dose rate, superimposed on the dynamic rotation of the linear accelerator gantry. The difficulties faced by IMRT for breast cancer lie primarily in two aspects: First, the shape of the tumor target (crescent-shaped, spindle-shaped), the delineation range (whether it includes internal mammary lymph nodes, whether it includes the supraclavicular region), and whether it includes a boost zone vary significantly among different breast cancer patients, and these factors significantly impact radiotherapy plan design. Second, radiotherapy plans currently primarily utilize inverse optimization algorithms, and unreasonable initial conditions can increase the difficulty of plan design and significantly increase optimization time. Furthermore, radiotherapy plan design primarily relies on the experience of physicists and planners. The quality of plan design and evaluation standards vary significantly between centers, making standardization impossible. This creates uncertainty and impacts the effectiveness of radiotherapy for breast cancer. Summary of the Invention

[0004] This invention addresses the shortcomings of existing technologies by providing an OVH-based automated breast cancer radiotherapy planning method and system based on overlapping volume histograms. This method uses overlapping volume histogram data to calculate the optimal OAR optimization objective function for the patient. This method not only rapidly assists in radiotherapy planning but also ensures the quality and consistency of radiotherapy plans, thereby reducing the risks of breast cancer radiotherapy.

[0005] Overlap volume histogram (OVH) is a descriptor of shape relationship, which is used to quantify the spatial position relationship between organs at risk and tumor target areas during radiotherapy. The breast tumor target area is close to the lung, heart, humeral head, and contralateral healthy breast. The relevant guidelines for breast cancer planning design have strict restrictions on these organs at risk, and therefore put forward high requirements for the planning design technology. Breast cancer planning after breast-conserving or modified radical mastectomy mainly uses tangent field-based intensity-modulated radiotherapy or hybrid intensity-modulated radiotherapy, which can maximize the protection of patients' organs at risk. Traditional dose-volume histograms (DVH) can only quantify the relationship between the irradiated dose and volume of normal tissue, and cannot effectively provide the impact of its spatial position relationship with the target area on the dose. In the present invention, we use OVH to quantify the spatial position relationship between the dose of organs at risk such as the lung and heart on the ipsilateral side of breast cancer radiotherapy and the target area, construct an automatic dose prediction model, and use it to guide the automatic design and optimization of radiotherapy plans.

[0006] In order to achieve the above object of the invention, the technical solution adopted by the present invention is as follows:

[0007] An OVH-based automatic radiotherapy planning method for breast cancer includes the following steps:

[0008] S1: A standardized breast cancer radiotherapy history database was constructed, categorized by surgery type (radiotherapy after breast-conserving surgery vs. radiotherapy after mastectomy) and lesion location (left or right breast). The database included information such as radiotherapy target volume, delineation of organs at risk (OARs), radiotherapy dose, and irradiated volume.

[0009] S2: Based on the RT Structure file generated by the radiotherapy planning system, the overlapping volume histogram (OVH) is used to calculate the spatial relationship between the radiotherapy target volume (PTV) and related organs at risk.

[0010] S3: Based on the RT Dose file and the outline structure file of the radiotherapy planning system, the dose volume histogram (DVH) data of each organ at risk is extracted;

[0011] S4: Based on the aforementioned OVH and DVH information, extract the target area expansion distance Lx corresponding to a specific volume point on the OVH map of the organ at risk of interest, and the dose value Dx corresponding to the same volume point on the DVH curve;

[0012] S5: With Lx as the horizontal coordinate and Dx as the vertical coordinate, draw a scatter plot and perform linear regression fitting to obtain the functional relationship between Lx and Dx;

[0013] S6: Use the constructed fitting function to predict the dose of radiotherapy plan for new patients, provide an optimization objective function, and assist in the realization of automated radiotherapy plan design.

[0014] Preferably, in step S1, the historical database needs to include standardized prescription dose limits, uniformly named structural contour information, and complete radiotherapy DICOM file data, including the patient's positioning CT image, three-dimensional structure file (RTStructure), three-dimensional dose distribution file (RTDose) and radiotherapy plan file (RTPlan) containing planning parameters.

[0015] Preferably, in step S2, the calculation of OVH depends on the three-dimensional contours of the target area and the organ at risk outlined in the imported structure file, and a self-written program is used to calculate the OVH curve of a certain organ to the target area according to the formula.

[0016] Preferably, in step S3, the organ name and ID in the RT Structure file are read, and the dvhcalc software package is called to read the dose data in the RT Dose file to calculate the differential DVH of the organ and further convert it into integral DVH data for subsequent analysis.

[0017] Preferably, in step S4, a Python program is used to automatically read the outer expansion distance Lx corresponding to a specified volume position in the OVH curve and the dose value Dx corresponding to the same volume point in the DVH curve. The read data is then plotted in a scatter plot with Lx as the horizontal axis and Dx as the vertical axis.

[0018] Preferably, in step S5, a linear regression analysis is performed on the above data using the Scipy software package to obtain a linear fitting function describing the relationship between Lx and Dx.

[0019] Preferably, in step S6, for a new patient, the three-dimensional structure of the target area and the organ at risk outlined by the doctor is obtained by simulating the positioning CT image, the OVH curve is calculated according to step S2, and the expansion distance L at a specific volume position is selected. X , substituting this value into the regression equation in step S5 to predict the dose value Dx. The predicted results can be used as the objective function parameters of the patient planning system to optimize the radiotherapy dose distribution. At the same time, the optimization results are fed back to the historical database to support continuous model updating.

[0020] The present invention also discloses an OVH-based automatic breast cancer radiotherapy planning design system, which can be used to implement the above-mentioned OVH-based automatic breast cancer radiotherapy planning design method, specifically including:

[0021] Data input module: used to input the patient's simulated positioning CT images and the target volume and OAR outline structure in RT Structure format;

[0022] OVH and DVH calculation module: Based on historical database data, it calculates the overlapping volume histogram of specific organs at risk and radiotherapy target areas, and reads RTdose files to extract the DVH curve of this organ;

[0023] Linear regression modeling module: plot Lx and Dx of all patients in the database and establish a linear regression model;

[0024] Historical database module: stores standardized radiotherapy information including patient CT, RT Structure, RT Dose, RT Plan, etc.

[0025] Radiotherapy dose prediction module: Based on the OVH data of a new patient's specific organ, a linear regression equation is input to calculate the dose volume parameters of this organ;

[0026] Patient data storage module: stores patient individualized radiotherapy plans and dose information;

[0027] Prediction result evaluation module: compares the predicted dose with the actual plan results, evaluates the model accuracy and optimizes the parameters

[0028] User interaction interface module: provides users with an interactive interface for displaying the OVH curve and DVH curve of the current patient's specific organ, the calculated predicted dose, and storing related data.

[0029] The present invention also discloses a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the automatic breast cancer radiotherapy plan is realized.

[0030] The present invention also discloses a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the automatic breast cancer radiotherapy planning method is implemented.

[0031] Compared with the prior art, the advantages of the present invention are:

[0032] (1) Using spatial geometry quantification to replace purely empirical target setting to achieve standardization of OAR dose prediction;

[0033] (2) Guide the optimization process, simplify manual trial and error iteration, and significantly improve planning efficiency and consistency;

[0034] (3) The model is lightweight and highly interpretable, avoiding the reliance of deep models on large amounts of labeled data;

[0035] (4) It is easy to integrate into the existing radiotherapy planning system and has good promotion and scalability. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 2 is a flow chart of an automatic radiotherapy plan design method for breast cancer based on OVH in an embodiment of the present invention.

[0037] Figure 2 2 is a schematic diagram of the OVH principle in an embodiment of the present invention.

[0038] Figure 3 This is an OVH image of the ipsilateral lung and the radiotherapy target area in an embodiment of the present invention.

[0039] Figure 4 This is an OVH image of the healthy lung and the radiotherapy target area in an embodiment of the present invention.

[0040] Figure 5 This is an OVH diagram of the heart and radiotherapy target area in an embodiment of the present invention.

[0041] Figure 6 The left lung L obtained based on OVH and DVH in the embodiment of the present invention 20 With D 20 Scatter plot and fitted curve.

[0042] Figure 7 The right lung L obtained based on OVH and DVH in the embodiment of the present invention 20 With D 20 Scatter plot and fitted curve.

[0043] Figure 8 is the heart L obtained based on OVH and DVH in the embodiment of the present invention 30 With D 30 Scatter plot and fitted curve.

[0044] Figure 9 It is a dose-volume histogram of the breast cancer radiotherapy plan designed initially based on experience.

[0045] Figure 10 It is the patient's final irradiated dose volume histogram calculated by the automatic radiotherapy planning system in the embodiment of the present invention.

[0046] Figure 11 4 is a diagram of a user interaction interface of a portable system in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and examples.

[0048] The present invention provides an OVH-based automatic radiotherapy plan design method for breast cancer based on overlapping volume histograms, such as Figure 1 As shown, including:

[0049] S1 Establish a historical database of radiotherapy plans for breast cancer patients.

[0050] S11 first distinguishes between breast-conserving therapy and post-mastectomy radiotherapy based on breast cancer examination reports and treatment records, and also distinguishes between left- and right-sided breast cancer. Based on this classification, a standardized prescription dose limit and a historical database of organs and radiotherapy techniques are established.

[0051] Each S12 database needs to contain radiotherapy plan data for breast cancer patients, including simulated CT positioning scans, RT Structure, a three-dimensional outline structure of the radiotherapy target area and organs at risk, RT Dose, a three-dimensional dose distribution data, and RT Plan files containing radiotherapy prescriptions.

[0052] S2 calculates the OVH data of the organ, the principle is as follows Figure 2 .

[0053] S21 imports CT images and the corresponding three-dimensional structural contours to determine the three-dimensional contours of the radiotherapy target volume (such as PTV) and the organs at risk of interest. If necessary, the image voxels can be resampled.

[0054] S22 The calculation formula of OVH in this invention is as follows:

[0055]

[0056] where T represents the tumor target, O is the organ at risk of interest, |O| represents the volume of organ O, d(p,T) is the shortest distance from voxel p to the edge of the tumor target, and |{p∈O|d(p,T)≤d}| is the volume occupied by voxels in the organ that are no more than d away from the tumor target.

[0057] S23 calculates the Euclidean distance d of each voxel of the organ at risk to the nearest point on the target surface. If the voxel overlaps with the target, the value of d is negative.

[0058] S24 sets the bin distance, traverses all OAR voxels, and assigns their corresponding distance values ​​to the corresponding bin. Calculate OVH for each bin O,T The result is the cumulative overlap volume ratio, such as Figure 3-Figure 5 shown.

[0059] S3 extracts the DVH data of the organ.

[0060] S31 reads the name and ID of a specific organ from the radiotherapy delineation data RT Structure. Based on this name and ID, the dvhcalc software package is used to read the differential DVH data from the three-dimensional dose distribution file RT Dose and convert it into integral DVH data.

[0061] S4 extracts Lx and Dx data.

[0062] S41 uses Python code to read the target expansion distance L corresponding to the organ at risk for a specific volume (such as 20% of the lung volume) on the OVH map 20 , and the dose D corresponding to the organ at risk (such as 20% volume of the lung) on ​​the DVH diagram 20 .

[0063] S42 draws a scatter plot of the data read above, with Lx as the horizontal coordinate and Dx as the vertical coordinate, as shown in FIG. Figure 6-8 shown.

[0064] S5 linear regression fitting analysis

[0065] S51 uses the Scipy software package to perform linear regression fitting analysis on the above data and obtains the equation description of the fitting curve, as shown in Figure 6-Figure 8 shown.

[0066] S6 for planning new patients

[0067] S61 obtains the simulated positioning CT of the new patient and the three-dimensional contour data of the target area and organs at risk drawn by the doctor, and uses the S2 step to calculate the corresponding OVH.

[0068] S62 calculates the distance L corresponding to the specific volume of the patient's organ at risk X Input the linear equation in step S5 to calculate and obtain the radiotherapy dose parameter Dx corresponding to the organ at risk under this volume.

[0069] S63 inputs this parameter into the radiotherapy planning system, uses it as the objective function to optimize the radiotherapy plan for this patient, and evaluates the quality of this automated plan based on the prescribed dose, and feeds this plan back to the database, such as Figure 10 shown.

[0070] The present invention also discloses an OVH-based automatic breast cancer radiotherapy planning system, which can be used to implement the above-mentioned automatic breast cancer planning, specifically including:

[0071] Data input module: used to input the patient's simulated positioning CT images and the target volume and OAR outline structure in RT Structure format;

[0072] OVH and DVH calculation module: Based on historical database data, it calculates the overlapping volume histogram of specific organs at risk and radiotherapy target areas, and reads RTdose files to extract the DVH curve of this organ;

[0073] Linear regression modeling module: plot Lx and Dx of all patients in the database and establish a linear regression model;

[0074] Historical database module: stores standardized radiotherapy information including patient CT, RT Structure, RT Dose, RT Plan, etc.

[0075] Radiotherapy dose prediction module: Based on the OVH data of a new patient's specific organ, a linear regression equation is input to calculate the dose volume parameters of this organ;

[0076] Patient data storage module: stores patient individualized radiotherapy plans and dose information;

[0077] Prediction result evaluation module: compares the predicted dose with the actual plan results, evaluates the model accuracy and optimizes the parameters

[0078] User interaction interface module: provides an interactive interface for users, such as Figure 11 As shown, it is used to display the OVH curve and DVH curve of the current patient's specific organ, the calculated predicted dose, and store related data.

Claims

1. The OVH-based automated breast cancer radiotherapy planning design method includes the following steps: S1: Build a standardized breast cancer radiotherapy history database, categorized by surgical type and lesion location. The database includes information on radiotherapy targets, outlined structures of organs at risk, radiotherapy doses, and irradiated volumes. S2: Based on the RT Structure file generated by the radiotherapy planning system, the spatial relationship between the radiotherapy target volume (PTV) and related organs at risk is calculated using the overlapping volume histogram (OVH). S3: Based on the 3D dose file RT Dose and the outline structure file of the radiotherapy planning system, the dose volume histogram DVH data of each organ at risk is extracted; S4: Based on the OVH and DVH information, extract the target area expansion distance Lx corresponding to a specific volume point on the OVH map of the organ at risk of interest, and the dose value Dx corresponding to the same volume point on the DVH curve; S5: With Lx as the horizontal coordinate and Dx as the vertical coordinate, draw a scatter plot and perform linear regression fitting to obtain the functional relationship between Lx and Dx; S6: Use the constructed fitting function to predict the dose of radiotherapy plan for new patients, provide an optimization objective function, and assist in the realization of automated radiotherapy plan design.

2. The OVH-based automatic breast cancer radiotherapy planning method according to claim 1, characterized in that: The historical database described in S1 must include standardized prescription dose limits, uniformly named structural contour information, and complete radiotherapy DICOM file data, including the patient's positioning CT image, three-dimensional structure file RT Structure, three-dimensional dose distribution file RT Dose, and radiotherapy plan file RT Plan containing planning parameters.

3. The OVH-based automatic breast cancer radiotherapy planning design method according to claim 1, characterized in that: The calculation of OVH in S2 depends on the three-dimensional contours of the target volume and organs at risk drawn in the imported structure file. The mathematical expression of OVH is as follows: where T represents the tumor target, O is the organ at risk of interest, |O| represents the volume of organ O, d(p,T) is the shortest distance from voxel p to the edge of the tumor target, and |{p∈O|d(p,T)≤d}| is the volume occupied by voxels in the organ that are no more than d away from the tumor target.

4. The OVH-based automatic breast cancer radiotherapy planning method according to claim 1, characterized in that: In S3, the organ name and ID in the RT Structure file are read, and the dvhcalc software package is called to read the dose data in the RT Dose file to calculate the differential DVH of the organ and further convert it into integral DVH data for subsequent analysis.

5. The OVH-based automatic breast cancer radiotherapy planning method according to claim 1, characterized in that: In S4, a Python program automatically reads the outer expansion distance Lx corresponding to a specified volume position in the OVH curve and the dose value Dx corresponding to the same volume point in the DVH curve. The read data is then plotted in a scatter plot with Lx as the horizontal axis and Dx as the vertical axis.

6. The OVH-based automatic breast cancer radiotherapy planning design method according to claim 1, characterized in that: In S5, the Scipy software package was used to perform linear regression analysis on the outward expansion distance Lx corresponding to the specified volume position in the OVH curve and the dose value Dx corresponding to the same volume point in the DVH curve to obtain a linear fitting function describing the relationship between Lx and Dx.

7. The OVH-based automatic breast cancer radiotherapy planning method according to claim 1, characterized in that: In S6, for new patients, the three-dimensional structure of the target area and organs at risk outlined by the doctor is obtained by simulating the positioning CT image, the OVH curve is calculated according to step S2, and the expansion distance L at a specific volume position is selected. X , and substitute this value into the regression equation in step S5 to predict the dose value Dx.

8. An OVH-based automated breast cancer radiotherapy planning system, suitable for implementing the method of any one of claims 1 to 7, specifically comprising: Data input module: used to input the patient's simulated positioning CT images and the target volume and OAR outline structure in RT Structure format; OVH and DVH calculation module: Based on historical database data, it calculates the overlapping volume histogram of specific organs at risk and radiotherapy target areas, and reads RTdose files to extract the DVH curve of this organ; Linear regression modeling module: plot Lx and Dx of all patients in the database and establish a linear regression model; Historical database module: stores standardized radiotherapy information including patient CT, RT Structure, RT Dose, and RT Plan; Radiotherapy dose prediction module: Based on the OVH data of a new patient's specific organ, a linear regression equation is input to calculate the dose volume parameters of this organ; Patient data storage module: stores patient individualized radiotherapy plans and dose information; Prediction result evaluation module: compares the predicted dose with the actual plan results, evaluates the model accuracy and optimizes the parameters User interaction interface module: provides users with an interactive interface for displaying the OVH curve and DVH curve of the current patient's specific organ, the calculated predicted dose, and storing related data.

9. A computer device, characterized in that: It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the OVH-based automatic radiotherapy planning design method for breast cancer according to one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, which, when executed by a processor, implements the OVH-based automatic breast cancer radiotherapy planning design method as described in one of claims 1 to 7.

Images