Target region automatic external expansion method for radiotherapy

Abstract

The invention discloses a target region automatic external expansion method for radiotherapy, and the method comprises the steps: S1, obtaining a CT image of a target external expansion region, determining the positions of a plurality of target regions of the target external expansion region and the position of an anatomical mark point corresponding to each target region according to the CT image, and determining the positions of a starting slice and an ending slice of each target region; s2, calculating the contour range of each target region; s3, setting an expansion coefficient array, determining the slice number of each target region according to the position of the target region ending slice, and dividing each target region into a plurality of continuous slice regions along the height direction according to the slice number; and S4, according to the expansion coefficient array and the data in the target region contour range set, performing external expansion on the plurality of slice regions in each target region to generate a target region slice region set. According to the method, multi-partition non-uniform external expansion is carried out on the complex target region, so that the risk that the target region at one end is close to high-risk normal tissues and cannot be integrally and continuously expanded due to uniform external expansion of the target region is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of radiotherapy, and particularly relates to an automatic external expansion method for a target area in radiotherapy. Background Art

[0002] In radiotherapy planning, the accurate definition of the radiotherapy target volume (PTV) is a key technical link affecting the tumor control rate and the protection of normal tissues. The traditional PTV definition method is based on standard operating specifications and adopts a geometric uniform external expansion mode, that is, the clinical target volume (CTV) is isotropically expanded by a fixed distance (usually 3-5 mm) to cover uncertain factors such as organ movement and setup errors.

[0003] However, in clinical practice, especially in the field of radiotherapy for head and neck tumors, the limitations of this method are becoming increasingly prominent: First, restricted by complex anatomical features such as the bony structure of the skull base, hollow organs, and important neurovascular bundles, uniform external expansion easily leads to overdose of the organs at risk (OARs); Second, the invasive growth pattern of tumor cells has significant directional heterogeneity characteristics. For example, biological behaviors such as nasopharyngeal carcinoma spreading along the foramina of the skull base and oropharyngeal carcinoma infiltrating through fascial spaces require that the target area expansion should reflect spatial heterogeneity; Third, the traditional method does not fully consider the differences in the organ movement amplitude of different anatomical subregions and the deformation characteristics between fractions, resulting in insufficient target area coverage or excessive irradiation of normal tissues in some cases. Summary of the Invention

[0004] To solve the problem that the uniform external expansion of the radiotherapy target area in the prior art has limitations, the present invention proposes an automatic external expansion method for a target area in radiotherapy.

[0005] The specific technical solution is as follows: An automatic external expansion method for a target area in radiotherapy, the steps include:

[0006] S1: Obtain the CT image of the target external expansion area, determine the positions of multiple target areas in the target external expansion area and the positions of the corresponding anatomical marker points for each target area according to the CT image, and determine the positions of the starting slice and the ending slice of each target area;

[0007] S2: Calculate the contour range of each target area, and statistically generate a set of target area contour ranges for each target area;

[0008] S3: Set an expansion coefficient array, determine the number of slices of each target area according to the position of the ending slice of the target area, and divide each target area into several continuous slice regions along the height direction according to the number of slices;

[0009] S4: According to the data in the expansion coefficient array and the set of target region contour ranges, expand each of the several slice regions in each target region respectively to generate a set of target region slice regions PTV.

[0010] Furthermore, the types of the target regions include high-risk region CTV1, medium-risk region CTV2, nasopharyngeal primary tumor region GTVnx, and lymph node metastatic region GTVnd. The high-risk region CTV1 and the medium-risk region CTV2 are clinical target regions.

[0011] Furthermore, the anatomical landmark points of each target region divide the current target region into an upper segment Super and a lower segment Infer along the cross-section of the target region height. The upper segment Super of the target region is uniformly expanded, and the lower segment Infer of the target region is non-uniformly expanded.

[0012] Furthermore, the expansion interval for the upper segment Super of the target region is 1 mm.

[0013] Furthermore, the number of slices of each target region is the same as the number of slices of the lower segment Infer of the target region.

[0014] Furthermore, the expansion coefficient array is an ascending array, and several slice regions in any target region are expanded non-decreasingly starting from the starting slice position.

[0015] Furthermore, the calculation formula for the number of slices of each target region is:

[0016] floor = ceil((z_max - cut_slice) / 5), where floor represents the number of slices of the current target region, z_max represents the position of the end slice of the current target region, and cut_slice represents the position of the anatomical landmark point of the current target region.

[0017] Furthermore, several sub-slice regions of each target region are equally spaced. The several sub-slice regions include:

[0018] Region 0: (z_min, cut_slice);

[0019] Region 1: (cut_slice, cut_slice + floor);

[0020] Region 2: (cut_slice + floor, cut_slice + 2*floor);

[0021] Region 3: (cut_slice + 2*floor, cut_slice + 3*floor);

[0022] Region 4: (cut_slice + 3*floor, cut_slice + 4*floor);

[0023] Region 5: (cut_slice + 4*floor, z_max);

[0024] Wherein, z_min represents the position where the current target area starts slicing.

[0025] Furthermore, the outer expansion of each target area in step S4 is carried out synchronously, and the expansion interval of the outer expansion of each target area is determined by the expansion coefficient array.

[0026] Furthermore, step S4 also includes: integrating the PTV set into a four-dimensional array ptv_all, converting the four-dimensional array ptv_all into DICOM RT format, and uploading it to the cloud directory for storage.

[0027] The above technical solution has the following advantages or technical effects:

[0028] 1. The present invention can adapt to the non-uniform outer expansion requirements of complex anatomical structures, has the ability to adjust three-dimensional non-uniform boundaries, can perform non-uniform adaptive outer expansion on high-risk clinical target areas (CTV), and aims at the complex anatomical structures unique to head and neck tumors (such as the target area of nasopharyngeal carcinoma adjacent to sensitive organs such as the brainstem and optic nerve), avoiding the risk that one end of the target area approaches high-risk normal tissues and cannot continue to expand due to uniform outer expansion of the target area.

[0029] 2. The outer expansion method of the present invention can achieve multi-target area coordination, ensure spatial consistency guarantee, and automatically generate an optimized outer expansion path that meets dosimetric constraints.

[0030] 3. In the present invention, this method can intelligently match different expansion coefficients according to the risk levels and spatial topological relationships of different target areas, can dynamically adjust the CTV-PTV outer expansion coefficients of each anatomical partition, realize "individualized non-uniform outer expansion", and minimize the risk brought by the outer expansion of the target area to the normal tissues (OAR) that need to be protected. Description of the Drawings

[0031] Figure 1 is the method flow chart of the present invention;

[0032] Figure 2 is the schematic diagram of the outer expansion of the target area slice region of the present invention;

[0033] Figure 3 is the schematic diagram of non-uniform outer expansion of the present invention;

[0034] Figure 4 is the layer coloring schematic diagram of the 3D modeling of the outer expansion of the target area slice region of the present invention. Detailed Embodiment

[0035] To make the technical solution of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0036] As Figure 1 shown, a method for automatically expanding the target area for radiotherapy includes the following steps:

[0037] S1: Obtain the CT image of the target expansion area, determine the positions of multiple target areas in the target expansion area and the positions of the corresponding anatomical marker points cut_slice for each target area according to the CT image, and determine the positions of the starting slice z_min and the ending slice z_max for each target area.

[0038] S2: Calculate the contour range of each target area, and statistically generate a set of target area contour ranges for each target area.

[0039] S3: Set an expansion coefficient array, determine the number of slices for each target area according to the position of the ending slice of the target area, and divide each target area into several continuous slice regions along the height direction according to the number of slices.

[0040] S4: According to the data in the expansion coefficient array and the set of target area contour ranges, expand each of the several slice regions in each target area respectively, and perform morphological expansion on different target areas using different sizes of expansion coefficients to generate a set of target area slice regions PTV.

[0041] In the prior art, for the expansion of the clinical target volume (CTV), uniform expansion is usually adopted, that is, a fixed value is used for expansion in the up-down, front-back, left-right directions. Since the target areas of nasopharyngeal carcinoma patients are relatively complex, for different target areas, their prescription doses are different, and the total prescription doses of different target areas span from the low-dose area (44 - 54 Gy) to the high-dose area (66 - 70 Gy). The conventional uniform expansion method in the prior art cannot cope with complex structures. At the same time, the upper part of the target area is close to the optic nerve, and the expansion space for high-risk normal tissues such as the skull base is limited, while there is more expansion space away from this position. Therefore, the present invention adopts a non-uniform target area automatic expansion method, and according to the anatomical marker points and the characteristics of the target area at different slice positions, morphological expansion is performed using different sizes of structural elements to automatically generate a radiotherapy target volume PTV (Planning Target Volume) that is more in line with complex structures and requirements.

[0042] Obtain the CT image from the HDF5 file, and determine multiple target areas at the target position in the CT image, such as Figure 2As shown, it includes multiple target regions determined based on the CT images of the target area of a patient. The types of target regions include the high-risk region CTV1, the medium-risk region CTV2, the primary nasopharyngeal tumor region GTVnx, and the lymph node metastatic region GTVnd. Among them, the high-risk region CTV1 and the medium-risk region CTV2 are clinical target volumes. Figure 2 As shown in the legend, from top to bottom are CTV1 and two CTV2s in sequence. Among them, CTV1 includes one GTVnx and two GTVnds, and one CTV2 includes one GTVnd. For different target regions, external expansion is performed to generate PTV1, PTV2, PTVX, and PTVD, corresponding to CTV1, CTV2, GTVnx, and GTVnd respectively. The PTV range in the figure is the range of the end slice.

[0043] As Figure 3 As shown, the anatomical landmark points of each target region divide the current target region into an upper segment Super and a lower segment Infer along the cross-section of the target region height. The upper segment Super of the target region is uniformly externally expanded, and its external expansion result generates PTV_Super. The expansion interval for the external expansion of the upper segment Super of the target region can be 1 mm. The lower segment Infer of the target region is non-uniformly externally expanded step by step from top to bottom, and its external expansion result generates PTV_Infer. The overall PTV can be conical in shape. The PTV set of the target region slice area is the set of all PTV_Infer and PTV_Super of the target regions. When the height of the anatomical landmark point moves down, the height of the lower segment Infer of the target region decreases accordingly. At this time, the overall PTV is in the shape of a cylinder-cone composite; when the anatomical landmark point coincides with the end slice, the overall PTV is cylindrical, and this is uniform external expansion.

[0044] In any target region, several slice regions are externally expanded non-decreasingly starting from the starting slice position. The number of slices of each target region is the number of slices of the lower segment Infer of the target region.

[0045] As an embodiment of the present invention, the calculation formula for the number of slices of each target region is:

[0046] floor = ceil((z_max - cut_slice) / 5), where floor represents the number of slices of the current target region, z_max represents the position of the end slice of the current target region, and cut_slice represents the position of the anatomical landmark point of the current target region. The actual number of slices can be adjusted according to actual needs.

[0047] Several sub-slice regions of each target region are equally spaced. The several sub-slice regions include:

[0048] Region 0: (z_min, cut_slice);

[0049] Region 1: (cut_slice, cut_slice + floor);

[0050] Region 2: (cut_slice + floor, cut_slice + 2*floor);

[0051] Region 3: (cut_slice + 2*floor, cut_slice + 3*floor);

[0052] Region 4: (cut_slice + 3*floor, cut_slice + 4*floor);

[0053] Region 5: (cut_slice + 4*floor, z_max);

[0054] Where z_min represents the starting position of slicing for the current target area.

[0055] The outer expansion of each target area is carried out synchronously, so that the outer expansion of multiple target areas has spatial consistency. The expansion coefficient array is an ascending array, and the expansion coefficient array can be [3, 5, 7, 9, 11]. The expansion interval for the outer expansion of different sub-slicing areas of each target area is determined by the expansion coefficient array, and different expansion coefficients are assigned to regions 1 to 5 in ascending order. At the same time, the expansion coefficient array is also used to adapt to the differences in the fixation effects brought by different body position fixation methods. By assigning different expansion coefficients, the fixation deviation is corrected to ensure that there is no deviation in the overall dose of the target area. By dynamically adjusting the CTV-PTV expansion coefficients of each anatomical region, the risk brought by the outer expansion of the target area to the normal tissues (OAR) that need to be protected is minimized.

[0056] Perform morphological expansion on each target area based on the original target area contour. Create a structural element kernel=np.ones((n, n)) for each target area. The expansion process is automated using a Python script, and the dilation operation is implemented using the dilate function of OpenCV in it: dilation = cv2.dilate(original contour, kernel). The original contour is obtained from the set of target area contour ranges. After the outer expansion of each target area is completed, the stratified color-coded schematic diagram of the 3D modeling of the outer expansion of each PTV target area slice region is as Figure 4 shown. Integrate the PTV set into a four-dimensional array ptv_all, ptv_all = stack(PTV1, PTV2, PTVx, PTVd). Convert the four-dimensional array ptv_all to the DICOM RT format and upload it to the cloud directory for storage.

[0057] The present invention provides a method for automatically expanding the target area for radiotherapy of nasopharyngeal carcinoma, which solves the problem that traditional uniform expansion cannot adapt to complex dose distributions through non-uniform morphological dilation. Based on CT images, the positions of multiple target areas and the anatomical landmark point cut_slice are identified, and the starting slice z_min and the ending slice z_max are divided. The target area is divided into an upper segment Super and a lower segment Infer along the height direction. The lower segment Infer is dynamically segmented into multiple equidistant sub-regions, and an ascending expansion coefficient array is used to achieve increasing expansion from top to bottom, forming a conical or cylindrical-conical composite PTV structure. The function is automatically called through a Python script to synchronously perform spatial consistency expansion on each target area. The method of the present invention further dynamically adjusts the expansion shape in combination with anatomical landmark points, and corrects the body position fixation deviation through differential structural elements, significantly improving the dose coverage accuracy of complex target areas.

[0058] The above embodiments only represent several implementation manners of the present invention. The description is relatively specific and detailed, but it cannot be understood as a limitation on the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several deformations and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention should be subject to the appended claims.

Claims

1. An automatic external expansion method for a target area in radiotherapy, characterized in that the steps Including: S1: Obtain the CT image of the target expansion region, determine the positions of multiple target areas in the target expansion region and the positions of the anatomical marker points corresponding to each target area, and determine the positions of the starting slice and the ending slice of each target area; S2: Calculate the contour range of each target area, and statistically generate a set of target area contour ranges from the contour ranges of each target area; S3: Set an expansion coefficient array, determine the number of slices of each target area according to the position of the ending slice of the target area, and divide each target area into several continuous slice regions along the height direction according to the number of slices; S4: Expand each of the several slice regions in each target area respectively according to the data in the expansion coefficient array and the set of target area contour ranges, and generate a set of target area slice regions PTV.

2. The automatic external expansion method for a target area used in radiotherapy according to claim 1, characterized in that, The types of the target areas include a high-risk region CTV1, a medium-risk region CTV2, a primary nasopharyngeal tumor region GTVnx, and a lymph node metastatic region GTVnd, and the high-risk region CTV1 and the medium-risk region CTV2 are clinical target areas.

3. A method for automatically expanding a target area for radiotherapy according to claim 1, characterized in that The anatomical marker point of each target area divides the current target area into an upper segment Super and a lower segment Infer along the cross-section of the height of the target area. The upper segment Super of the target area is uniformly expanded, and the lower segment Infer of the target area is non-uniformly expanded.

4. The automatic external expansion method for a target area used in radiotherapy according to claim 3, characterized in that, The expansion interval for expanding the upper segment Super of the target area is 1 mm.

5. A method for automatically expanding a target area for radiotherapy according to claim 3, characterized in that The number of slices of each target area is the same as the number of slices of the lower segment Infer of the target area.

6. A method for automatically expanding a target area for radiotherapy according to claim 1, characterized in that, The expansion coefficient array is an ascending array, and several slice regions in any target area are expanded non-decreasingly starting from the starting slice position.

7. A method for automatically expanding a target area for radiotherapy according to claim 5, characterized in that The calculation formula for the number of slices of each target area is: floor = ceil((z_max - cut_slice) / 5), where floor represents the number of slices of the current target area, z_max represents the position of the ending slice of the current target area, and cut_slice represents the position of the anatomical marker point of the current target area.

8. A method for automatically expanding a target area for radiotherapy according to claim 7, characterized in that Each of the several sub-slice regions of each target area is equally divided, and the several sub-slice regions include: Region 0: (z_min, cut_slice); Region 1: (cut_slice, cut_slice + floor); Region 2: (cut_slice + floor, cut_slice + 2*floor); Region 3: (cut_slice + 2*floor, cut_slice + 3*floor); Region 4: (cut_slice + 3*floor, cut_slice + 4*floor); Region 5: (cut_slice + 4*floor, z_max); where z_min represents the position of the starting slice of the current target area.

9. A method for automatically expanding a target area for radiotherapy according to claim 1, characterized in that, In step S4, the expansion of each target area is carried out synchronously, and the expansion interval for expanding each target area is determined by the expansion coefficient array.

10. A method for automatically expanding a target area for radiotherapy according to claim 1, characterized in that, The step S4 further includes: integrating the PTV set into a four-dimensional array ptv_all, converting the four-dimensional array ptv_all into the DICOM RT format, and uploading it to a cloud directory for storage.

Images