A matching device for radiotherapy positioning device

Abstract

The present invention provides a matching device for a radiotherapy positioning device, which establishes communication with a hospital imaging system, a processing device, and a doctor terminal respectively, and includes a reference database establishment module, a key part feature extraction module, a feature classification module, an optimal model solution module, a key feature extraction module to be matched, and a positioning device matching module. The present invention provides a matching device for a radiotherapy positioning device, which ensures that the positioning devices pre-set by the hospital have sufficient quantity and high-quality accuracy. When a tumor patient needs to match a positioning device, the key features to be matched in the medical imaging data of the patient to be matched are extracted, and it is only necessary to match the key features to be matched with all the optimal feature samples, so as to match the most suitable optimal positioning device as the radiotherapy positioning device, thereby improving the matching accuracy of the positioning device and ensuring the accuracy of positioning during radiotherapy.

Description

Technical Field

[0001] The invention relates to the field of radiotherapy, and in particular to a matching device for a radiotherapy positioning device. Background Art

[0002] Radiotherapy is one of the three major methods of treating tumors. According to statistics, about 65%-75% of tumor patients need radiotherapy. With the continuous development of radiotherapy technology, radiotherapy has entered the "three precisions" era, namely precise positioning, precise planning, and precise treatment. Among them, precise positioning is the basis for achieving precise planning and precise treatment.

[0003] During radiotherapy, cancer patients usually need a positioning device for auxiliary positioning. Currently, the positioning devices used in radiotherapy are mostly universal positioning devices with single or several fixed models pre-stored in the radiotherapy department. The models of these universal positioning devices are limited. During radiotherapy, doctors or technicians manually select a relatively suitable positioning device for cancer patients based on their own experience. Because the number of fixed positioning devices is too small and the selection for cancer patients is entirely based on manual experience, it is difficult to ensure that the selected positioning device is accurately matched to the patient, which can easily cause large errors. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a matching device for a radiotherapy positioning device, which can solve the problem that the current positioning device is selected for tumor patients from a small number of fixed positioning devices stored in the hospital entirely based on the doctor's manual experience, and it is difficult to ensure that the selected positioning device is accurately matched to the patient, which is easy to cause large errors.

[0005] The purpose of the present invention is achieved by the following technical solutions:

[0006] A matching device for a radiotherapy positioning device, wherein the matching device establishes communication with a hospital imaging system, a processing device, and a doctor terminal, respectively, and comprises the following modules:

[0007] A benchmark database establishment module is used to obtain diagnostic imaging data corresponding to a number of tumor patients who need radiotherapy and have undergone radiotherapy from the hospital imaging system, and store all diagnostic imaging data as benchmark imaging data in a preset benchmark database;

[0008] The key part feature extraction module is used to extract the key part features of all the reference images in the preset reference database, take each key part feature as a feature sample, and calculate the sample distance between any two feature samples according to the preset distance calculation rule;

[0009] A feature classification module, wherein the feature classification module uses a k-means algorithm to cluster all feature samples according to the sample distance to obtain a number of cluster feature sample sets, each of which contains a number of different feature samples;

[0010] An optimal model solving module, wherein the optimal model solving module solves the optimal feature model for each cluster feature sample set according to a preset optimal solution function to obtain a number of optimal feature samples, and the optimal model solving module inputs each optimal feature sample into a processing device to process a corresponding optimal positioning device;

[0011] A key feature extraction module to be matched, wherein the key feature extraction module to be matched obtains medical image data of the patient to be matched from the hospital imaging system, and extracts key features in the medical image data as key features to be matched;

[0012] A positioning device matching module, wherein the positioning device matching module is used to match the key features to be matched with all optimal feature samples according to preset matching rules, match the corresponding optimal feature samples as positioning feature samples, and use the optimal positioning device processed according to the positioning feature samples as a radiotherapy positioning device and send it to the doctor terminal for selecting a radiotherapy positioning device for the patient to be matched.

[0013] Furthermore, the matching of the key feature to be matched with all the optimal feature samples according to the preset matching rules is specifically as follows: calculating the distance between the key feature to be matched and each optimal feature sample and using it as the matching distance, judging whether the matching distance with the smallest value is less than the preset matching distance threshold, and if so, using the optimal feature sample corresponding to the smallest matching distance as the positioning feature sample, and using the optimal positioning device processed according to the positioning feature sample as the radiotherapy positioning device, and if not, using the medical image data of the turning point to be matched as the new benchmark image data and storing it in the preset benchmark database.

[0014] Furthermore, when the patient to be matched is a patient with a head tumor and needs to match a head positioning device, the diagnostic image data is a head diagnostic image, and the key part feature is a lower edge boundary feature of the head.

[0015] Furthermore, when the patient to be matched is a breast tumor patient and needs to match a radiotherapy positioning compensation model, the diagnostic image data is a chest diagnostic image containing the patient to be matched, and the key part feature is a chest contour feature.

[0016] Furthermore, each of the feature samples contains a number of coordinate points, and the preset distance calculation rule is specifically as follows: when calculating the distance between two feature samples, the distance between each coordinate point on one of the feature samples to the other feature sample is calculated respectively and used as the first distance, and the first distance with the largest value is used as the sample distance between the two feature samples.

[0017] Furthermore, the calculation process of the first distance is specifically as follows: taking the distance from a single coordinate point on one feature sample to all coordinate points on another feature sample as the second distance, and taking the second distance with the smallest value as the first distance.

[0018] Furthermore, when the patient to be matched is an oral tumor patient and needs to be matched with an oral stent, the diagnostic image data is an oral diagnostic image containing teeth, and the key part features are tooth features. The sample distance between any two feature samples is calculated according to the preset distance calculation rules as follows: the left endpoint distance and the right endpoint distance between the two feature samples are calculated respectively, and the average of the left endpoint distance and the right endpoint distance is used as the sample distance between the two feature samples.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows: a matching device for a radiotherapy positioning device in the present application collects reference image data, extracts key features of the reference image data, and calculates the sample distance between any two feature samples, classifies all feature samples according to the sample distance using the k-means algorithm, solves the optimal model for the feature samples of each cluster, obtains the optimal feature samples, and inputs the optimal feature samples corresponding to each cluster into a processing device to process the corresponding optimal positioning device, thereby ensuring that the positioning devices pre-set by the hospital have sufficient quantity and high-quality accuracy. When a tumor patient needs to match a positioning device, the key features to be matched in the medical image data of the patient to be matched are extracted, and it is only necessary to match the key features to be matched with all the optimal feature samples, thereby matching the most suitable optimal positioning device as the radiotherapy positioning device, thereby improving the matching accuracy of the positioning device and ensuring the accuracy of positioning during radiotherapy.

[0020] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings. The specific implementation of the present invention is given in detail by the following embodiments and their accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0022] Figure 1 A schematic diagram of the working process of a matching device for a radiotherapy positioning device of the present invention;

[0023] Figure 2 is a schematic diagram of reference image data in a matching device of a radiotherapy positioning device of the present invention when the positioning device to be matched is a head positioning device suitable for patients with head tumors;

[0024] Figure 3 A schematic diagram of key parts of a matching device of a radiotherapy positioning device of the present invention when the positioning device to be matched is a head positioning device suitable for patients with head tumors;

[0025] Figure 4 It is a schematic diagram of reference image data in a matching device of a radiotherapy positioning device of the present invention when the positioning device to be matched is an oral stent suitable for oral tumor patients;

[0026] Figure 5 The present invention is a schematic diagram of the body of a tumor patient after placing a radiotherapy positioning compensation mold in a matching device of a radiotherapy positioning device when the positioning device to be matched is a radiotherapy positioning compensation mold suitable for breast tumor patients. DETAILED DESCRIPTION

[0027] The present invention is further described below in conjunction with the accompanying drawings and specific implementation methods. It should be noted that, under the premise of no conflict, the various embodiments or technical features described below can be arbitrarily combined to form a new embodiment.

[0028] like Figure 1 As shown, a matching device of a radiotherapy positioning device in the present application, wherein the matching device establishes communication with a hospital imaging system, a processing device, and a doctor terminal respectively, comprises the following modules:

[0029] A reference database establishment module, the reference database establishment module is used to obtain diagnostic image data corresponding to a number of tumor patients who need radiotherapy and have undergone radiotherapy from the hospital imaging system, and store all diagnostic image data as reference image data in a preset reference database. In this embodiment, the reference image data is different according to the type of positioning device required. For example, when the patient to be matched is a head tumor patient and needs to match the head positioning device, the diagnostic image data is a head diagnostic image; when the patient to be matched is a breast tumor patient and needs to match the radiotherapy positioning compensation model, the diagnostic image data is a chest diagnostic image containing the patient to be matched; when the patient to be matched is an oral tumor patient and needs to match the oral stent, the diagnostic image data is an oral diagnostic image containing teeth. Therefore, different diagnostic image data are selected in advance according to the different positioning devices required, and the diagnostic image data can also cover the above three types of influencing data.

[0030] The key part feature extraction module is used to extract the key part features of all the reference images in the preset reference database, and each key part feature is used as a feature sample, and the sample distance between any two feature samples is calculated according to the preset distance calculation rule. In this embodiment, when the patient to be matched is a patient with a head tumor and needs to match a head positioning device, the key part feature is the lower edge boundary feature of the head; when the patient to be matched is a patient with a breast tumor and needs to match a radiotherapy positioning compensation model, the key part feature is the chest contour feature; when the patient to be matched is a patient with an oral tumor and needs to match an oral stent, the key part feature is the tooth feature. When the patient to be matched is a patient with a head tumor and needs to match a head positioning device, or when the patient to be matched is a patient with a breast tumor and needs to match a radiotherapy positioning compensation module, the above-mentioned calculation of the sample distance between any two feature samples according to the preset distance calculation rule is specifically as follows: each of the feature samples contains a number of coordinate points, and the above-mentioned calculation of the sample distance according to the preset distance calculation rule is specifically as follows: when calculating the distance between two feature samples, the distance between each coordinate point on one of the feature samples and the other feature sample is calculated and used as the first distance, and the first distance with the largest value is used as the sample distance between the two feature samples. The calculation process of the first distance is specifically as follows: the distance from a single coordinate point on one of the feature samples to all coordinate points on the other feature sample is used as the second distance, and the second distance with the smallest value is used as the first distance. When the patient to be matched is a patient with an oral tumor and needs to match an oral stent, the above-mentioned calculation of the sample distance between any two feature samples according to the preset distance calculation rule is specifically as follows: the left endpoint distance and the right endpoint distance between the two feature samples are calculated respectively, and the average of the left endpoint distance and the right endpoint distance is used as the sample distance between the two feature samples.

[0031] Feature classification module, the feature classification module is used to cluster all feature samples according to the sample distance using the k-means algorithm to obtain a number of clustered feature sample sets, each of which contains a number of different feature samples. The k-means algorithm is specifically as follows: the input of the algorithm is a sample set, through which the samples can be clustered, and samples with similar features are clustered into one category. For each point, the center point closest to all center points is calculated, and then the point is classified into the cluster represented by the center point. After one iteration is completed, the center point is recalculated for each cluster, and then for each point, the center point closest to itself is re-searched. This cycle is repeated until the clusters between the previous and subsequent iterations do not change.

[0032] The optimal model solving module is used to solve the optimal feature model for each cluster feature sample set according to a preset optimal solution function to obtain a number of optimal feature samples, and each optimal feature sample is input into a processing device to process a corresponding optimal positioning device. The processing device in this embodiment can be a 3D printing device or a common processing device.

[0033] The key feature extraction module to be matched obtains the medical image data of the patient to be matched from the hospital imaging system, and extracts the key features in the medical image data as the key features to be matched.

[0034] The positioning device matching module is used to match the key features to be matched with all the optimal feature samples according to the preset matching rules, match the corresponding optimal feature samples as the positioning feature samples, and use the optimal positioning device processed according to the positioning feature samples as the radiotherapy positioning device. Specifically, the distance between the key features to be matched and each optimal feature sample is calculated as the matching distance, and it is determined whether the matching distance with the smallest value is less than the preset matching distance threshold. If so, the optimal feature sample corresponding to the smallest matching distance is used as the positioning feature sample, and the optimal positioning device processed according to the positioning feature sample is used as the radiotherapy positioning device. If not, the medical image data of the turning point to be matched is used as the new reference image data and stored in the preset reference database.

[0035] In this embodiment, there are three application scenarios for the above device: 1. When the positioning device to be matched is a head positioning device suitable for patients with head tumors, usually a headrest. 2. When the positioning device to be matched is a radiotherapy positioning compensation mold suitable for patients with breast tumors. The following is a detailed description of the technical solutions in this application under the above application scenarios: 3. When the positioning device to be matched is an oral stent suitable for patients with oral tumors.

[0036] 1. When the positioning device to be matched is a head positioning device suitable for patients with head tumors:

[0037] In the above step of establishing a reference database, head positioning CT or head diagnostic CT of tumor patients who have used or need to use head restraints are collected as reference image data. The head positioning CT or head diagnostic CT is as follows: Figure 2 As shown, the benchmark image data is stored in a preset benchmark database.

[0038] In the above extraction of key part features, the key part feature at this time is the lower edge boundary feature of the head, such as Figure 3 As shown in the figure, the lower edge boundary feature of the head is the external point cloud image of the back of the human head and the neck. After extracting the key part features of all the reference images, each key part feature is used as a feature sample. It is known that the curvature between each two feature samples is the same, close, or completely different. At this time, it is necessary to determine whether the feature samples overlap, that is, it is necessary to calculate the sample distance between any two feature samples according to the preset distance calculation rule. The preset distance calculation rule is shown in formula (1), which is as follows:

[0039]

[0040] Among them, l P,Q and f(P,Q) are the sample distances between any feature sample P and any feature sample Q, p i is the i-th point on the feature sample P, i is a positive integer, q j is the jth point on the feature sample Q, j is a positive integer, d(p i ,q j ) is the distance from the i-th point on the feature sample P to the j-th point on the feature sample Q, that is, the second distance (also known as the Euclidean distance), g(p i ,Q)) is q j is the distance from the i-th point on the feature sample P to the feature sample Q, that is, the first distance. The above formula (1) is illustrated by taking the feature sample P and the feature sample Q as an example, and the distance calculation method between any other two feature samples is the same as the above.

[0041] In the above optimal model solving steps, the following example is used to illustrate: Assume that the Kth clustering feature sample set is clustered, and let the Kth clustering feature sample set be C k And including m feature samples, C k The feature samples contained in k1 ,P k2 ,...,P km , at this time, let the optimal feature sample corresponding to the Kth cluster feature sample set be X k , then the optimal feature model solution is shown in formula (2):

[0042]

[0043] Among them, X k is the optimal feature sample, P ki is the i-th feature sample in the K-th cluster feature sample set, f(X k ,P ki ) is the distance between the optimal feature sample and the i-th feature sample in the K-th cluster feature sample set. Therefore, through the above formula and the way of defining the distance between two samples, the optimal feature sample can be found for each cluster feature sample set, that is, the optimal feature sample is the most similar to all feature samples in the corresponding cluster feature sample set.

[0044] 2. When the positioning device to be matched is a radiotherapy positioning compensation model suitable for breast tumor patients:

[0045] In the above step of establishing a reference database, chest diagnostic images (ie, body surface scan images containing chest contours) of tumor patients who have previously used or need to use headrests are collected as reference image data.

[0046] In this application scenario, the method for calculating the sample distance between any two feature samples in extracting key part features is also shown in the above formula (1).

[0047] The solution process of the optimal model solution step in this application scenario is shown in the above formula (2), which will not be repeated here. Figure 5 The figure shows that when a tumor patient undergoes radiotherapy, he needs to place a radiotherapy positioning compensation mold on his chest. The radiotherapy positioning compensation mold is Figure 5 The square membrane in the.

[0048] 3. When the positioning device to be matched is an oral stent suitable for oral tumor patients:

[0049] In the step of establishing the reference database, oral diagnostic images (including patient teeth) of tumor patients who have used or need to use headrests are collected as reference image data. Figure 4 As shown, the key feature at this time is the tooth feature.

[0050] In this scenario, the key part feature is the tooth feature. When extracting the tooth feature, it is necessary to segment the teeth in the oral diagnostic image and extract the curve corresponding to the teeth. At this time, the tooth feature is a curve. Let the curve be P. The curve is composed of several points. Therefore, let P = {p i}, i is a positive integer, p i is the i-th point of the curve P. At this time, the distance between the two feature samples is the distance between the two corresponding curves, as shown in formula (3):

[0051]

[0052] Among them, l P,Q and f(P,Q) are the sample distances between any feature sample P and any feature sample Q, p left is the left endpoint of feature sample P, p right is the right endpoint of the feature sample P, q left is the left endpoint of the feature sample Q, q right is the right endpoint of the feature sample Q, d(p left -q left ) is the left endpoint distance between feature sample P and feature sample Q, d(p right -q right ) The right endpoint distance between feature sample P and feature sample Q. The above formula (3) is illustrated by taking feature sample P and feature sample Q as an example, and the distance calculation method between any other two feature samples is the same as the above.

[0053] In the above optimal model solving steps, the following example is used to illustrate: Assume that the Kth clustering feature sample set is clustered, and let the Kth clustering feature sample set be C k And including m feature samples, C k The feature samples contained in k1 ,P k2 ,...,P km , at this time, let the optimal feature sample corresponding to the Kth cluster feature sample set be X k , then the optimal feature model solution is shown in formula (4):

[0054]

[0055] Among them, X k is the optimal feature sample, P ki is the i-th feature sample in the K-th cluster feature sample set. Therefore, through the above formula (4) and the way of defining the distance between two samples, the optimal feature sample can be found for each cluster feature sample set, that is, the optimal feature sample is similar to all feature samples in the corresponding cluster feature sample set to the greatest extent.

[0056] A matching device for a radiotherapy positioning device in the present application collects reference image data, extracts key features from the reference image data, and calculates the sample distance between any two feature samples. According to the sample distance, all feature samples are classified using the k-means algorithm, and the optimal model is solved for the feature samples of each cluster to obtain the optimal feature samples. The optimal feature samples corresponding to each cluster are respectively input into a processing device to process the corresponding optimal positioning device. This ensures that the positioning devices pre-set by the hospital have sufficient quantity and high-quality accuracy. When a tumor patient needs to match a positioning device, the key features to be matched in the medical image data of the patient to be matched are extracted. It is only necessary to match the key features to be matched with all the optimal feature samples, thereby matching the most suitable optimal positioning device as the radiotherapy positioning device, thereby improving the matching accuracy of the positioning device and ensuring the accuracy of positioning during radiotherapy. In the prior art, the generation process of the customized positioning device involves obtaining the patient's image data, extracting individual features, and then printing the individualized device through 3D printing. This process, from data modeling to printing after obtaining the patient's image data, often takes one day or even several days to complete the production of the individualized device. The entire positioning device generation process cycle is relatively long, which prolongs the waiting time for patient treatment. In the present application, only matching is required when the patient undergoes radiotherapy to quickly obtain a positioning device that is highly suitable for tumor patients, without the need for a customization process, thereby improving the efficiency of radiotherapy.

[0057] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in any form. Any ordinary technician in the industry can smoothly implement the present invention as shown in the drawings and above. However, any equivalent changes, modifications and evolutions made by technicians familiar with the profession without departing from the scope of the technical solution of the present invention using the technical content disclosed above are all equivalent embodiments of the present invention. At the same time, any equivalent changes, modifications and evolutions made to the above embodiments based on the essential technology of the present invention are still within the protection scope of the technical solution of the present invention.

Claims

1. A matching device for a radiotherapy positioning device, characterized in that: The matching device establishes communication with the hospital imaging system, the processing device and the doctor terminal respectively, and includes the following modules: A reference database establishment module, which is used to obtain diagnostic image data corresponding to a number of tumor patients who need radiotherapy and have undergone radiotherapy from the hospital imaging system, and store all diagnostic image data as reference image data in a preset reference database; A key part feature extraction module, which is used to extract key part features of all reference images in a preset reference database, take each key part feature as a feature sample, and calculate the sample distance between any two feature samples according to a preset distance calculation rule; A feature classification module, wherein the feature classification module is used to perform clustering processing on all feature samples using a k-means algorithm according to the sample distance to obtain a number of cluster feature sample sets, each of which contains a number of different feature samples; An optimal model solving module, the optimal model solving module is used to solve the optimal feature model for each cluster feature sample set according to a preset optimal solution function to obtain a number of optimal feature samples, and the optimal model solving module inputs each optimal feature sample into a processing device to process a corresponding optimal positioning device; A key feature extraction module to be matched, wherein the key feature extraction module to be matched obtains medical image data of the patient to be matched from the hospital imaging system, and extracts key features in the medical image data as key features to be matched; A positioning device matching module, wherein the positioning device matching module is used to match the key features to be matched with all optimal feature samples according to preset matching rules, match the corresponding optimal feature samples as positioning feature samples, and use the optimal positioning device processed according to the positioning feature samples as a radiotherapy positioning device and send it to the doctor terminal for selecting a radiotherapy positioning device for the patient to be matched.

2. A matching device for a radiotherapy positioning device as claimed in claim 1, characterized in that: The method of matching the key feature to be matched with all the optimal feature samples according to the preset matching rules is specifically as follows: calculating the distance between the key feature to be matched and each optimal feature sample and using it as the matching distance, judging whether the matching distance with the smallest value is less than the preset matching distance threshold, if so, using the optimal feature sample corresponding to the smallest matching distance as the positioning feature sample, and using the optimal positioning device processed according to the positioning feature sample as the radiotherapy positioning device, if not, using the medical image data of the turning point to be matched as the new benchmark image data and storing it in the preset benchmark database.

3. A matching device for a radiotherapy positioning device as claimed in claim 1, characterized in that: When the patient to be matched is a patient with a head tumor and needs to match a head positioning device, the diagnostic image data is a head diagnostic image, and the key part feature is a head lower edge boundary feature.

4. A matching device for a radiotherapy positioning device as claimed in claim 1, characterized in that: When the patient to be matched is a breast tumor patient and needs to match a radiotherapy positioning compensation model, the diagnostic image data is a chest diagnostic image containing the patient to be matched, and the key part feature is a chest contour feature.

5. A matching device for a radiotherapy positioning device as claimed in claim 3 or 4, characterized in that: Each of the feature samples contains a number of coordinate points, and the preset distance calculation rule is specifically as follows: when calculating the distance between two feature samples, the distance between each coordinate point on one of the feature samples and the other feature sample is calculated respectively and used as the first distance, and the first distance with the largest value is used as the sample distance between the two feature samples.

6. A matching device for a radiotherapy positioning device as claimed in claim 5, characterized in that: The calculation process of the first distance is specifically as follows: the distance from a single coordinate point on one feature sample to all coordinate points on another feature sample is taken as the second distance, and the second distance with the smallest value is taken as the first distance.

7. A matching device for a radiotherapy positioning device as claimed in claim 1, characterized in that: When the patient to be matched is an oral tumor patient and needs to be matched with an oral stent, the diagnostic image data is an oral diagnostic image containing teeth, the key part features are tooth features, and the sample distance between any two feature samples is calculated according to the preset distance calculation rules. Specifically, the left endpoint distance and the right endpoint distance between the two feature samples are calculated respectively, and the average of the left endpoint distance and the right endpoint distance is used as the sample distance between the two feature samples.

Images