Method and system for detecting whether bladder of patient is full or not during radiotherapy

By using transfer learning technology to automatically segment the bladder region on CBCT images, the problem of accuracy in assessing bladder fullness during radiotherapy was solved, enabling efficient and precise control of the radiotherapy process and improving treatment efficacy and safety.

CN122048851APending Publication Date: 2026-05-15FUJIAN CANCER HOSPITAL (FUJIAN CANCER INST FUJIAN CANCER PREVENTION & CONTROL CENT)
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610120702.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-01-28
Publication Date
2026-05-15

AI Technical Summary

Technical Problem

Existing technologies struggle to efficiently and accurately assess and control bladder fullness during radiotherapy preparation, resulting in low radiotherapy precision and efficiency. Furthermore, traditional medical image processing methods rely on manual intervention and are prone to errors.

Method used

By employing transfer learning technology, a deep learning model trained on CT images is adapted to CBCT images to achieve automatic segmentation and volume estimation of the bladder region. The Dice similarity coefficient is used to assess the consistency of the filling state and provide real-time feedback to adjust the radiotherapy plan.

Benefits of technology

It improves the efficiency and precision of radiotherapy preparation, reduces human intervention, ensures the safety and accuracy of treatment, and protects surrounding normal tissues.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122048851A_ABST
    Figure CN122048851A_ABST
Patent Text Reader

Abstract

The invention relates to a method for detecting whether the bladder of a patient is full or not during radiotherapy, and the method comprises the steps: model preparation and transfer learning: obtaining a deep learning model which is trained on a CT image and is used for automatically segmenting an ROI, and carrying out the transfer learning fine tuning of the deep learning model through a CBCT image data set marked with the bladder ROI, a bladder segmentation model suitable for the CBCT image is obtained; image acquisition and preprocessing: acquiring a CBCT image of a patient and preprocessing the CBCT image; automatic segmentation and volume estimation: inputting the preprocessed CBCT image into a bladder segmentation model to obtain a segmentation result of the bladder ROI, and then calculating the volume of the bladder based on the segmentation result; and filling state judgment: based on the calculated bladder volume and the bladder volume preset by the treatment plan, calculating an overlapping degree index of the two so as to obtain a bladder filling state judgment result. According to the invention, accurate segmentation and volume estimation of the bladder area can be realized, the filling state is further judged, and an instant and objective basis is provided for formulation and adjustment of a radiotherapy scheme.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of medical image processing technology, specifically to a method, system, and related equipment for detecting whether a patient's bladder is full during radiotherapy. Background Technology

[0002] Radiotherapy is one of the three main methods for treating tumors and plays a crucial role in cancer treatment. For some tumors, such as nasopharyngeal carcinoma, cervical cancer, and breast cancer, radiotherapy has been proven to achieve high cure rates. Currently, approximately 70% of cancer patients require radical or adjuvant radiotherapy. The core objective of radiotherapy is to maximize the therapeutic gain ratio, that is, to deliver a sufficiently high radiation dose to the tumor target area while minimizing radiation to surrounding healthy tissues, thereby reducing treatment-related complications. With the continuous development of radiotherapy technology, medical linear accelerators have become the main equipment in clinical radiotherapy. New-generation radiotherapy technologies, such as intensity-modulated radiotherapy (IMRT) and volumetric modulated radiotherapy (VMAT), are widely used globally. The core advantage of these technologies lies in their high-precision target dose distribution, but this also increases the mechanical precision requirements of the accelerator. While these advanced technologies can significantly improve treatment outcomes, they also make the accuracy and uncertainties of radiotherapy more complex. Especially in the transition zone between the high-dose region and the region of rapid dose drop, the precision control of the accelerator is crucial to the final effect of radiotherapy.

[0003] In radiotherapy for cervical cancer and other pelvic regions, bladder fullness has a significant impact on treatment efficacy and the protection of surrounding normal tissues. A full bladder effectively increases the distance between the bladder and the cervical region, thereby reducing radiation damage to surrounding normal tissues. Therefore, accurate assessment and control of bladder fullness are crucial for improving radiotherapy efficacy and protecting surrounding normal tissues.

[0004] Currently, medical imaging technologies (such as CT or CBCT) are widely used in the radiotherapy preparation phase for assessing patients' anatomical structures and developing radiotherapy plans. However, traditional medical image processing methods often rely on manual intervention, such as manually segmenting regions of interest (ROIs) and estimating bladder volume and filling status. These processes are time-consuming and error-prone, which can easily affect the efficiency and accuracy of radiotherapy.

[0005] With the development of artificial intelligence technology, deep learning models have shown great potential in medical image segmentation. Several automatic segmentation models based on CT images already exist. However, CBCT images typically have lower contrast and higher noise levels than diagnostic CT images. Directly applying models trained on CT images to CBCT images results in a significant decrease in segmentation accuracy. Completely re-collecting and annotating a large number of CBCT images to train a new model is costly and time-consuming.

[0006] Therefore, there is an urgent need for a technology that can efficiently, accurately, and automatically assess bladder filling status based on CBCT images in order to optimize the radiotherapy process and achieve personalized and precise treatment. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and system for detecting whether a patient's bladder is full during radiotherapy. This invention utilizes transfer learning technology to quickly adapt a deep learning model that has been trained on CT images to bladder segmentation based on CBCT images, thereby achieving automatic and accurate segmentation and volume estimation of the bladder region, and then judging the fullness status in real time, providing an immediate and objective basis for the formulation and adjustment of radiotherapy plans.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: a method for detecting whether a patient's bladder is full during radiotherapy, comprising the following steps: (1) Model preparation and transfer learning: Obtain a deep learning model that has been trained on CT images for automatic segmentation of regions of interest (ROIs). Use a CBCT image dataset labeled with bladder ROIs to perform transfer learning fine-tuning on the deep learning model to obtain a bladder segmentation model suitable for CBCT images. (2) Image acquisition and preprocessing: Acquire the patient's CBCT images and perform preprocessing; (3) Automatic segmentation and volume estimation: The preprocessed CBCT image is input into the bladder segmentation model to obtain the segmentation result of the bladder ROI, and then the bladder volume is calculated based on the segmentation result; (4) Bladder filling status determination: Based on the calculated bladder volume and the bladder volume preset in the treatment plan, the overlap index between the two is calculated to quantitatively assess the consistency between the current actual bladder filling status and the preset status in the treatment plan, thereby obtaining the bladder filling status determination result.

[0009] Further, in step (1), the transfer learning fine-tuning includes: adjusting the parameters of the deep learning model using the CBCT image dataset at a learning rate lower than that of the pre-training stage.

[0010] Further, in step (2), the preprocessing includes denoising, contrast enhancement and normalization; the denoising uses Gaussian filtering or median filtering; the contrast enhancement uses histogram equalization.

[0011] Furthermore, in step (3), the method for calculating the bladder volume based on the segmentation results is as follows: The number of pixels belonging to the bladder ROI in the statistical segmentation results is counted, and the actual volume of the bladder is calculated based on the spatial resolution of the CBCT image.

[0012] Furthermore, in step (4), the Dice similarity coefficient is used as an overlap index to quantitatively assess the consistency between the current actual bladder filling state and the pre-set state of the treatment plan. Its calculation formula is as follows: DSC=2×|V plan ∩V treatment | / (|V plan |+|V treatment |) Where DSC is the Dice similarity coefficient, V plan V represents the bladder volume pre-planned in the treatment plan. treatment This represents the calculated bladder volume.

[0013] Furthermore, the method also includes step (5): during radiotherapy, steps (2) to (4) are repeated to monitor changes in bladder filling status in real time and provide feedback information based on the changes to assist in adjusting the radiotherapy plan.

[0014] The present invention also provides a bladder filling detection system for implementing the above method, comprising: The model processing module is used to store and execute bladder segmentation models adapted for CBCT images after transfer learning fine-tuning. The image processing module is used to preprocess the input CBCT images; The segmentation and calculation module is used to call the bladder segmentation model in the model processing module to segment the bladder ROI in the preprocessed CBCT image and calculate the bladder volume based on the segmentation results. The determination module is used to calculate the overlap index between the calculated bladder volume and the bladder volume preset in the treatment plan, so as to quantitatively evaluate the consistency between the current actual bladder filling state and the preset state in the treatment plan, and output the bladder filling state determination result.

[0015] Furthermore, the bladder filling detection system also includes a feedback module, which generates prompts for adjusting the radiotherapy plan based on the real-time changes in filling status output by the judgment module during radiotherapy.

[0016] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method.

[0017] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method.

[0018] Compared with existing technologies, this invention has the following beneficial effects: By using transfer learning technology, existing ROI automatic segmentation models based on CT images are adapted to CBCT images, significantly reducing the need for training models from scratch, shortening model preparation time, and improving the accuracy of CBCT image segmentation. By achieving automatic segmentation and volume estimation, this invention not only reduces manual intervention and improves the efficiency of radiotherapy preparation, but also provides clinicians with rapid and accurate bladder filling information, thereby improving the precision and safety of radiotherapy. Real-time monitoring of bladder filling status and its dynamic feedback mechanism can help clinicians adjust treatment plans in a timely manner, maximizing the protection of surrounding normal tissues, demonstrating high practical value and innovation. Attached Figure Description

[0019] Figure 1 This is a flowchart of a method for detecting whether a patient's bladder is full during radiotherapy, provided in an embodiment of the present invention. Figure 2 This is an example diagram illustrating the determination of bladder fullness status in an embodiment of the present invention. Detailed Implementation

[0020] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0021] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0022] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0023] like Figure 1 As shown in the figure, this embodiment provides a method for detecting whether a patient's bladder is full during radiotherapy, and the implementation steps are as follows.

[0024] (1) Model preparation and transfer learning: Obtain a deep learning model that has been trained on CT images for automatic segmentation of regions of interest (ROIs). Use a CBCT image dataset labeled with bladder ROIs to perform transfer learning fine-tuning on the deep learning model to obtain a bladder segmentation model suitable for CBCT images.

[0025] In this process, transfer learning fine-tuning uses a learning rate lower than that of the pre-training phase to adjust the parameters of the deep learning model using the CBCT image dataset.

[0026] (2) Image acquisition and preprocessing: Acquire the patient's CBCT images and perform preprocessing.

[0027] The preprocessing includes denoising, contrast enhancement, and normalization. Denoising is performed using Gaussian filtering or median filtering. Contrast enhancement is performed using histogram equalization.

[0028] (3) Automatic segmentation and volume estimation: The preprocessed CBCT image is input into the bladder segmentation model to obtain the segmentation results of the bladder ROI, and then the bladder volume is calculated based on the segmentation results.

[0029] The specific implementation method for calculating bladder volume based on segmentation results is as follows: count the number of pixels belonging to the bladder ROI in the segmentation results, and calculate the actual bladder volume according to the spatial resolution of the CBCT image.

[0030] (4) Bladder filling status determination: Based on the calculated bladder volume and the bladder volume preset in the treatment plan, the overlap index between the two is calculated to quantitatively assess the consistency between the current actual bladder filling status and the preset status in the treatment plan, thereby obtaining the bladder filling status determination result.

[0031] Among them, the Dice Similarity Coefficient (DSC) was used as an overlap index to quantitatively assess the spatial consistency between the current actual bladder filling state and the pre-set state of the treatment plan during radiotherapy.

[0032] The formula for calculating DSC is: DSC=2×|V plan ∩V treatment | / (|V plan |+|V treatment |) Among them, V plan This represents the bladder volume pre-planned in the treatment plan, specifically the bladder volume delineated based on CT images during the treatment planning phase; V treatmentThis represents the calculated bladder volume, which is the bladder volume automatically calculated using this method based on CBCT images on the day of treatment. |⋅| represents the number of voxels or physical volume value of the calculated volume, and ∩ represents the spatial intersection of the two regions.

[0033] In actual operation, the system presets a clinically acceptable overlap threshold of 0.85 or 0.90. When the real-time calculated DSC value is lower than this threshold, the system automatically triggers an alert mechanism and generates a structured prompt message on the physician's terminal: "Insufficient bladder volume consistency (DSC=<calculated value>), it is recommended to check the bladder filling status, and if necessary, supplement filling or re-optimize the treatment plan." This ensures the accuracy of target area dose coverage and effectively avoids the risk of excessive irradiation of organs endangered by anatomical deformation, thereby improving the safety of the radiotherapy process and the reliability of the plan execution.

[0034] (5) During radiotherapy, repeat steps (2) to (4) to monitor changes in bladder filling status in real time and provide feedback information based on the changes to assist in adjusting the radiotherapy plan.

[0035] Figure 2 An example of bladder fullness determination is shown. Figure 2 In the image, (a) shows the bladder volume delineated based on CT images, (b) shows the bladder volume automatically calculated using this method based on CBCT images, and (c) displays the bladder differences between the two. Figure 2 As can be seen, this method can accurately calculate the current bladder volume, thereby calculating the overlap index and determining the bladder filling status.

[0036] The specific implementation process of these steps will be described in detail below.

[0037] 1. Model preparation and transfer learning 1.1 Selection of Pre-trained Models A deep learning model for automatic ROI segmentation, trained and validated on a large number of computed tomography (CT) images, was selected. This model should possess high recognition capabilities for various anatomical structures (such as the bladder) and high segmentation accuracy, enabling effective segmentation across different patients.

[0038] 1.2 Transfer Learning Scheme Design 1.2.1 Data Preparation A certain amount of CBCT image data, especially images of the bladder region, will be collected and processed. These images will be used for training and validation of transfer learning, and the data will be labeled to accurately determine the ROI of the bladder.

[0039] 1.2.2 Model Fine-tuning The pre-trained model is fine-tuned using collected CBCT image data. By setting an appropriate learning rate, the model can effectively learn from new data while retaining the useful information it gained during training on CT images.

[0040] 1.2.3 Model Validation The model was validated on an independent validation dataset to ensure that the fine-tuned model could accurately segment the bladder ROI in CBCT images, thus guaranteeing the stability and accuracy of the segmentation.

[0041] 2. Image Acquisition and Preprocessing 2.1 Image Acquisition When patients undergo CBCT scans, it is essential to acquire high-quality image data, ensuring that the image clarity, resolution, and contrast meet the requirements of subsequent processing, thus providing a solid foundation for segmentation and volume estimation.

[0042] 2.2 Image Preprocessing Scheme Design 2.2.1 Noise Reduction Denoising algorithms such as Gaussian filtering or median filtering are applied to remove noise from images, improve image quality, and ensure the accuracy of subsequent processing.

[0043] 2.2.2 Contrast Enhancement Histogram equalization and other algorithms are used to enhance image contrast, making the edges of the bladder region clearer and facilitating subsequent segmentation operations.

[0044] 2.2.3 Standardization Standardize image data to a uniform range (e.g., between 0 and 1) to ensure that the model can perform efficient and accurate segmentation within a uniform data range.

[0045] 3. Automatic segmentation and volume estimation 3.1 Automatic Segmentation Execution A deep learning model fine-tuned through transfer learning was used to automatically segment the bladder region of interest (ROI) in preprocessed CBCT images. The segmentation results, including the boundaries and location information of the bladder ROI, were recorded to provide necessary input data for subsequent volumetric calculations.

[0046] 3.2 Design of Volume Estimation Scheme Based on the boundary information of the bladder's ROI, a volume calculation algorithm is applied to estimate the bladder's volume. For example, the volume of the bladder can be calculated by counting the number of pixels within the segmented region and combining this with the spatial resolution of the image, thus providing a basis for subsequent determination of its fullness.

[0047] 4. Determination of fullness status 4.1 Design of the Filling State Determination Scheme 4.1.1 Judgment Criteria The Dice similarity coefficient (DSC) was used as an overlap index to quantitatively assess the spatial consistency between the current actual bladder filling state and the pre-set state of the treatment plan during radiotherapy, thereby determining whether the bladder filling state meets the requirements of radiotherapy.

[0048] 4.1.2 Real-time determination Based on the calculated bladder volume and the bladder volume preset in the treatment plan, the overlap index between the two is calculated to determine the bladder filling status in real time. This information will be immediately fed back to clinicians, enabling them to adjust the treatment plan in a timely manner according to changes in bladder filling status during radiotherapy, ensuring treatment accuracy and reducing radiation damage to normal tissues.

[0049] This embodiment also provides a bladder filling detection system for implementing the above method, including: The model processing module is used to store and execute bladder segmentation models adapted for CBCT images after transfer learning fine-tuning. The image processing module is used to preprocess the input CBCT images; The segmentation and calculation module is used to call the bladder segmentation model in the model processing module to segment the bladder ROI in the preprocessed CBCT image and calculate the bladder volume based on the segmentation results. The judgment module is used to calculate the overlap index between the calculated bladder volume and the bladder volume preset in the treatment plan, so as to quantitatively evaluate the consistency between the current actual bladder filling status and the preset state in the treatment plan, and output the bladder filling status judgment result. The feedback module is used to generate prompts for adjusting the radiotherapy plan based on the real-time changes in the filling status output by the judgment module during radiotherapy.

[0050] This embodiment also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method.

[0051] This embodiment also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described above.

[0052] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0053] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0054] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0055] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0056] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for detecting bladder fullness in patients during radiotherapy, characterized in that, Includes the following steps: (1) Model preparation and transfer learning: Obtain a deep learning model that has been trained on CT images for automatic segmentation of regions of interest (ROIs). Use a CBCT image dataset labeled with bladder ROIs to perform transfer learning fine-tuning on the deep learning model to obtain a bladder segmentation model suitable for CBCT images. (2) Image acquisition and preprocessing: Acquire the patient's CBCT images and perform preprocessing; (3) Automatic segmentation and volume estimation: The preprocessed CBCT image is input into the bladder segmentation model to obtain the segmentation result of the bladder ROI, and then the bladder volume is calculated based on the segmentation result; (4) Bladder filling status determination: Based on the calculated bladder volume and the bladder volume preset in the treatment plan, the overlap index between the two is calculated to quantitatively assess the consistency between the current actual bladder filling status and the preset status in the treatment plan, thereby obtaining the bladder filling status determination result.

2. The method for detecting bladder fullness during radiotherapy according to claim 1, characterized in that, In step (1), the transfer learning fine-tuning includes: adjusting the parameters of the deep learning model using the CBCT image dataset at a learning rate lower than that of the pre-training stage.

3. The method for detecting bladder fullness during radiotherapy according to claim 1, characterized in that, In step (2), the preprocessing includes denoising, contrast enhancement and normalization; the denoising uses Gaussian filtering or median filtering; the contrast enhancement uses histogram equalization.

4. The method for detecting bladder fullness during radiotherapy according to claim 1, characterized in that, In step (3), the method for calculating bladder volume based on the segmentation results is as follows: The number of pixels belonging to the bladder ROI in the statistical segmentation results is counted, and the actual volume of the bladder is calculated based on the spatial resolution of the CBCT image.

5. The method for detecting bladder fullness during radiotherapy according to claim 1, characterized in that, In step (4), the Dice similarity coefficient is used as an overlap index to quantitatively assess the consistency between the current actual bladder filling state and the pre-set state of the treatment plan. The calculation formula is as follows: DSC = 2 x |V plan ∩V treatment | / (|V plan |+|V treatment |) Where DSC is the Dice similarity coefficient, V plan V represents the bladder volume pre-planned in the treatment plan. treatment This represents the calculated bladder volume.

6. The method for detecting bladder fullness during radiotherapy according to claim 1, characterized in that, It also includes step (5): During radiotherapy, repeat steps (2) to (4) to monitor changes in bladder filling status in real time and provide feedback information based on the changes to assist in adjusting the radiotherapy plan.

7. A bladder fullness detection system for implementing the method according to any one of claims 1-6, characterized in that, include: The model processing module is used to store and execute bladder segmentation models adapted for CBCT images after transfer learning fine-tuning. The image processing module is used to preprocess the input CBCT images; The segmentation and calculation module is used to call the bladder segmentation model in the model processing module to segment the bladder ROI in the preprocessed CBCT image and calculate the bladder volume based on the segmentation results. The determination module is used to calculate the overlap index between the calculated bladder volume and the bladder volume preset in the treatment plan, so as to quantitatively evaluate the consistency between the current actual bladder filling state and the preset state in the treatment plan, and output the bladder filling state determination result.

8. The bladder fullness detection system according to claim 7, characterized in that, It also includes a feedback module, which generates prompts for adjusting the radiotherapy plan based on the real-time changes in filling status output by the judgment module during radiotherapy.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-6.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1-6.