Autoimmune pancreatitis auxiliary diagnosis system based on multi-modal medical image

Abstract

The invention discloses an autoimmune pancreatitis auxiliary diagnosis system based on a multi-modal medical image, and relates to the field of medical images. Comprising a multi-temporal image processing module, a lesion evolution trajectory construction module and the like. The multi-temporal image processing module is used for carrying out automatic segmentation, cross-time-point alignment and modal standardization on the multi-modal medical images at a plurality of time points to obtain image sequences with consistent time sequences; the lesion evolution trajectory construction module is used for extracting quantitative characteristics such as pancreatic morphology, main pancreatic duct morphology, signal or metabolic parameters and the like at each time point, and constructing a lesion evolution trajectory; the trajectory comparison and analysis module is used for comparing the evolutionary trajectory of the subject with a pre-constructed typical evolutionary trajectory set to generate a similarity analysis result; the state updating module is used for fusing newly added time point features and maintaining and updating illness state representation; and the auxiliary diagnosis module outputs an illness state trend prompt and a follow-up visit risk prompt based on the similarity analysis result and the illness state representation.

Description

Technical Field

[0001] This invention relates to the field of medical imaging, and in particular to an auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging. Background Technology

[0002] Autoimmune pancreatitis (AIP) is an immune-related inflammatory disease of the pancreas characterized by diffuse or focal pancreatic enlargement, membranous changes, and stenosis of the main pancreatic duct. Clinically, it often requires comprehensive diagnosis using multimodal imaging modalities such as CT, MRI, DWI, MRCP, and PET-CT. AIP exhibits a long-term fluctuating course; while patients generally achieve good remission after initial treatment, they still have a high risk of relapse during follow-up. Therefore, longitudinal analysis of multi-temporal images is crucial. However, current imaging diagnostic methods are mainly based on qualitative observations at single time points, lacking quantitative descriptions of changes in pancreatic structure, main pancreatic duct morphology, and metabolic indicators. They also lack systematic processing and comparison mechanisms for the relationships between multimodal and multi-time-point images, failing to effectively reflect the overall trend of disease evolution over time. Furthermore, current technologies lack a system for extracting disease evolution trajectories from long-term follow-up images and comparing them with typical evolutionary patterns, making it difficult to assist in judging trends of disease improvement, stabilization, or exacerbation. Therefore, an auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging is proposed. Summary of the Invention

[0003] The main objective of this invention is to provide an auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging, which can effectively solve the problems in the background art.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0005] A multimodal medical imaging-based auxiliary diagnostic system for autoimmune pancreatitis, including:

[0006] The multi-temporal image processing module is used to acquire multimodal medical images of the subject at multiple time points, and to perform automatic segmentation, cross-time point spatial alignment and modal standardization processing on the multimodal medical images to obtain a temporally consistent image sequence.

[0007] The lesion evolution trajectory construction module is used to extract quantitative features such as pancreatic morphology, main pancreatic duct morphology, membranous changes, image signals or metabolic parameters at each time point based on the time-consistent image sequence, and construct a multidimensional evolution trajectory reflecting the changes of the lesion over time in chronological order.

[0008] The trajectory comparison analysis module is used to compare and analyze the multidimensional evolution trajectory reflecting the changes of the lesion over time with a pre-constructed set of typical lesion evolution trajectories, and generate analysis results that characterize the similarity between the subject's trajectory and different lesion types.

[0009] The status update module is used to fuse the trajectory features of the current image at the corresponding time point with the existing disease evolution status representation during the follow-up examination of the examinee, so as to form an updated disease evolution status representation.

[0010] The auxiliary diagnostic module is used to output auxiliary information for judging the diagnostic tendency of autoimmune pancreatitis and providing follow-up risk warnings based on the results of the trajectory comparison analysis module and the updated disease evolution status.

[0011] Furthermore, the multimodal medical images include at least two of CT, MRI, DWI, MRCP, or PET-CT, and the multi-temporal image processing module is specifically used for:

[0012] Automatic segmentation of pancreatic body, main pancreatic duct and surrounding membranous structures in images at different time points;

[0013] Images from each time point are aligned to a unified reference coordinate system through non-rigid body registration, so that the anatomical correspondence remains consistent in the time dimension.

[0014] Standardize the signal or density values ​​of different modes using a unified scale.

[0015] Furthermore, the lesion evolution trajectory construction module is used to extract at least one of the following features at each time point:

[0016] Morphological characteristics of the pancreas, including its size, extent of enlargement, thickness and continuity of the membranous changes;

[0017] Characteristics of the main pancreatic duct, including the length of the stenosis, the average diameter of the stenotic segment, and the degree of upstream dilation;

[0018] Signal or density characteristics, including CT values, MRI T1 / T2 signal intensity, or DWI / ADC values;

[0019] Metabolic characteristics, including PET uptake value, metabolic volume, or total metabolic amount.

[0020] Furthermore, the lesion evolution trajectory construction module further determines the evolution trend of the lesion based on the characteristic changes at adjacent time points, including: whether it shows continuous improvement; whether it tends to stabilize after improvement; whether it aggravates again after stabilization; or whether it shows intermittent fluctuations, and takes the evolution trend as part of the evolution trajectory.

[0021] Furthermore, the trajectory comparison and analysis module includes a set and analysis unit for typical lesion evolution trajectories:

[0022] The typical disease evolution trajectory set is formed by constructing trajectories and clustering patterns in previous cases with long-term follow-up information, and includes trajectory types that show continuous improvement, delayed improvement, stable residual disease, or recurrence of aggravation.

[0023] The analysis unit is used to compare the lesion evolution trajectory of the examinee with the set of typical lesion evolution trajectories and generate similarity analysis results.

[0024] Furthermore, the analysis unit analyzes the differences between the subject's trajectory and each typical trajectory by comprehensively considering the magnitude, direction of change, and rhythm of change of the feature value, and generates similarity analysis results to distinguish different disease evolution patterns.

[0025] Furthermore, the state update module is used for:

[0026] To maintain a disease status representation for each examinee that reflects the process of previous disease changes;

[0027] During each follow-up examination, the lesion characteristics at the newly added time points are fused with the disease status representation;

[0028] Recent data is given higher weight during the fusion process to highlight the impact of recent changes in the patient's condition on the judgment of diagnostic tendency.

[0029] Furthermore, the lesion evolution trajectory construction module further aligns the patient's treatment-related information with the image timeline. The treatment-related information includes treatment dose, dose changes, maintenance treatment time, and drug discontinuation time. Then, based on the correspondence between the treatment plan and image changes, a treatment response index is generated and incorporated into the lesion evolution trajectory.

[0030] Furthermore, the auxiliary diagnostic module is used to output the following based on the disease state representation and similarity analysis results:

[0031] Analysis suggestions indicating whether the condition is more likely to improve, stabilize, or worsen;

[0032] Typical trajectory types indicating similar lesion trajectories among examinees;

[0033] And auxiliary suggestions on follow-up frequency or re-examination intervals generated based on the analysis prompts.

[0034] A method for auxiliary diagnosis of autoimmune pancreatitis based on multimodal medical imaging includes the following steps:

[0035] Multimodal medical images of subjects at multiple time points were acquired, and the pancreas and related structures were segmented, aligned across time points, and normalized modally.

[0036] Extract lesion-related features at each time point and construct the lesion evolution trajectory;

[0037] The lesion evolution trajectory is compared and analyzed with a pre-established set of typical lesion evolution trajectories to form similarity analysis results;

[0038] The patient's condition status is updated based on the trajectory information at newly added time points;

[0039] Based on the disease status representation and similarity analysis results, output prompts to assist in diagnosis.

[0040] The present invention has the following beneficial effects:

[0041] Compared with existing technologies, this solution utilizes a multi-temporal image processing module to segment, register, and standardize multimodal medical images at different time points, providing a unified reference coordinate for subsequent quantitative analysis. A lesion evolution trajectory construction module quantitatively extracts multidimensional indicators such as pancreatic morphology, main pancreatic duct morphology, capsule changes, signal or metabolic characteristics, and constructs a complete lesion evolution trajectory, reflecting the changing trend of the lesion throughout the follow-up process. A trajectory comparison analysis module compares the subject's lesion evolution trajectory with a set of typical trajectories, generating analysis results representing the degree of similarity, enabling the differentiation of different types of disease evolution patterns. A status update module dynamically integrates the subject's follow-up information, increasing the weight of recent disease changes in the final judgment. Finally, an auxiliary diagnostic module outputs disease trend prompts and follow-up risk warnings. Compared with existing diagnostic methods that rely on single-time-point images, this invention enables systematic processing and longitudinal analysis of multimodal, multi-time-point image data, assisting in judging the tendency of disease improvement, stabilization, or aggravation from the perspective of continuous lesion evolution, which is beneficial to improving the objectivity of the diagnosis process and the effectiveness of follow-up management for autoimmune pancreatitis. Attached Figure Description

[0042] Figure 1 This is a structural diagram of the system modules of the present invention. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0044] Example

[0045] This embodiment provides an auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging. Its structure includes a multi-temporal image processing module, a lesion evolution trajectory construction module, a trajectory comparison analysis module, a status update module, and an auxiliary diagnostic module. The modules are sequentially connected through data channels to form a complete information processing flow.

[0046] Implementation of the multi-temporal image processing module:

[0047] This module is used to preprocess multimodal medical images acquired at different time points, making them comparable in spatial scale and data characteristics. Specifically, it includes the following steps:

[0048] Data acquisition: The system receives CT, MRI, DWI, MRCP, or PET-CT image data collected from the subject at multiple time points. The image data at each time point form a natural time series, such as T0 (initial diagnosis), T1 (follow-up), T2 (subsequent re-examination), etc.

[0049] The module automatically segments the pancreas and related structures; it performs automatic segmentation of the pancreas body, main pancreatic duct, and membranous changes in images at each time point, generating masks of the organ and lesion areas. The segmentation results are used for subsequent analysis to avoid discrepancies caused by inconsistencies in structural extent between different time points.

[0050] Spatial alignment across time points: Images from various time points are aligned to a unified reference coordinate system using a non-rigid registration method. The spatial positions of the pancreas and main pancreatic duct at different time points are calibrated using a deformation field, ensuring that the corresponding anatomical regions at time points T0, T1, T2, etc., maintain a matching relationship in three-dimensional space.

[0051] Modal standardization processing: CT values, MRI signal intensity, DWI / ADC values, etc. are uniformly scaled to eliminate signal deviations caused by different equipment and different acquisition parameters.

[0052] After the above processing, the images at each time point are organized into a temporally consistent image sequence.

[0053] Implementation method of the disease evolution trajectory construction module:

[0054] This module is used to extract quantitative indicators from temporally consistent image sequences and construct trajectories reflecting the changes of lesions over time. Specifically, it includes:

[0055] Time-point quantization feature extraction;

[0056] At each time point, at least one of the following quantitative indicators was extracted: pancreatic morphological characteristics: volume, extent of enlargement, thickness and continuity of membranous changes; main pancreatic duct characteristics: length of stenosis segment, diameter of stenosis and degree of dilation upstream of stenosis; signal / density characteristics: CT value, MRI T1 / T2 signal, DWI / ADC value; metabolic characteristics: PET uptake value SUVmax, metabolic volume or total metabolic amount.

[0057] Evolutionary trend judgment: Based on the characteristic changes at adjacent time points, determine whether the lesion belongs to: continuous improvement; stabilization after improvement; relapse after stabilization; or intermittent fluctuation.

[0058] A multidimensional lesion evolution trajectory was constructed; the characteristics and corresponding trends at the above time points were arranged in chronological order to form a multidimensional trajectory data structure reflecting the dynamic changes of the lesions. An independent lesion evolution trajectory was generated for each subject for subsequent comparative analysis.

[0059] Implementation method of the trajectory comparison and analysis module:

[0060] This module is used to compare the lesion evolution trajectory of the examinee with typical trajectory patterns in order to distinguish different lesion evolution types.

[0061] A set of typical disease evolution trajectories is constructed. The system pre-utilizes a large number of past cases with long-term follow-up information, extracts corresponding time series features and performs pattern clustering to form typical evolution patterns including: continuous improvement type; delayed improvement type; stable residual disease type; relapsed severity type; etc.

[0062] Trajectory comparison analysis: The analysis unit receives the subject's lesion evolution trajectory and compares it one by one with the aforementioned typical trajectory set. During the comparison process, the magnitude of characteristic values, the direction of change, and the rhythm of change are comprehensively considered to quantitatively assess the differences between the subject's trajectory and various typical trajectories.

[0063] Generate similarity analysis results; based on the above comparison results, give the degree of similarity between the subject's trajectory and different patterns, which is used to distinguish the trend of the condition: improvement, stabilization or aggravation.

[0064] Implementation of the state update module:

[0065] This module is used to dynamically maintain the patient's condition status, enabling the system to continuously update the judgment criteria over time.

[0066] Status representation maintenance; establish a disease status representation structure for each examinee to record the disease evolution information at all previous time points.

[0067] New information fusion: During each follow-up visit, the quantitative features of the new time points are extracted and fused with the disease status representation to form the latest status data.

[0068] Recent weighting has been increased; during the fusion process, data from recent time periods are given higher weight, making the system's output more consistent with the latest changes in the condition and improving the real-time nature of the auxiliary judgment.

[0069] Implementation method of auxiliary diagnostic module:

[0070] This module generates information for clinical reference based on the aforementioned results, specifically including:

[0071] Disease trend indication; based on the disease status representation and similarity analysis results, it indicates whether the disease is more likely to improve, stabilize or worsen.

[0072] Typical pattern matching prompts indicate which type of typical trajectory pattern (e.g., delayed improvement type) the subject's lesion trajectory is closer to.

[0073] Follow-up frequency recommendations: Based on the above assessment, supplementary suggestions are provided regarding follow-up frequency and re-examination intervals to support the formulation of clinical management strategies.

[0074] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. An auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging, characterized in that, include, The multi-temporal image processing module is used to acquire multimodal medical images of the subject at multiple time points, and to perform automatic segmentation, cross-time point spatial alignment and modal standardization processing on the multimodal medical images to obtain a temporally consistent image sequence. The lesion evolution trajectory construction module is used to extract quantitative features such as pancreatic morphology, main pancreatic duct morphology, membranous changes, image signals or metabolic parameters at each time point based on the time-consistent image sequence, and construct a multidimensional evolution trajectory reflecting the changes of the lesion over time in chronological order. The trajectory comparison analysis module is used to compare and analyze the multidimensional evolution trajectory reflecting the changes of the lesion over time with a pre-constructed set of typical lesion evolution trajectories, and generate analysis results that characterize the similarity between the subject's trajectory and different lesion types. The status update module is used to fuse the trajectory features of the current image at the corresponding time point with the existing disease evolution status representation during the follow-up examination of the examinee, so as to form an updated disease evolution status representation. The auxiliary diagnostic module is used to output auxiliary information for judging the diagnostic tendency of autoimmune pancreatitis and providing follow-up risk warnings based on the results of the trajectory comparison analysis module and the updated disease evolution status.

2. The auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging according to claim 1, characterized in that, The multimodal medical images include at least two of CT, MRI, DWI, MRCP, or PET-CT, and the multi-temporal image processing module is specifically used for: Automatic segmentation of pancreatic body, main pancreatic duct and surrounding membranous structures in images at different time points; Images from each time point are aligned to a unified reference coordinate system through non-rigid body registration, so that the anatomical correspondence remains consistent in the time dimension. Standardize the signal or density values ​​of different modes using a unified scale.

3. The auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging according to claim 1, characterized in that, The lesion evolution trajectory construction module is used to extract at least one of the following features at each time point: Morphological characteristics of the pancreas, including its size, extent of enlargement, thickness and continuity of the membranous changes; Characteristics of the main pancreatic duct, including the length of the stenosis, the average diameter of the stenotic segment, and the degree of upstream dilation; Signal or density characteristics, including CT values, MRI T1 / T2 signal intensity, or DWI / ADC values; Metabolic characteristics, including PET uptake value, metabolic volume, or total metabolic amount.

4. The auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging according to claim 1, characterized in that, The lesion evolution trajectory construction module further determines the evolution trend of the lesion based on the characteristic changes at adjacent time points, including: whether it shows continuous improvement; whether it tends to stabilize after improvement; whether it aggravates again after stabilization; or whether it shows intermittent fluctuations, and takes the evolution trend as part of the evolution trajectory.

5. The auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging according to claim 1, characterized in that, The trajectory comparison and analysis module includes a set and analysis unit for typical lesion evolution trajectories: The typical disease evolution trajectory set is formed by constructing trajectories and clustering patterns in previous cases with long-term follow-up information, and includes trajectory types that show continuous improvement, delayed improvement, stable residual disease, or recurrence of aggravation. The analysis unit is used to compare the lesion evolution trajectory of the examinee with the set of typical lesion evolution trajectories and generate similarity analysis results.

6. The auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging according to claim 5, characterized in that, The analysis unit analyzes the differences between the subject's trajectory and each typical trajectory by comprehensively considering the magnitude, direction of change and rhythm of the feature value, and generates similarity analysis results to distinguish different disease evolution patterns.

7. The auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging according to claim 1, characterized in that, The status update module is used for: To maintain a disease status representation for each examinee that reflects the process of previous disease changes; During each follow-up examination, the lesion characteristics at the newly added time points are fused with the disease status representation; Recent data is given higher weight during the fusion process to highlight the impact of recent changes in the patient's condition on the judgment of diagnostic tendency.

8. The auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging according to claim 1, characterized in that, The lesion evolution trajectory construction module further aligns the patient's treatment-related information with the image timeline. The treatment-related information includes treatment dose, dose changes, maintenance treatment time, and drug discontinuation time. Then, based on the correspondence between the treatment plan and image changes, a treatment response index is generated and incorporated into the lesion evolution trajectory.

9. The auxiliary diagnostic system for autoimmune pancreatitis based on multimodal medical imaging according to claim 1, characterized in that, The auxiliary diagnostic module is used to output the following based on the disease state representation and similarity analysis results: Analysis suggestions indicating whether the condition is more likely to improve, stabilize, or worsen; Typical trajectory types indicating similar lesion trajectories among examinees; And auxiliary suggestions on follow-up frequency or re-examination intervals generated based on the analysis prompts.

10. A method for auxiliary diagnosis of autoimmune pancreatitis based on multimodal medical imaging, characterized in that, Includes the following steps: Multimodal medical images of subjects at multiple time points were acquired, and the pancreas and related structures were segmented, aligned across time points, and normalized modally. Extract lesion-related features at each time point and construct the lesion evolution trajectory; The lesion evolution trajectory is compared and analyzed with a pre-established set of typical lesion evolution trajectories to form similarity analysis results; The patient's condition status is updated based on the trajectory information at newly added time points; Based on the disease status representation and similarity analysis results, output prompts to assist in diagnosis.

Images