Intelligent skeletal muscle disease diagnosis system based on multi-modal data fusion

Abstract

According to the skeletal muscle disease intelligent diagnosis system based on multi-modal data fusion provided by the invention, the types of skeletal muscle diseases are accurately diagnosed by adopting a machine learning and integrated learning method by combining clinical characteristics of a patient, laboratory biochemical examination data and radiomics characteristics. The system comprises a data acquisition module, a feature extraction module, a model construction module, a diagnosis module and a visual decision support module. The data acquisition module collects multi-modal data of a patient, the feature extraction module extracts radiomics features, and the model construction module constructs a clinical feature model, a biochemical model and a radiomics model through a machine learning algorithm and fuses results into a comprehensive diagnosis model. And the diagnosis module diagnoses the skeletal muscle diseases according to the model, and provides a graphical diagnosis result through the visualization module. According to the invention, the accuracy of skeletal muscle disease diagnosis can be improved, intelligent decision support is provided for clinicians, and the risks of misdiagnosis and missed diagnosis are reduced.

Description

Technical Field

[0001] The present invention relates to the field of medical diagnosis technology, and in particular to a multimodal data fusion technology that combines clinical characteristics, laboratory biochemical examination indicators and medical imaging data, and is applied to an intelligent diagnosis system for musculoskeletal diseases. Background Art

[0002] Musculoskeletal diseases are a group of diseases that affect the human motor system, mainly manifested in persistent pain and discomfort, which seriously affect the patient's quality of life and may lead to functional impairment, disability, and increased medical and socioeconomic burdens. Traditional methods of diagnosing musculoskeletal diseases mainly rely on the doctor's personal experience, which leads to a high degree of subjectivity in the diagnosis results and is prone to inconsistency or misdiagnosis.

[0003] In addition, single-modality data (such as using only clinical features or imaging data) often cannot provide sufficient accuracy and reliability in the diagnosis of musculoskeletal diseases. In recent years, with the development of artificial intelligence and medical imaging technology, researchers have begun to try to fuse multimodal data (such as clinical features, laboratory biochemical test indicators, and medical images) to improve the accuracy of diagnosis. However, most current studies are limited to a single modality or a simple feature combination, lacking effective data fusion methods and efficient intelligent diagnostic systems.

[0004] Therefore, developing an intelligent diagnosis system for musculoskeletal diseases that combines multimodal data fusion and machine learning technology can not only improve the accuracy of musculoskeletal disease diagnosis, but also provide strong support for clinical decision-making, which has important clinical application value and research significance. Summary of the invention

[0005] In order to overcome the deficiencies of the above-mentioned prior art, the present invention provides an intelligent diagnosis system for musculoskeletal diseases based on multimodal data fusion. The present invention aims to solve the problems of insufficient accuracy of existing diagnostic methods, complex data processing, and lack of effective diagnostic tools. By combining clinical characteristics, laboratory biochemical examination data, and imaging genomics features, the present invention can accurately and efficiently classify musculoskeletal diseases and assist clinicians in making more accurate diagnostic decisions.

[0006] The technical solution adopted by the present invention includes: Data acquisition module: used to collect clinical data, laboratory biochemical examination data and imaging data of patients. Clinical data include basic information such as age, gender, onset time, symptoms and signs of patients; laboratory biochemical examination data include blood test indicators such as tumor markers; imaging data include CT, MRI and ultrasound images. Feature extraction module: extract imaging omics features from imaging data, analyze images using computer vision and image processing technology, and extract multiple features such as texture features, morphological features, grayscale matrix, etc. Model construction module: use machine learning algorithms (SVM, random forest, XGBoost) to construct clinical feature models, biochemical models and imaging omics models respectively. On this basis, the output results of these three models are integrated to construct a comprehensive musculoskeletal disease diagnosis model. Diagnosis module: by combining multimodal data (including clinical features, laboratory biochemical examination features and imaging omics features), using integrated learning technology, intelligent prediction of musculoskeletal disease classification is provided. Visualization decision support module: based on the constructed diagnostic model, a graphical interface is provided to help doctors quickly evaluate the type of musculoskeletal disease and assist in formulating treatment plans.

[0007] Compared with the existing single-modality diagnosis method, the intelligent diagnosis system for musculoskeletal diseases of the present invention not only improves the accuracy of musculoskeletal disease diagnosis through multimodal data fusion, but also has high operability and clinical application value. The present invention can more comprehensively consider the clinical characteristics and imaging data of patients, improve the reliability of musculoskeletal disease diagnosis, and reduce the risk of misdiagnosis and missed diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 It is a schematic diagram of the implementation process of the present invention, Figure 2 Schematic diagram of the implementation of the medical image feature extraction module. DETAILED DESCRIPTION

[0009] Data acquisition module: This module is used to collect multimodal data of patients, including clinical data, laboratory biochemical examination indicators and imaging data. Clinical data includes basic information of patients (such as gender, age, weight, etc.), and laboratory biochemical examination indicators include blood test indicators (such as CA125, CA199, etc.). Imaging data is obtained through medical imaging equipment such as ultrasound, CT and MRI. These data are stored in the database for subsequent processing.

[0010] Feature extraction module: Based on the image data collected by the data acquisition module, the feature extraction module uses image processing technology (such as gray matrix analysis, texture analysis, morphological analysis, etc.) to extract imaging features. Through these features, important information such as tumor morphology, texture and density can be effectively captured to assist in the analysis and classification of tumor types.

[0011] Model building module: The present invention uses machine learning methods to build classification models. First, based on the collected clinical data, laboratory biochemical examination data and imaging omics features, clinical feature models, biochemical models and imaging omics models are constructed respectively. The clinical feature model and biochemical model are modeled using support vector machine (SVM), random forest, and XGBoost algorithms; the imaging omics model is trained by extracting image features and modeled using similar machine learning algorithms. Then, the output results of the clinical feature model, biochemical model and imaging omics model are fused, and the final comprehensive diagnostic model is obtained through integrated learning. This model can integrate information from multiple data sources to improve the accuracy of predicting the category of musculoskeletal diseases.

[0012] Diagnosis module: The diagnosis module predicts the type of musculoskeletal disease based on the fusion model. In actual application, the diagnosis module first inputs new patient data (including clinical data, laboratory biochemical test data, and imaging data) into the system, and then calculates the type of musculoskeletal disease of the patient through the fusion model. The module can give prediction results in real time to assist doctors in formulating personalized treatment plans.

[0013] Visual Decision Support Module: This module displays the diagnosis results of musculoskeletal diseases through a graphical interface, including the distribution diagram of the various possibilities of musculoskeletal diseases. Doctors can make decisions quickly based on these visual results. At the same time, the system will also provide relevant diagnostic suggestions and treatment plans to provide effective support for clinical treatment.

Claims

1. An intelligent diagnosis system for musculoskeletal diseases based on multimodal data fusion, characterized in that: The system includes: a data acquisition module for collecting clinical data, laboratory biochemical examination data and imaging data of patients; a feature extraction module for extracting radiomics features from imaging data; a model construction module for constructing a clinical feature model, a biochemical model and a radiomics model, and fusing the results of the three to construct a comprehensive diagnostic model; a diagnosis module for predicting the category of musculoskeletal diseases based on the comprehensive diagnostic model; and a visual decision support module for providing graphical evaluation results of musculoskeletal disease types to assist clinicians in making decisions.

2. The intelligent diagnosis system for musculoskeletal diseases according to claim 1, wherein: The clinical data include the patient's basic information, blood test indicators, and symptoms and signs.

3. The intelligent diagnosis system for musculoskeletal diseases according to claim 1, wherein: The laboratory biochemical examination data include the patient's blood routine, inflammatory indicators, tumor markers, bone metabolism markers, rheumatism and rheumatoid markers.

4. The intelligent diagnosis system for musculoskeletal diseases according to claim 1, wherein: The image data includes ultrasound images, X-ray images, CT images, and MRI images.

5. The intelligent diagnosis system for musculoskeletal diseases according to claim 1, wherein: The diagnostic module uses an integrated learning method to combine clinical data, laboratory biochemical examination data and imaging omics features for diagnostic prediction.

6. The intelligent diagnosis system for musculoskeletal diseases according to claim 1, wherein: The visual decision support module provides probability distribution and decision curve analysis of musculoskeletal disease types.

Images