Intelligent report system and method for automatic comparison and diagnosis of bedside chest x-ray in ICU
Patent Information
- Application Number
- CN202211049097.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-08-30
- Publication Date
- 2026-09-15
- Estimated Expiration
- 2042-08-30
AI Technical Summary
床旁胸片的摄影设备是移动式X线机,常规胸片是固定式X线机,移动式X线机的曝光能力稍低,对软组织的穿透能力稍差于固定式X线机,因此部分患者的图像质量欠佳
[0019] This invention applies AI models and rule-based programs to automatically compare and analyze current ICU bedside chest X-rays with previous bedside chest X-rays, resulting in an intelligent diagnostic reporting system for bedside chest X-ray examinations. This system is integrated with PACS/RIS to automatically acquire patient medical history, previous images, and previous reports. After image acquisition, the system segments the images and uses the segmented data, along with previous images and reports, to automatically output changes in mediastinal morphology, pleural cavity lesions, and lung lesions through registered images and registered segmented data regions. This automatic assessment and transmission of results to a structured report improves physician efficiency. More importantly, when the intelligent system detects potentially risky situations, it immediately issues warnings to relevant medical staff, enabling them to promptly address patient abnormalities and prioritize effective treatment, ensuring patient safety.
Smart Images

Figure CN115359878B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of medical information, and more specifically, to an intelligent reporting system and method for automatic comparative diagnosis of ICU bedside chest X-ray films. Background Technology
[0002] In modern general hospitals, chest X-rays are among the most frequently performed imaging diagnostic tasks. A significant proportion of these are bedside chest X-rays for critically ill patients in intensive care units (ICUs). In most ICUs, daily chest X-rays are standard procedure. However, multiple bedside chest X-rays may be performed within a single day when a patient's condition changes or after certain medical procedures. Because the intervals between consecutive bedside chest X-rays are very short, comparisons must be made with both the most recent and earlier consecutive X-rays to detect progressive changes. The purpose of imaging examinations is to detect critical findings immediately, identify signs that may alter treatment plans, and track imaging changes. The sheer number of examinations and the repeated comparisons between multiple examinations constitute a heavy diagnostic burden for radiologists.
[0003] Meanwhile, bedside chest X-rays present greater diagnostic challenges. Unlike routine chest X-rays, bedside chest X-rays have distinct characteristics in terms of equipment, patient positioning, and patient condition. Bedside chest X-rays utilize mobile X-ray machines, while routine chest X-rays use stationary machines. Mobile machines have slightly lower exposure capabilities and less penetration into soft tissues, resulting in sometimes poorer image quality. Bedside chest X-rays are performed in an anteroposterior view, while routine chest X-rays are performed in a posteroanterior view. The anteroposterior view can magnify the cardiac image, making it difficult to assess mediastinal and cardiac morphological abnormalities. Patients receiving bedside chest X-rays tend to have more severe conditions and are equipped with numerous medical devices supporting and monitoring physiological functions, while patients receiving routine chest X-rays generally have milder conditions and fewer such devices. The increased interference from these devices on bedside chest X-rays can affect the detection of important signs. Therefore, bedside chest X-ray diagnosis presents certain difficulties. Because bedside chest X-rays are of great diagnostic significance for critically ill patients, involve a large number of comparative images, and have poor image quality, radiologists need considerable experience to diagnose bedside chest X-rays, and patients' conditions can change rapidly, requiring quick diagnoses. Currently, radiologists diagnose bedside X-ray images one by one. Due to the heavy workload and diagnostic difficulty, misdiagnosis or delayed treatment is easily caused, compromising patient safety. Summary of the Invention
[0004] In view of this, the main objective of the present invention is to provide an intelligent reporting system and method for automatic comparison and diagnosis of ICU bedside chest X-rays. This system can automatically compare and diagnose the current ICU bedside chest X-ray with previous bedside chest X-rays immediately after the ICU bedside chest X-ray is taken, and immediately issue warnings to relevant medical staff when potential risks are detected, so as to prioritize effective treatment for these patients. This solves the problems of misdiagnosis or delayed treatment caused by the high difficulty and workload of diagnosis in the prior art.
[0005] To achieve the above objectives, the technical solution of the present invention is implemented as follows:
[0006] On one hand, this invention provides an intelligent reporting system for automatic comparative diagnosis of ICU bedside chest X-ray images. This system includes: a patient information acquisition module, a segmentation data acquisition module, an image registration module, a comparison analysis module, and a structured report module. The patient information acquisition module is connected to the segmentation data acquisition module, image registration module, comparison analysis module, and structured report module, and is used to acquire the patient's current ICU bedside chest X-ray image, patient medical history information, previous images, and previous report information. The previous images include the patient's most recent ICU bedside chest X-ray image and all ICU bedside chest X-ray images prior to the most recent one. The segmentation data acquisition module is connected to the patient information acquisition module, image registration module, and structured report module, and is used to input the current ICU bedside chest X-ray image and previous images into multiple deep learning models for image segmentation, obtaining... The system includes corresponding segmentation data; the types of segmentation data include segmentation data of chest structures and segmentation data of chest lesions; an image registration module, connected to the patient information acquisition module, segmentation data acquisition module, comparison analysis module, and structured report module, is used to perform image registration between the current ICU bedside chest X-ray image and previous images based on the segmentation data, and output all registered images and the registered segmentation data; a comparison analysis module, connected to the patient information acquisition module, image registration module, and structured report module, is used to analyze the changes of related diseases on the current ICU bedside chest X-ray image based on all registered images, registered segmentation data, and previous report information; and a structured report module, connected to the patient information acquisition module, segmentation data acquisition module, image registration module, and comparison analysis module, is used to automatically output the final diagnostic data based on the changes of related diseases.
[0007] Preferably, the segmentation data of the chest structure includes: the entire chest imaging area, lung field area, and mediastinal area of the current ICU bedside chest X-ray image and previous images; the segmentation data of the chest lesions includes: the pleural cavity lesion area and the abnormal high-density lesion area of the lung in the current ICU bedside chest X-ray image and the most recent ICU bedside chest X-ray image; wherein, the pleural cavity lesion area includes the pneumothorax area and the pleural effusion area.
[0008] Preferably, the registered images are: the registered current ICU bedside chest X-ray image, the registered most recent ICU bedside chest X-ray image, and all previously registered previous ICU bedside chest X-ray images; the registered segmentation data are: the mediastinal region in the registered current ICU bedside chest X-ray image, the pleural cavity lesion region in the registered current ICU bedside chest X-ray image, the abnormal high-density lesion region in the lungs in the registered current ICU bedside chest X-ray image, the mediastinal region in the registered most recent ICU bedside chest X-ray image, the pleural cavity lesion region in the registered most recent ICU bedside chest X-ray image, the abnormal high-density lesion region in the lungs in the registered most recent ICU bedside chest X-ray image, and the mediastinal region in all previously registered previous ICU bedside chest X-ray images.
[0009] Preferably, the comparative analysis module further includes: a mediastinal morphology analysis unit, used to query the evaluation of the mediastinum in the most recent report in the previous report information, and to judge the changes in mediastinal morphology on the current ICU bedside chest X-ray image based on the evaluation of the mediastinum, and output the changes in mediastinal morphology; wherein, the evaluation of the mediastinum is normal or abnormal; the changes in mediastinal morphology include: significantly enlarged mediastinal shadow, enlarged mediastinal shadow, reduced mediastinal shadow, and no change in mediastinal shadow.
[0010] Preferably, the comparative analysis module further includes: a pleural cavity lesion analysis unit, used to query the evaluation of the pleural cavity in the most recent report in the previous report information, and based on the evaluation of the pleural cavity, to judge the changes of pleural cavity lesions on the current ICU bedside chest X-ray image, and output the changes of pleural cavity lesions; wherein, the evaluation of the pleural cavity is normal or abnormal; the changes of pleural cavity lesions include: newly added pleural cavity lesions, normal pleural cavity, significantly enlarged pleural cavity lesions, enlarged pleural cavity lesions, shrunken pleural cavity lesions, and no change in pleural cavity lesions.
[0011] Preferably, the comparative analysis module further includes: a lung lesion analysis unit, used to query the evaluation of lung lesions in the most recent report in the previous report information, and to judge the changes of lung lesions on the current ICU bedside chest X-ray image based on the evaluation of lung lesions, and output the changes of lung lesions; wherein, the evaluation of lung lesions is normal or abnormal; the changes of lung lesions include: new lung lesions, normal lungs, significantly enlarged lung lesions, enlarged lung lesions, shrunken lung lesions, and no changes in lung lesions.
[0012] Preferably, the judgment of changes in mediastinal morphology on the current ICU bedside chest X-ray image includes: when the mediastinal evaluation is normal, the mediastinal morphology analysis unit compares the mediastinal region in the most recent registered ICU bedside chest X-ray image with the mediastinal region in all previous ICU bedside chest X-ray images after registration, finds the previous minimum mediastinal region, and compares the previous minimum mediastinal region with the mediastinal region in the current registered ICU bedside chest X-ray image. If the mediastinal region in the current registered ICU bedside chest X-ray image is larger than the previous minimum mediastinal region by more than a first preset threshold, it is judged as a significant increase in mediastinal shadow; if the mediastinal region in the current registered ICU bedside chest X-ray image is larger than the previous minimum mediastinal region by more than a second preset threshold, it is judged as an increase in mediastinal shadow; if the mediastinal region in the current registered ICU bedside chest X-ray image is smaller than or equal to the previous minimum mediastinal region by less than the second preset threshold, it is judged as no change in mediastinal shadow; when the mediastinal evaluation is abnormal, the mediastinal morphology analysis... The unit compares and analyzes the mediastinal region in the most recent registered ICU bedside chest X-ray image with the mediastinal region in the current registered ICU bedside chest X-ray image. If the mediastinal region in the current registered ICU bedside chest X-ray image is larger than a first preset threshold compared to the mediastinal region in the most recent registered ICU bedside chest X-ray image, it is determined that the mediastinal shadow is significantly enlarged; if the mediastinal region in the current registered ICU bedside chest X-ray image is larger than a second preset threshold compared to the mediastinal region in the most recent registered ICU bedside chest X-ray image, it is determined that the mediastinal shadow is enlarged; if the mediastinal region in the current registered ICU bedside chest X-ray image is smaller than a second preset threshold compared to the mediastinal region in the most recent registered ICU bedside chest X-ray image, it is determined that the mediastinal shadow is smaller; if the change in the mediastinal region in the current registered ICU bedside chest X-ray image is less than or equal to the second preset threshold, it is determined that the mediastinal shadow has not changed.
[0013] Preferably, the assessment of changes in pleural cavity lesions on the current ICU bedside chest X-ray image includes: when the pleural cavity assessment is normal, the pleural cavity lesion analysis unit detects whether there are pleural cavity lesions in the registered current ICU bedside chest X-ray image; if so, it is determined to be a newly added pleural cavity lesion; if not, it is determined to be a normal pleural cavity; when the pleural cavity assessment is abnormal, the pleural cavity lesion analysis unit compares the pleural cavity lesion area in the most recent registered ICU bedside chest X-ray image with the pleural cavity lesion area in the registered current ICU bedside chest X-ray image, and if the pleural cavity lesion area in the registered current ICU bedside chest X-ray image is larger than that in the most recent registered ICU bedside chest X-ray image, the pleural cavity lesion area is larger than that in the third... If the threshold is reached, the pleural cavity lesion is judged to have significantly increased in size; if the pleural cavity lesion area in the current registered ICU bedside chest X-ray image is larger than the pleural cavity lesion area in the most recent registered ICU bedside chest X-ray image than the fourth preset threshold, the pleural cavity lesion is judged to have increased in size; if the pleural cavity lesion area in the current registered ICU bedside chest X-ray image is smaller than the pleural cavity lesion area in the most recent registered ICU bedside chest X-ray image than the fourth preset threshold, the pleural cavity lesion is judged to have decreased in size; if the change in the pleural cavity lesion area in the current registered ICU bedside chest X-ray image is less than or equal to the fourth preset threshold, the pleural cavity lesion is judged to have remained unchanged.
[0014] Preferably, the changes in lung lesions on the current ICU bedside chest X-ray image are assessed, including: when the evaluation of lung lesions is normal, the lung lesion analysis unit detects whether there are lung lesions in the registered current ICU bedside chest X-ray image; if so, it is determined to be a newly added lung lesion; if not, it is determined to be normal lungs; when the evaluation of lung lesions is abnormal, the lung lesion analysis unit compares the abnormal high-density lesion area in the most recent registered ICU bedside chest X-ray image with the abnormal high-density lesion area in the current registered ICU bedside chest X-ray image; if the abnormal high-density lesion area in the current registered ICU bedside chest X-ray image is larger than the abnormal high-density lesion area in the most recent registered ICU bedside chest X-ray image by more than a fifth preset threshold, it is determined to be a lung lesion. If the lesion area significantly increases; or if the area of abnormal high-density lesions in the pleural cavity of the current registered ICU bedside chest X-ray image increases by more than the sixth preset threshold compared to the area of abnormal high-density lesions in the most recent registered ICU bedside chest X-ray image, then the lung lesion is judged to have increased; if the area of abnormal high-density lesions in the pleural cavity of the current registered ICU bedside chest X-ray image decreases by less than the sixth preset threshold compared to the area of abnormal high-density lesions in the most recent registered ICU bedside chest X-ray image, then the lung lesion is judged to have decreased; if the change in the area of abnormal high-density lesions in the pleural cavity of the current registered ICU bedside chest X-ray image compared to the area of abnormal high-density lesions in the most recent registered ICU bedside chest X-ray image is less than or equal to the sixth preset threshold, then the lung lesion is judged to have remained unchanged.
[0015] On the other hand, the present invention also provides an intelligent reporting method for automatic comparison and diagnosis of ICU bedside chest X-ray images, comprising: acquiring the patient's current ICU bedside chest X-ray image, patient medical history information, previous images, and previous report information; wherein, the previous images are the patient's most recent ICU bedside chest X-ray image and all ICU bedside chest X-ray images prior to the most recent one; inputting the current ICU bedside chest X-ray image and previous images into multiple deep learning models for image segmentation, respectively obtaining corresponding segmentation data; the types of segmentation data include segmentation data of chest structures and segmentation data of chest lesions; based on the segmentation data, performing image registration between the current ICU bedside chest X-ray image and previous images, and outputting all registered images and registered segmentation data; based on all registered images, registered segmentation data, and previous report information, analyzing the changes of related diseases on the current ICU bedside chest X-ray image; and automatically outputting final diagnostic data based on the changes of related diseases.
[0016] Preferably, the segmentation data of the chest structure includes: the entire chest imaging area, lung field area, and mediastinal area of the current ICU bedside chest X-ray image and previous images; the segmentation data of the chest lesions includes: the pleural cavity lesion area and the abnormal high-density lesion area of the lung in the current ICU bedside chest X-ray image and the most recent ICU bedside chest X-ray image; wherein, the pleural cavity lesion area includes the pneumothorax area and the pleural effusion area.
[0017] Preferably, the registered images are: the registered current ICU bedside chest X-ray image, the registered most recent ICU bedside chest X-ray image, and all previously registered previous ICU bedside chest X-ray images; the registered segmentation data are: the mediastinal region in the registered current ICU bedside chest X-ray image, the pleural cavity lesion region in the registered current ICU bedside chest X-ray image, the abnormal high-density lesion region in the lungs in the registered current ICU bedside chest X-ray image, the mediastinal region in the registered most recent ICU bedside chest X-ray image, the pleural cavity lesion region in the registered most recent ICU bedside chest X-ray image, the abnormal high-density lesion region in the lungs in the registered most recent ICU bedside chest X-ray image, and the mediastinal region in all previously registered previous ICU bedside chest X-ray images.
[0018] Technical effects of the present invention:
[0019] This invention applies AI models and rule-based programs to automatically compare and analyze current ICU bedside chest X-rays with previous bedside chest X-rays, resulting in an intelligent diagnostic reporting system for bedside chest X-ray examinations. This system is integrated with PACS / RIS to automatically acquire patient medical history, previous images, and previous reports. After image acquisition, the system segments the images and uses the segmented data, along with previous images and reports, to automatically output changes in mediastinal morphology, pleural cavity lesions, and lung lesions through registered images and registered segmented data regions. This automatic assessment and transmission of results to a structured report improves physician efficiency. More importantly, when the intelligent system detects potentially risky situations, it immediately issues warnings to relevant medical staff, enabling them to promptly address patient abnormalities and prioritize effective treatment, ensuring patient safety. Attached Figure Description
[0020] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:
[0021] Figure 1A schematic diagram of the intelligent reporting system for automatic comparison and diagnosis of chest X-ray films at the ICU bedside according to Embodiment 1 of the present invention is shown.
[0022] Figure 2 A schematic diagram of the intelligent reporting system for automatic comparison and diagnosis of chest X-ray images at the ICU bedside according to Embodiment 2 of the present invention is shown.
[0023] Figure 3 A schematic diagram of the intelligent reporting system for automatic comparison and diagnosis of chest X-ray images at the ICU bedside according to Embodiment 3 of the present invention is shown.
[0024] Figure 4 This diagram illustrates the current chest X-ray image in the ICU bedside chest X-ray automatic comparison diagnosis intelligent reporting system according to Embodiment 3 of the present invention.
[0025] Figure 5 This diagram illustrates the chest imaging region of the current chest X-ray image in the intelligent reporting system for automatic comparison and diagnosis of ICU bedside chest X-ray films according to Embodiment 3 of the present invention.
[0026] Figure 6 This diagram illustrates the current chest X-ray image after registration in the ICU bedside chest X-ray automatic comparison diagnosis intelligent reporting system according to Embodiment 3 of the present invention.
[0027] Figure 7 This diagram illustrates the chest imaging region of the current chest X-ray image after registration in the intelligent reporting system for automatic comparison and diagnosis of ICU bedside chest X-rays according to Embodiment 3 of the present invention.
[0028] Figure 8 This diagram illustrates the pleural effusion region of the current chest X-ray image after registration in the intelligent reporting system for automatic comparison and diagnosis of ICU bedside chest X-rays according to Embodiment 3 of the present invention.
[0029] Figure 9 This diagram illustrates a previous chest X-ray image before registration in the ICU bedside chest X-ray automatic comparison diagnostic intelligent reporting system according to Embodiment 3 of the present invention.
[0030] Figure 10 A schematic diagram of the chest imaging region of a previous chest X-ray image before registration is shown in the intelligent reporting system for automatic comparison and diagnosis of chest X-ray images at the ICU bedside according to Embodiment 3 of the present invention.
[0031] Figure 11 This diagram illustrates a previously registered chest X-ray image in the ICU bedside chest X-ray automatic comparison and diagnostic intelligent reporting system according to Embodiment 3 of the present invention.
[0032] Figure 12A schematic diagram of the chest imaging area of a previous chest X-ray image after registration is shown in the intelligent reporting system for automatic comparison and diagnosis of chest X-ray images at the ICU bedside according to Embodiment 3 of the present invention.
[0033] Figure 13 This diagram illustrates the pleural effusion region of a previous chest X-ray image after registration in the intelligent reporting system for automatic comparison and diagnosis of ICU bedside chest X-rays according to Embodiment 3 of the present invention.
[0034] Figure 14 A schematic diagram of the structure of the ICU bedside chest X-ray automatic comparison diagnosis intelligent reporting system according to Embodiment 4 of the present invention is shown;
[0035] Figure 15 A flowchart of an intelligent reporting method for automatic comparative diagnosis of ICU bedside chest X-ray films according to Embodiment 5 of the present invention is shown. Detailed Implementation
[0036] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.
[0037] Example 1
[0038] Figure 1 A schematic diagram of an intelligent reporting system for automatic comparison and diagnosis of ICU bedside chest X-ray films according to Embodiment 1 of the present invention is shown; Figure 1 As shown, the system includes: a patient information acquisition module 10, a segmentation data acquisition module 20, an image registration module 30, a comparison analysis module 40, and a structured report module 50.
[0039] Before acquiring patient information, the system must first identify the nature of the patient's image to determine whether it was a bedside chest X-ray examination. The qualitative judgment result (whether it was a bedside chest X-ray or not) is then output to the corresponding control in the structured report's "Technical Assessment." If the examination is determined to be an ICU bedside chest X-ray, the patient's X-ray image is output for subsequent AI diagnostic procedures. If the examination is determined to be not an ICU bedside chest X-ray, the AI diagnostic process is aborted, a prompt message is sent, and the relevant personnel are responsible for handling the matter, which is also recorded in the database.
[0040] The patient information acquisition module 10 is connected to the segmentation data acquisition module 20, the image registration module 30, the comparison analysis module 40, and the structured report module 50, respectively, and is used to acquire the patient's current ICU bedside chest X-ray image, patient medical history information, previous images, and previous report information; among which, the previous images are the patient's most recent ICU bedside chest X-ray image and all ICU bedside chest X-ray images before the most recent one;
[0041] The system retrieves the patient's current primary disease diagnosis and clinical manifestations from the RIS (Reliability, Infectious Disease) and electronic medical records. It outputs qualitative judgment data—the patient's current primary disease diagnosis from the RIS and electronic medical records—and qualitative judgments—the clinical manifestations retrieved from the RIS and electronic medical records. This is implemented using an AI model and program.
[0042] The patient's current primary disease diagnosis obtained from RIS and electronic medical records is returned to the corresponding controls in the structured report "Clinical Assessment" for the diagnostic physician's reference;
[0043] Clinical manifestations obtained from RIS and electronic medical records are returned to the corresponding controls in the structured report "Clinical Manifestations" for the diagnosing physician's reference.
[0044] The program retrieves previous images and reports, including images from the same patient's ICU admission and similar examinations from the PACS / RIS database, as well as the most recent bedside chest X-ray report from the RIS system, extracting positive data such as mediastinal abnormalities, pleural cavity abnormalities, and lung abnormalities.
[0045] The system outputs all images of the same patient's bedside chest X-ray examination after admission to the ICU from the PACS, and the most recent bedside chest X-ray examination report from the structured reporting system; it also outputs the images of the most recent bedside chest X-ray examination, all images of the previous bedside chest X-ray examinations in the ICU, the mediastinal evaluation in the most recent report, the pleural cavity evaluation in the most recent report, and the lung evaluation in the most recent report.
[0046] The image of the most recent bedside chest X-ray in the ICU, and all images of the previous bedside chest X-ray in the ICU are used for image comparison;
[0047] The mediastinal evaluation, pleural cavity evaluation, and lung evaluation from the most recent report are used for subsequent image comparison and report generation.
[0048] The segmentation data acquisition module 20 is connected to the patient information acquisition module 10, the image registration module 30, and the structured report module 50, respectively. It is used to input the current ICU bedside chest X-ray image and previous images into multiple deep learning models for image segmentation, and obtain the corresponding segmentation data. The types of segmentation data include segmentation data of chest structures and segmentation data of chest lesions.
[0049] The segmentation data for chest structures includes: the entire chest imaging area, lung field area, and mediastinal area of the current ICU bedside chest X-ray image and previous images; the segmentation data for chest lesions includes: the pleural cavity lesion area and the abnormal high-density lesion area of the lung in the current ICU bedside chest X-ray image and the most recent ICU bedside chest X-ray image; among which, the pleural cavity lesion area includes the pneumothorax area and the pleural effusion area.
[0050] The segmentation data of the thoracic structures includes the thoracic imaging region, lung field, and mediastinal region.
[0051] The segmentation data for chest lesions includes the pleural cavity lesion area and the area of abnormal high-density lesions in the lungs.
[0052] Segment the chest imaging region from all images of the same patient's bedside chest X-ray examination in PACS upon admission to the ICU, and output the chest imaging region data for subsequent image registration and chest structure segmentation.
[0053] From all images of the same patient's bedside chest X-ray upon admission to the ICU via PACS, the lung fields and mediastinal region are segmented. The lung field region data is output for subsequent segmentation and comparison of intrapulmonary lesions. The mediastinal region is used for subsequent comparison of mediastinal morphology.
[0054] Segmentation of pleural cavity lesion areas: Input the current ICU bedside chest X-ray image, the most recent ICU bedside chest X-ray image, and the chest imaging area; Output the pneumothorax and pleural effusion areas, qualitative judgment data - pneumothorax, qualitative judgment data - pleural effusion, and qualitative judgment data - hydropneumothorax;
[0055] If pneumothorax is present, the results are returned to the "Pneumothorax" control in the structured report. The pneumothorax region is used for subsequent image comparison.
[0056] If pleural effusion is present, the results are returned to the controls in the structured report "Pleural Effusion". The pleural effusion area is used for subsequent image comparison.
[0057] If pneumothorax is present, the results are returned to the controls in the structured report "Pneumothorax". The pneumothorax and pleural effusion areas are used for subsequent image comparison.
[0058] Segmentation of abnormal high-density lesions in the lungs: Input the current ICU bedside chest X-ray image, the most recent ICU bedside chest X-ray image, and chest imaging; output the abnormal high-density lesion region in the lungs and a qualitative judgment – abnormal high-density lesion in the lungs; if abnormal high-density lesions in the lungs are present, the results are returned to the "Lung Lesions" control in the structured report. The abnormal high-density lesion region in the lungs is used for subsequent image comparison.
[0059] The image registration module 30 is connected to the patient information acquisition module 10, the segmentation data acquisition module 20, the comparison analysis module 40, and the structured report module 50, respectively. It is used to register the current ICU bedside chest X-ray image with previous images based on the segmentation data, and output all registered images and registered segmentation data.
[0060] The complete set of images after registration includes: the current registered ICU bedside chest X-ray image, the most recent registered ICU bedside chest X-ray image, and all previous registered ICU bedside chest X-ray images.
[0061] The segmented data after registration includes: the mediastinal region in the current registered ICU bedside chest X-ray image, the pleural cavity lesion region in the current registered ICU bedside chest X-ray image, the abnormal high-density lesion region in the lungs in the current registered ICU bedside chest X-ray image, the mediastinal region in the most recent registered ICU bedside chest X-ray image, the pleural cavity lesion region in the most recent registered ICU bedside chest X-ray image, the abnormal high-density lesion region in the lungs in the most recent registered ICU bedside chest X-ray image, and the mediastinal region in all previous ICU bedside chest X-ray images after registration.
[0062] Specifically: Input the current ICU bedside chest X-ray image, the most recent ICU bedside chest X-ray image, all images from previous ICU bedside chest X-ray images, and the chest imaging region; Output the registered current ICU bedside chest X-ray image, the registered most recent ICU bedside chest X-ray image, all images from previous ICU bedside chest X-ray images, the mediastinum in the registered current ICU bedside chest X-ray image, the pneumothorax in the registered current ICU bedside chest X-ray image, and the registered current ICU bedside chest X-ray image. The registered segmented data includes pleural effusion in the current ICU bedside chest X-ray image, abnormal high-density lesions in the lungs in the most recent ICU bedside chest X-ray image, mediastinum in the most recent ICU bedside chest X-ray image, pneumothorax in the most recent ICU bedside chest X-ray image, pleural effusion in the most recent ICU bedside chest X-ray image, abnormal high-density lesions in the lungs in the most recent ICU bedside chest X-ray image, and the mediastinal region in all images from previous ICU bedside chest X-ray examinations. The registered segmented data is used for subsequent lesion comparison.
[0063] The comparison and analysis module 40 is connected to the patient information acquisition module 10, the image registration module 30, and the structured report module 50, respectively. It is used to analyze the changes of related diseases on the current ICU bedside chest X-ray images based on all registered images, registered segmentation data, and previous report information.
[0064] The structured report module 50 is connected to the patient information acquisition module 10, the segmentation data acquisition module 20, the image registration module 30, and the comparison analysis module 40, and is used to automatically output the final diagnostic data based on the changes in related diseases.
[0065] The structured report module integrates findings from all functional modules to arrive at an overall diagnostic impression. Based on the rules built into the structured report, the final diagnostic data is automatically obtained and returned to the diagnostic impression in the structured report.
[0066] If a critical value is detected in the diagnostic impression, an alert message will be sent, which will be handled by relevant personnel and recorded in the database.
[0067] All data and images are stored in a structured report database.
[0068] This invention applies AI models and rule-based programs to automatically compare and analyze current ICU bedside chest X-rays with previous bedside chest X-rays, resulting in an intelligent diagnostic reporting system for bedside chest X-ray examinations. This system is integrated with PACS / RIS to automatically acquire patient medical history, previous images, and previous reports. After image acquisition, the system segments the images and uses the segmented data, along with previous images and reports, to automatically output changes in mediastinal morphology, pleural cavity lesions, and lung lesions through registered images and registered segmented data regions. This automatic assessment and transmission of results to a structured report improves physician efficiency. More importantly, when the intelligent system detects potentially risky situations, it immediately issues warnings to relevant medical staff, enabling them to promptly address patient abnormalities and prioritize effective treatment, ensuring patient safety.
[0069] Example 2
[0070] Figure 2 A schematic diagram of the intelligent reporting system for automatic comparison and diagnosis of ICU bedside chest X-ray films according to Embodiment 2 of the present invention is shown; Figure 2 As shown, the comparative analysis module 40 also includes: a mediastinal morphology analysis unit 402, which is used to query the evaluation of the mediastinum in the most recent report in the previous report information, and judge the changes in mediastinal morphology on the current ICU bedside chest X-ray image based on the evaluation of the mediastinum, and output the changes in mediastinal morphology; wherein, the evaluation of the mediastinum is normal or abnormal; the changes in mediastinal morphology include: significantly increased mediastinal shadow, increased mediastinal shadow, reduced mediastinal shadow, and no change in mediastinal shadow.
[0071] Specifically, the mediastinal area in the current registered ICU bedside chest X-ray image is compared with the mediastinal area in the most recent registered ICU bedside chest X-ray image and the mediastinal area in all images of the most recent registered ICU bedside chest X-ray image to determine whether there are significant changes in mediastinal morphology.
[0072] Input the current registered ICU bedside chest X-ray image, the most recent registered ICU bedside chest X-ray image, all images from previous ICU bedside chest X-ray examinations, the mediastinal region in the current registered ICU bedside chest X-ray image, the mediastinum in the most recent registered ICU bedside chest X-ray image, the mediastinal region in all images from previous ICU bedside chest X-ray examinations, and the mediastinal evaluation from the most recent report; Output the previous minimum mediastinal region, qualitative judgment - changes in mediastinal morphology, qualitative judgment - significant increase in mediastinal size, and prompt information - significant increase in mediastinal shadow.
[0073] When changes in mediastinal morphology are present, the results are returned to the "Changes in Mediastinal Morphology" control in the structured report. When the mediastinum significantly enlarges, a prompt message is sent, which is handled by relevant personnel and recorded in the database.
[0074] Analysis of changes in mediastinal morphology:
[0075] When the mediastinum is assessed as normal, the mediastinal morphology analysis unit 402 compares the mediastinal region in the most recent registered ICU bedside chest X-ray image with the mediastinal region in all previous ICU bedside chest X-ray images after registration, identifies the previous minimum mediastinal region, and compares the previous minimum mediastinal region with the mediastinal region in the current registered ICU bedside chest X-ray image. If the mediastinal region in the current registered ICU bedside chest X-ray image is larger than the previous minimum mediastinal region by more than a first preset threshold (e.g., 20%), it is judged as a significant increase in mediastinal shadow; if the mediastinal region in the current registered ICU bedside chest X-ray image is larger than the previous minimum mediastinal region by more than a second preset threshold (e.g., 10%), it is judged as an increase in mediastinal shadow; if the mediastinal region in the current registered ICU bedside chest X-ray image is smaller than or equal to the previous minimum mediastinal region, it is judged as no change in mediastinal shadow.
[0076] When the mediastinum is assessed as abnormal, the mediastinal morphology analysis unit compares the mediastinal region in the most recent registered ICU bedside chest X-ray image with the mediastinal region in the current registered ICU bedside chest X-ray image. If the mediastinal region in the current registered ICU bedside chest X-ray image is larger than a first preset threshold compared to the mediastinal region in the most recent registered ICU bedside chest X-ray image, it is determined that the mediastinal shadow is significantly enlarged; if the mediastinal region in the current registered ICU bedside chest X-ray image is larger than a second preset threshold compared to the mediastinal region in the most recent registered ICU bedside chest X-ray image, it is determined that the mediastinal shadow is enlarged; if the mediastinal region in the current registered ICU bedside chest X-ray image is smaller than a second preset threshold compared to the mediastinal region in the most recent registered ICU bedside chest X-ray image, it is determined that the mediastinal shadow is smaller; if the change in the mediastinal region in the current registered ICU bedside chest X-ray image is less than or equal to the second preset threshold, it is determined that the mediastinal shadow has not changed.
[0077] Example 3
[0078] Figure 3 A schematic diagram of the intelligent reporting system for automatic comparison and diagnosis of ICU bedside chest X-ray films according to Embodiment 3 of the present invention is shown; Figure 3 As shown, the comparative analysis module 40 also includes: a pleural cavity lesion analysis unit 404, used to query the evaluation of the pleural cavity in the most recent report in the previous report information, and to judge the changes of pleural cavity lesions on the current ICU bedside chest X-ray image based on the evaluation of the pleural cavity, and output the changes of pleural cavity lesions; wherein, the evaluation of the pleural cavity is normal or abnormal; the changes of pleural cavity lesions include: newly added pleural cavity lesions, normal pleural cavity, significantly enlarged pleural cavity lesions, enlarged pleural cavity lesions, shrunken pleural cavity lesions, and no change in pleural cavity lesions.
[0079] Specifically: Compare the area of pleural lesions in the current registered ICU bedside chest X-ray image with that in the most recent registered ICU bedside chest X-ray image to determine if there are significant changes in pleural lesions. Input: current registered ICU bedside chest X-ray image, most recent registered ICU bedside chest X-ray image, pneumothorax and pleural effusion areas in the current registered ICU bedside chest X-ray image, pneumothorax and pleural effusion areas in the most recent registered ICU bedside chest X-ray image, and pleural region evaluation from the most recent report; Output: Qualitative judgment - Change in pleural lesions, Qualitative judgment - Significantly enlarged pleural lesions, and Warning message - Significantly enlarged pleural lesions.
[0080] When changes in the pleural cavity morphology are present, the results are returned to the "Changes in Pleural Cavity Morphology" control in the structured report. When a pleural cavity lesion significantly enlarges, an alert is sent for relevant personnel to handle, and the incident is recorded in the database.
[0081] Analysis of pleural cavity lesions:
[0082] The system assesses changes in pleural cavity lesions on current ICU bedside chest X-ray images, including: when the pleural cavity is assessed as normal, the pleural cavity lesion analysis unit detects whether there are pleural cavity lesions in the registered current ICU bedside chest X-ray images. If there are, they are identified as newly added pleural cavity lesions; if not, the pleural cavity is identified as normal.
[0083] When the pleural cavity assessment is abnormal, the pleural cavity lesion analysis unit compares the pleural cavity lesion area in the most recent registered ICU bedside chest X-ray image with the pleural cavity lesion area in the current registered ICU bedside chest X-ray image. If the pleural cavity lesion area in the current registered ICU bedside chest X-ray image is larger than the pleural cavity lesion area in the most recent registered ICU bedside chest X-ray image by more than a third preset threshold (e.g., 20%), it is judged as a significant increase in pleural cavity lesion size. If the pleural cavity lesion area in the registered ICU bedside chest X-ray image increases by more than the fourth preset threshold (e.g., 10%), it is judged as an enlarged pleural cavity lesion; if the pleural cavity lesion area in the current registered ICU bedside chest X-ray image decreases by less than the fourth preset threshold compared to the most recent registered ICU bedside chest X-ray image, it is judged as a shrunken pleural cavity lesion; if the change in the pleural cavity lesion area in the current registered ICU bedside chest X-ray image compared to the most recent registered ICU bedside chest X-ray image is less than or equal to the fourth preset threshold, it is judged as no change in the pleural cavity lesion.
[0084] The following example illustrates this:
[0085] Figure 4 This diagram illustrates the current chest X-ray image in the ICU bedside chest X-ray automatic comparison diagnosis intelligent reporting system according to Embodiment 3 of the present invention.
[0086] Figure 5 This diagram illustrates the chest imaging region of the current chest X-ray image in the intelligent reporting system for automatic comparison and diagnosis of ICU bedside chest X-ray films according to Embodiment 3 of the present invention.
[0087] Figure 6This diagram illustrates the current chest X-ray image after registration in the ICU bedside chest X-ray automatic comparison diagnosis intelligent reporting system according to Embodiment 3 of the present invention.
[0088] Figure 7 This diagram illustrates the chest imaging region of the current chest X-ray image after registration in the intelligent reporting system for automatic comparison and diagnosis of ICU bedside chest X-rays according to Embodiment 3 of the present invention.
[0089] Figure 8 This diagram illustrates the pleural effusion region of the current chest X-ray image after registration in the intelligent reporting system for automatic comparison and diagnosis of ICU bedside chest X-rays according to Embodiment 3 of the present invention.
[0090] Figure 9 This diagram illustrates a previous chest X-ray image before registration in the ICU bedside chest X-ray automatic comparison diagnostic intelligent reporting system according to Embodiment 3 of the present invention.
[0091] Figure 10 A schematic diagram of the chest imaging region of a previous chest X-ray image before registration is shown in the intelligent reporting system for automatic comparison and diagnosis of chest X-ray images at the ICU bedside according to Embodiment 3 of the present invention.
[0092] Figure 11 This diagram illustrates a previously registered chest X-ray image in the ICU bedside chest X-ray automatic comparison and diagnostic intelligent reporting system according to Embodiment 3 of the present invention.
[0093] Figure 12 A schematic diagram of the chest imaging area of a previous chest X-ray image after registration is shown in the intelligent reporting system for automatic comparison and diagnosis of chest X-ray images at the ICU bedside according to Embodiment 3 of the present invention.
[0094] Figure 13 This diagram illustrates the pleural effusion region in a previously recorded chest X-ray image after registration in the intelligent reporting system for automatic comparison and diagnosis of ICU bedside chest X-rays according to Embodiment 3 of the present invention; as shown. Figure 4-13 As shown, the analysis of effusion lesions
[0095] Based on the patient's medical history, the current primary diagnosis is: pleural effusion.
[0096] Obtain previous report information, among which the positive data is: pleural effusion.
[0097] Image registration is then performed.
[0098] The pleural effusion area in the registered previous chest X-ray image was compared with the pleural effusion area in the registered current chest X-ray image. The current pleural effusion area was increased by more than 20% compared with the previous pleural effusion area, which was considered to be a significant increase in the size of the pleural cavity lesion.
[0099] Example 4
[0100] Figure 14 A schematic diagram of the intelligent reporting system for automatic comparison and diagnosis of ICU bedside chest X-ray films according to Embodiment 4 of the present invention is shown; Figure 14 As shown, the comparison and analysis module 40 also includes a lung lesion analysis unit 406, which is used to query the evaluation of lung lesions in the most recent report in the previous report information, judge the changes of lung lesions on the current ICU bedside chest X-ray image based on the evaluation of lung lesions, and output the changes of lung lesions; wherein, the evaluation of lung lesions is normal or abnormal; the changes of lung lesions include: new lung lesions, normal lungs, significantly enlarged lung lesions, enlarged lung lesions, shrunken lung lesions, and no changes in lung lesions.
[0101] Specifically: Compare the area of lung lesions in the current registered ICU bedside chest X-ray image with that in the most recent registered ICU bedside chest X-ray image to determine if there is a significant change in lung lesions. Input: current registered ICU bedside chest X-ray image, most recent registered ICU bedside chest X-ray image, lung lesions in the current registered ICU bedside chest X-ray image, lung lesion area in the most recent registered ICU bedside chest X-ray image, and lung lesion evaluation from the most recent report; Output: Qualitative judgment - change in lung lesions, Qualitative judgment - significantly increased lung lesions, and prompt information - significantly increased lung lesions.
[0102] When changes in the morphology of lung lesions are observed, the results are returned to the "Changes in the Morphology of Lung Lesions" control in the structured report. When lung lesions significantly enlarge, an alert message is sent, which is then handled by relevant personnel and recorded in the database.
[0103] Analysis of changes in lung lesions:
[0104] The system assesses changes in lung lesions on current ICU bedside chest X-ray images, including: When the lung lesion assessment is normal, the lung lesion analysis unit checks for lung lesions in the registered current ICU bedside chest X-ray image; if present, it is considered a new lung lesion; otherwise, the lungs are considered normal. When the lung lesion assessment is abnormal, the lung lesion analysis unit compares the abnormal high-density lesion area in the most recent registered ICU bedside chest X-ray image with the abnormal high-density lesion area in the current registered ICU bedside chest X-ray image. If the abnormal high-density lesion area in the current registered ICU bedside chest X-ray image increases by more than a fifth preset threshold (e.g., 20%) compared to the abnormal high-density lesion area in the most recent registered ICU bedside chest X-ray image, it is considered a lung lesion. The lung lesion is considered to have increased significantly if the area of abnormal high-density lesions in the pleural cavity of the current registered ICU bedside chest X-ray image is larger than the sixth preset threshold (10%) compared to the area of abnormal high-density lesions in the most recent registered ICU bedside chest X-ray image; the lung lesion is considered to have shrunk if the area of abnormal high-density lesions in the pleural cavity of the current registered ICU bedside chest X-ray image is smaller than the sixth preset threshold; and the lung lesion is considered to have remained unchanged if the change in the area of abnormal high-density lesions in the pleural cavity of the current registered ICU bedside chest X-ray image compared to the area of abnormal high-density lesions in the most recent registered ICU bedside chest X-ray image is less than or equal to the sixth preset threshold.
[0105] Example 5
[0106] Figure 15 A flowchart of an intelligent reporting method for automatic comparative diagnosis of ICU bedside chest X-ray films according to Embodiment 5 of the present invention is shown; Figure 15 As shown, the method includes the following steps:
[0107] Before acquiring patient information, the system must first identify the nature of the patient's image to determine whether it was a bedside chest X-ray examination. The qualitative judgment result (whether it was a bedside chest X-ray or not) is then output to the corresponding control in the structured report's "Technical Assessment." If the examination is determined to be an ICU bedside chest X-ray, the patient's X-ray image is output for subsequent AI diagnostic procedures. If the examination is determined to be not an ICU bedside chest X-ray, the AI diagnostic process is aborted, a prompt message is sent, and the relevant personnel are responsible for handling the matter, which is also recorded in the database.
[0108] Step S501: Obtain the patient's current ICU bedside chest X-ray image, patient medical history information, previous images, and previous report information; wherein, the previous images are the patient's most recent ICU bedside chest X-ray image and all ICU bedside chest X-ray images prior to the most recent one;
[0109] The system retrieves the patient's current primary disease diagnosis and clinical manifestations from the RIS (Reliability, Infectious Disease) and electronic medical records. It outputs qualitative judgment data—the patient's current primary disease diagnosis from the RIS and electronic medical records—and qualitative judgments—the clinical manifestations retrieved from the RIS and electronic medical records. This is implemented using an AI model and program.
[0110] The patient's current primary disease diagnosis obtained from RIS and electronic medical records is returned to the corresponding controls in the structured report "Clinical Assessment" for the diagnostic physician's reference;
[0111] Clinical manifestations obtained from RIS and electronic medical records are returned to the corresponding controls in the structured report "Clinical Manifestations" for the diagnosing physician's reference.
[0112] The program retrieves previous images and reports, including images from the same patient's ICU admission and similar examinations from the PACS / RIS database, as well as the most recent bedside chest X-ray report from the RIS system, extracting positive data such as mediastinal abnormalities, pleural cavity abnormalities, and lung abnormalities.
[0113] The system outputs all images of the same patient's bedside chest X-ray examination after admission to the ICU from the PACS, and the most recent bedside chest X-ray examination report from the structured reporting system; it also outputs the images of the most recent bedside chest X-ray examination, all images of the previous bedside chest X-ray examinations in the ICU, the mediastinal evaluation in the most recent report, the pleural cavity evaluation in the most recent report, and the lung evaluation in the most recent report.
[0114] The image of the most recent bedside chest X-ray in the ICU, and all images of the previous bedside chest X-ray in the ICU are used for image comparison;
[0115] The mediastinal evaluation, pleural cavity evaluation, and lung evaluation from the most recent report are used for subsequent image comparison and report generation.
[0116] Step S502: Input the current ICU bedside chest X-ray image and previous images into multiple deep learning models for image segmentation to obtain corresponding segmentation data; the types of segmentation data include segmentation data of chest structures and segmentation data of chest lesions.
[0117] The segmentation data for chest structures includes: the entire chest imaging area, lung field area, and mediastinal area of the current ICU bedside chest X-ray image and previous images; the segmentation data for chest lesions includes: the pleural cavity lesion area and the abnormal high-density lesion area of the lung in the current ICU bedside chest X-ray image and the most recent ICU bedside chest X-ray image; among which, the pleural cavity lesion area includes the pneumothorax area and the pleural effusion area.
[0118] The segmentation data of the thoracic structures includes the thoracic imaging region, lung field, and mediastinal region.
[0119] The segmentation data for chest lesions includes the pleural cavity lesion area and the area of abnormal high-density lesions in the lungs.
[0120] Segment the chest imaging region from all images of the same patient's bedside chest X-ray examination in PACS upon admission to the ICU, and output the chest imaging region data for subsequent image registration and chest structure segmentation.
[0121] From all images of the same patient's bedside chest X-ray upon admission to the ICU via PACS, the lung fields and mediastinal region are segmented. The lung field region data is output for subsequent segmentation and comparison of intrapulmonary lesions. The mediastinal region is used for subsequent comparison of mediastinal morphology.
[0122] Segmentation of pleural cavity lesion areas: Input the current ICU bedside chest X-ray image, the most recent ICU bedside chest X-ray image, and the chest imaging area; Output the pneumothorax and pleural effusion areas, qualitative judgment data - pneumothorax, qualitative judgment data - pleural effusion, and qualitative judgment data - hydropneumothorax;
[0123] If pneumothorax is present, the results are returned to the "Pneumothorax" control in the structured report. The pneumothorax region is used for subsequent image comparison.
[0124] If pleural effusion is present, the results are returned to the controls in the structured report "Pleural Effusion". The pleural effusion area is used for subsequent image comparison.
[0125] If pneumothorax is present, the results are returned to the controls in the structured report "Pneumothorax". The pneumothorax and pleural effusion areas are used for subsequent image comparison.
[0126] Segmentation of abnormal high-density lesions in the lungs: Input the current ICU bedside chest X-ray image, the most recent ICU bedside chest X-ray image, and chest imaging; output the abnormal high-density lesion region in the lungs and a qualitative judgment – abnormal high-density lesion in the lungs; if abnormal high-density lesions in the lungs are present, the results are returned to the "Lung Lesions" control in the structured report. The abnormal high-density lesion region in the lungs is used for subsequent image comparison.
[0127] Step S503: Based on the segmentation data, perform image registration between the current ICU bedside chest X-ray image and previous images, and output all registered images and the registered segmentation data;
[0128] The complete set of images after registration includes: the current registered ICU bedside chest X-ray image, the most recent registered ICU bedside chest X-ray image, and all previous registered ICU bedside chest X-ray images.
[0129] The segmented data after registration are: the mediastinal region in the current registered ICU bedside chest X-ray image, the pleural cavity lesion region in the current registered ICU bedside chest X-ray image, the abnormal high-density lesion region in the lungs in the current registered ICU bedside chest X-ray image, the mediastinal region in the most recent registered ICU bedside chest X-ray image, the pleural cavity lesion region in the most recent registered ICU bedside chest X-ray image, the abnormal high-density lesion region in the lungs in the most recent registered ICU bedside chest X-ray image, and the mediastinal region in all previous ICU bedside chest X-ray images after registration.
[0130] Specifically: Input the current ICU bedside chest X-ray image, the most recent ICU bedside chest X-ray image, all images from previous ICU bedside chest X-ray images, and the chest imaging region; Output the registered current ICU bedside chest X-ray image, the registered most recent ICU bedside chest X-ray image, all images from previous ICU bedside chest X-ray images, the mediastinum in the registered current ICU bedside chest X-ray image, the pneumothorax in the registered current ICU bedside chest X-ray image, and the registered current ICU bedside chest X-ray image. The registered segmented data includes pleural effusion in the current ICU bedside chest X-ray image, abnormal high-density lesions in the lungs in the most recent ICU bedside chest X-ray image, mediastinum in the most recent ICU bedside chest X-ray image, pneumothorax in the most recent ICU bedside chest X-ray image, pleural effusion in the most recent ICU bedside chest X-ray image, abnormal high-density lesions in the lungs in the most recent ICU bedside chest X-ray image, and the mediastinal region in all images from previous ICU bedside chest X-ray examinations. The registered segmented data is used for subsequent lesion comparison.
[0131] Step S504: Based on all registered images, registered segmentation data, and previous report information, analyze the changes in related diseases on the current ICU bedside chest X-ray images;
[0132] The mediastinal morphology analysis function in the comparative analysis module queries the evaluation of the mediastinum in the most recent report. Based on the evaluation, it judges the changes in mediastinal morphology on the current ICU bedside chest X-ray image and outputs the changes in mediastinal morphology. The evaluation of the mediastinum is normal or abnormal. The changes in mediastinal morphology include: significantly enlarged mediastinal shadow, enlarged mediastinal shadow, reduced mediastinal shadow, and no change in mediastinal shadow.
[0133] Specifically, the mediastinal area in the current registered ICU bedside chest X-ray image is compared with the mediastinal area in the most recent registered ICU bedside chest X-ray image and the mediastinal area in all images of the most recent registered ICU bedside chest X-ray image to determine whether there are significant changes in mediastinal morphology.
[0134] Input the current registered ICU bedside chest X-ray image, the most recent registered ICU bedside chest X-ray image, all images from previous ICU bedside chest X-ray examinations, the mediastinal region in the current registered ICU bedside chest X-ray image, the mediastinum in the most recent registered ICU bedside chest X-ray image, the mediastinal region in all images from previous ICU bedside chest X-ray examinations, and the mediastinal evaluation from the most recent report; Output the previous minimum mediastinal region, qualitative judgment - changes in mediastinal morphology, qualitative judgment - significant increase in mediastinal size, and prompt information - significant increase in mediastinal shadow.
[0135] When changes in mediastinal morphology are present, the results are returned to the "Changes in Mediastinal Morphology" control in the structured report. When the mediastinum significantly enlarges, a prompt message is sent, which is handled by relevant personnel and recorded in the database.
[0136] Analysis of changes in mediastinal morphology:
[0137] When the mediastinum is assessed as normal, the mediastinal morphology analysis unit 402 compares the mediastinal region in the most recent registered ICU bedside chest X-ray image with the mediastinal region in all previous ICU bedside chest X-ray images after registration, identifies the previous minimum mediastinal region, and compares the previous minimum mediastinal region with the mediastinal region in the current registered ICU bedside chest X-ray image. If the mediastinal region in the current registered ICU bedside chest X-ray image is larger than the previous minimum mediastinal region by more than a first preset threshold (e.g., 20%), it is judged as a significant increase in mediastinal shadow; if the mediastinal region in the current registered ICU bedside chest X-ray image is larger than the previous minimum mediastinal region by more than a second preset threshold (e.g., 10%), it is judged as an increase in mediastinal shadow; if the mediastinal region in the current registered ICU bedside chest X-ray image is smaller than or equal to the previous minimum mediastinal region, it is judged as no change in mediastinal shadow.
[0138] When the mediastinum is assessed as abnormal, the mediastinal morphology analysis unit compares the mediastinal region in the most recent registered ICU bedside chest X-ray image with the mediastinal region in the current registered ICU bedside chest X-ray image. If the mediastinal region in the current registered ICU bedside chest X-ray image is larger than a first preset threshold compared to the mediastinal region in the most recent registered ICU bedside chest X-ray image, it is determined that the mediastinal shadow is significantly enlarged; if the mediastinal region in the current registered ICU bedside chest X-ray image is larger than a second preset threshold compared to the mediastinal region in the most recent registered ICU bedside chest X-ray image, it is determined that the mediastinal shadow is enlarged; if the mediastinal region in the current registered ICU bedside chest X-ray image is smaller than a second preset threshold compared to the mediastinal region in the most recent registered ICU bedside chest X-ray image, it is determined that the mediastinal shadow is smaller; if the change in the mediastinal region in the current registered ICU bedside chest X-ray image is less than or equal to the second preset threshold, it is determined that the mediastinal shadow has not changed.
[0139] The pleural cavity lesion analysis unit in the comparative analysis module queries the evaluation of the pleural cavity in the most recent report from previous reports. Based on the evaluation, it judges the changes in pleural cavity lesions on the current ICU bedside chest X-ray image and outputs the changes in pleural cavity lesions. The evaluation status of the pleural cavity is normal or abnormal. The changes in pleural cavity lesions include: newly added pleural cavity lesions, normal pleural cavity, significantly enlarged pleural cavity lesions, enlarged pleural cavity lesions, shrunken pleural cavity lesions, and no change in pleural cavity lesions.
[0140] Specifically: Compare the area of pleural lesions in the current registered ICU bedside chest X-ray image with that in the most recent registered ICU bedside chest X-ray image to determine if there are significant changes in pleural lesions. Input: current registered ICU bedside chest X-ray image, most recent registered ICU bedside chest X-ray image, pneumothorax and pleural effusion areas in the current registered ICU bedside chest X-ray image, pneumothorax and pleural effusion areas in the most recent registered ICU bedside chest X-ray image, and pleural region evaluation from the most recent report; Output: Qualitative judgment - Change in pleural lesions, Qualitative judgment - Significantly enlarged pleural lesions, and Warning message - Significantly enlarged pleural lesions.
[0141] When changes in the pleural cavity morphology are present, the results are returned to the "Changes in Pleural Cavity Morphology" control in the structured report. When a pleural cavity lesion significantly enlarges, an alert is sent for relevant personnel to handle, and the incident is recorded in the database.
[0142] Analysis of pleural cavity lesions:
[0143] The system assesses changes in pleural cavity lesions on current ICU bedside chest X-ray images, including: when the pleural cavity is assessed as normal, the pleural cavity lesion analysis unit detects whether there are pleural cavity lesions in the registered current ICU bedside chest X-ray images. If there are, they are identified as newly added pleural cavity lesions; if not, the pleural cavity is identified as normal.
[0144] When the pleural cavity assessment is abnormal, the pleural cavity lesion analysis unit compares the pleural cavity lesion area in the most recent registered ICU bedside chest X-ray image with the pleural cavity lesion area in the current registered ICU bedside chest X-ray image. If the pleural cavity lesion area in the current registered ICU bedside chest X-ray image is larger than the pleural cavity lesion area in the most recent registered ICU bedside chest X-ray image by more than a third preset threshold (e.g., 20%), it is judged as a significant increase in pleural cavity lesion size. If the pleural cavity lesion area in the registered ICU bedside chest X-ray image increases by more than the fourth preset threshold (e.g., 10%), it is judged as an enlarged pleural cavity lesion; if the pleural cavity lesion area in the current registered ICU bedside chest X-ray image decreases by less than the fourth preset threshold compared to the most recent registered ICU bedside chest X-ray image, it is judged as a shrunken pleural cavity lesion; if the change in the pleural cavity lesion area in the current registered ICU bedside chest X-ray image compared to the most recent registered ICU bedside chest X-ray image is less than or equal to the fourth preset threshold, it is judged as no change in the pleural cavity lesion.
[0145] The lung lesion analysis unit in the comparative analysis module queries the evaluation of lung lesions in the most recent report from previous reports. Based on the evaluation, it judges the changes in lung lesions on the current ICU bedside chest X-ray image and outputs the changes in lung lesions. The evaluation of lung lesions is normal or abnormal. The changes in lung lesions include: new lung lesions, normal lungs, significantly enlarged lung lesions, enlarged lung lesions, shrunken lung lesions, and no change in lung lesions.
[0146] Specifically: Compare the area of lung lesions in the current registered ICU bedside chest X-ray image with that in the most recent registered ICU bedside chest X-ray image to determine if there is a significant change in lung lesions. Input: current registered ICU bedside chest X-ray image, most recent registered ICU bedside chest X-ray image, lung lesions in the current registered ICU bedside chest X-ray image, lung lesion area in the most recent registered ICU bedside chest X-ray image, and lung lesion evaluation from the most recent report; Output: Qualitative judgment - change in lung lesions, Qualitative judgment - significantly increased lung lesions, and prompt information - significantly increased lung lesions.
[0147] When changes in the morphology of lung lesions are observed, the results are returned to the "Changes in the Morphology of Lung Lesions" control in the structured report. When lung lesions significantly enlarge, an alert message is sent, which is then handled by relevant personnel and recorded in the database.
[0148] Analysis of changes in lung lesions:
[0149] The system assesses changes in lung lesions on current ICU bedside chest X-ray images, including: When the lung lesion assessment is normal, the lung lesion analysis unit checks for lung lesions in the registered current ICU bedside chest X-ray image; if present, it is considered a new lung lesion; otherwise, the lungs are considered normal. When the lung lesion assessment is abnormal, the lung lesion analysis unit compares the abnormal high-density lesion area in the most recent registered ICU bedside chest X-ray image with the abnormal high-density lesion area in the current registered ICU bedside chest X-ray image. If the abnormal high-density lesion area in the current registered ICU bedside chest X-ray image increases by more than a fifth preset threshold (e.g., 20%) compared to the abnormal high-density lesion area in the most recent registered ICU bedside chest X-ray image, it is considered a lung lesion. The lung lesion is considered to have increased significantly if the area of abnormal high-density lesions in the pleural cavity of the current registered ICU bedside chest X-ray image is larger than the sixth preset threshold (10%) compared to the area of abnormal high-density lesions in the most recent registered ICU bedside chest X-ray image; the lung lesion is considered to have shrunk if the area of abnormal high-density lesions in the pleural cavity of the current registered ICU bedside chest X-ray image is smaller than the sixth preset threshold; and the lung lesion is considered to have remained unchanged if the change in the area of abnormal high-density lesions in the pleural cavity of the current registered ICU bedside chest X-ray image compared to the area of abnormal high-density lesions in the most recent registered ICU bedside chest X-ray image is less than or equal to the sixth preset threshold.
[0150] Step S505: Automatically output final diagnostic data based on changes in the relevant diseases.
[0151] The structured report module integrates findings from all functional modules to arrive at an overall diagnostic impression. Based on the rules built into the structured report, the final diagnostic data is automatically obtained and returned to the diagnostic impression in the structured report.
[0152] If a critical value is detected in the diagnostic impression, an alert message will be sent, which will be handled by relevant personnel and recorded in the database.
[0153] All data and images are stored in a structured report database.
[0154] This invention applies AI models and rule-based programs to automatically compare and analyze current ICU bedside chest X-rays with previous bedside chest X-rays, resulting in an intelligent diagnostic reporting system for bedside chest X-ray examinations. This system is integrated with PACS / RIS to automatically acquire patient medical history, previous images, and previous reports. After image acquisition, the system segments the images and uses the segmented data, along with previous images and reports, to automatically output changes in mediastinal morphology, pleural cavity lesions, and lung lesions through registered images and registered segmented data regions. This automatic assessment and transmission of results to a structured report improves physician efficiency. More importantly, when the intelligent system detects potentially risky situations, it immediately issues warnings to relevant medical staff, enabling them to promptly address patient abnormalities and prioritize effective treatment, ensuring patient safety.
[0155] As can be seen from the above description, the embodiments of the present invention achieve the following technical effects: The embodiments of the present invention apply AI models and rule-based programs to the automatic comparative analysis of current ICU bedside chest X-rays and previous bedside chest X-rays, resulting in an intelligent diagnostic reporting system for bedside chest X-ray examinations. This system is integrated with PACS / RIS to automatically acquire patient medical history, previous images, and previous report information. After image acquisition, the images are segmented. Using the segmented data, previous images, and previous report information, along with the registered images and registered segmented data regions, the system can automatically output changes in mediastinal morphology, pleural cavity lesions, and lung lesions, automatically completing the assessment and automatically transmitting the results to a structured report, improving the efficiency of doctors' work. More importantly, when the intelligent system detects a potentially risky situation, it will immediately issue a warning to relevant medical staff, helping them to promptly pay attention to the patient's abnormal condition, prioritize effective treatment for these patients, and ensure patient safety.
[0156] Obviously, those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device, or fabricating them separately as individual integrated circuit modules, or fabricating multiple modules or steps as a single integrated circuit module. Thus, the present invention is not limited to any particular hardware and software combination.
[0157] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. An automated comparative diagnostic intelligent reporting system for ICU bedside chest X-rays, characterized in that, The system includes: a patient information acquisition module, a segmentation data acquisition module, an image registration module, a comparison analysis module, and a structured report module. The patient information acquisition module is connected to the segmentation data acquisition module, the image registration module, the comparison analysis module, and the structured report module, respectively, and is used to acquire the patient's current ICU bedside chest X-ray image, patient medical history information, previous images, and previous report information; wherein, the previous images are the patient's most recent ICU bedside chest X-ray image and all ICU bedside chest X-ray images before the most recent one; The segmentation data acquisition module is connected to the patient information acquisition module, the image registration module, and the structured report module, respectively. It is used to input the current ICU bedside chest X-ray image and the previous images into multiple deep learning models for image segmentation, obtaining two corresponding types of segmentation data. The two types of segmentation data include segmentation data of chest structures and segmentation data of chest lesions. Specifically, the segmentation data of chest structures includes the entire chest imaging region, lung field region, and mediastinal region of the current ICU bedside chest X-ray image and the previous images. The segmentation data of chest lesions includes the pleural cavity lesion region and the abnormal high-density lung lesion region of the current ICU bedside chest X-ray image and the most recent ICU bedside chest X-ray image. The pleural cavity lesion region further includes the pneumothorax region and the pleural effusion region. The image registration module is connected to the patient information acquisition module, the segmentation data acquisition module, the comparison analysis module, and the structured report module, respectively. It is used to perform image registration between the current ICU bedside chest X-ray image and previous images based on the segmentation data, and output all registered images and the registered segmentation data. The all registered images include: the current ICU bedside chest X-ray image, the most recent ICU bedside chest X-ray image, and all previously registered ICU bedside chest X-ray images. The registered segmentation data... The data includes: the mediastinal region in the current registered ICU bedside chest X-ray image, the pleural cavity lesion region in the current registered ICU bedside chest X-ray image, the abnormal high-density lesion region in the lungs in the current registered ICU bedside chest X-ray image, the mediastinal region in the most recent registered ICU bedside chest X-ray image, the pleural cavity lesion region in the most recent registered ICU bedside chest X-ray image, the abnormal high-density lesion region in the lungs in the most recent registered ICU bedside chest X-ray image, and the mediastinal region in all previous ICU bedside chest X-ray images after registration. The comparative analysis module is connected to the patient information acquisition module, the image registration module, and the structured report module, respectively, and is used to analyze the changes of related diseases on the current ICU bedside chest X-ray image based on the registered images, the registered segmentation data, and the previous report information. The structured report module is connected to the patient information acquisition module, the segmentation data acquisition module, the image registration module, and the comparison analysis module, and is used to automatically output final diagnostic data based on the changes in the relevant diseases.
2. The ICU bedside chest X-ray automatic comparison diagnostic intelligent reporting system according to claim 1, characterized in that, The comparative analysis module further includes a mediastinal morphology analysis unit, used to query the evaluation of the mediastinum in the most recent report in the previous report information, and based on the evaluation of the mediastinum, to judge the changes in mediastinal morphology on the current ICU bedside chest X-ray image, and output the changes in mediastinal morphology; wherein, the evaluation of the mediastinum is normal or abnormal; the changes in mediastinal morphology include: significantly enlarged mediastinal shadow, enlarged mediastinal shadow, reduced mediastinal shadow, and no change in mediastinal shadow.
3. The ICU bedside chest X-ray automatic comparison diagnostic intelligent reporting system according to claim 1, characterized in that, The comparative analysis module further includes a pleural cavity lesion analysis unit, used to query the evaluation of the pleural cavity in the most recent report in the previous report information, and based on the evaluation of the pleural cavity, to judge the changes in pleural cavity lesions on the current ICU bedside chest X-ray image, and output the changes in pleural cavity lesions; wherein, the evaluation of the pleural cavity is normal or abnormal; the changes in pleural cavity lesions include: newly added pleural cavity lesions, normal pleural cavity, significantly enlarged pleural cavity lesions, enlarged pleural cavity lesions, shrunken pleural cavity lesions, and no change in pleural cavity lesions.
4. The ICU bedside chest X-ray automatic comparison and diagnostic intelligent reporting system according to claim 1, characterized in that, The comparative analysis module further includes a lung lesion analysis unit, used to query the evaluation of lung lesions in the most recent report in the previous report information, and based on the evaluation of lung lesions, to judge the changes in lung lesions on the current ICU bedside chest X-ray image, and output the changes in lung lesions; wherein, the evaluation of lung lesions is normal or abnormal; the changes in lung lesions include: new lung lesions, normal lungs, significantly enlarged lung lesions, enlarged lung lesions, shrunken lung lesions, and no changes in lung lesions.
5. The ICU bedside chest X-ray automatic comparison diagnostic intelligent reporting system according to claim 2, characterized in that, The assessment of changes in mediastinal morphology on the current bedside chest X-ray images in the ICU includes: When the mediastinum is evaluated as normal, the mediastinum morphology analysis unit compares the mediastinal region in the most recent registered ICU bedside chest X-ray image with the mediastinal region in all previous ICU bedside chest X-ray images after registration, identifies the previous minimum mediastinal region, and compares the previous minimum mediastinal region with the mediastinal region in the current registered ICU bedside chest X-ray image. If the mediastinal region in the current registered ICU bedside chest X-ray image is larger than the previous minimum mediastinal region by more than a first preset threshold, it is determined that the mediastinal shadow has significantly increased; if the mediastinal region in the current registered ICU bedside chest X-ray image is larger than the previous minimum mediastinal region by more than a second preset threshold, it is determined that the mediastinal shadow has increased; if the mediastinal region in the current registered ICU bedside chest X-ray image is smaller than or equal to the previous minimum mediastinal region by less than the second preset threshold, it is determined that the mediastinal shadow has not changed. When the evaluation of the mediastinum is abnormal, the mediastinal morphology analysis unit compares the mediastinal region in the most recent registered ICU bedside chest X-ray image with the mediastinal region in the current registered ICU bedside chest X-ray image. If the mediastinal region in the current registered ICU bedside chest X-ray image is larger than the mediastinal region in the most recent registered ICU bedside chest X-ray image by more than the first preset threshold, it is determined that the mediastinal shadow is significantly enlarged; if the mediastinal region in the current registered ICU bedside chest X-ray image is larger than the mediastinal region in the most recent registered ICU bedside chest X-ray image by more than the first preset threshold, it is determined that the mediastinal shadow is significantly enlarged. If the mediastinal region in the ICU bedside chest X-ray image increases by more than the second preset threshold, it is determined that the mediastinal shadow has increased; if the mediastinal region in the registered current ICU bedside chest X-ray image shrinks by less than the second preset threshold compared to the most recent registered ICU bedside chest X-ray image, it is determined that the mediastinal shadow has shrunk; if the change in the mediastinal region in the registered current ICU bedside chest X-ray image compared to the most recent registered ICU bedside chest X-ray image is less than or equal to the second preset threshold, it is determined that the mediastinal shadow has not changed.
6. The ICU bedside chest X-ray automatic comparison diagnostic intelligent reporting system according to claim 3, characterized in that, Assessing changes in pleural cavity lesions on current bedside chest X-ray images in the ICU, including: When the evaluation of the pleural cavity is normal, the pleural cavity lesion analysis unit detects whether there is a pleural cavity lesion in the registered current ICU bedside chest X-ray image. If there is, it is determined to be a newly added pleural cavity lesion; if not, it is determined that the pleural cavity is normal. When the evaluation of the pleural cavity is abnormal, the pleural cavity lesion analysis unit compares the pleural cavity lesion area in the most recent registered ICU bedside chest X-ray image with the pleural cavity lesion area in the current registered ICU bedside chest X-ray image. If the pleural cavity lesion area in the current registered ICU bedside chest X-ray image is larger than the pleural cavity lesion area in the most recent registered ICU bedside chest X-ray image by more than a third preset threshold, then it is determined that the pleural cavity lesion has significantly increased. If the pleural cavity lesion area in the registered ICU bedside chest X-ray image increases by more than a fourth preset threshold, it is determined that the pleural cavity lesion has increased; if the pleural cavity lesion area in the registered current ICU bedside chest X-ray image decreases by less than the fourth preset threshold compared to the pleural cavity lesion area in the most recent registered ICU bedside chest X-ray image, it is determined that the pleural cavity lesion has decreased; if the change in the pleural cavity lesion area in the registered current ICU bedside chest X-ray image compared to the pleural cavity lesion area in the most recent registered ICU bedside chest X-ray image is less than or equal to the fourth preset threshold, it is determined that the pleural cavity lesion has not changed.
7. The ICU bedside chest X-ray automatic comparison diagnostic intelligent reporting system according to claim 4, characterized in that, Assessing changes in lung lesions on the current bedside chest X-ray images in the ICU, including: When the evaluation of the lung lesions is normal, the lung lesion analysis unit detects whether there are lung lesions in the registered current ICU bedside chest X-ray image. If there are, it is determined to be the newly added lung lesion; if not, it is determined that the lungs are normal. When the evaluation of the lung lesions is abnormal, the lung lesion analysis unit compares the abnormal high-density lesion area in the most recent registered ICU bedside chest X-ray image with the abnormal high-density lesion area in the current registered ICU bedside chest X-ray image. If the abnormal high-density lesion area in the current registered ICU bedside chest X-ray image is larger than the abnormal high-density lesion area in the most recent registered ICU bedside chest X-ray image by more than a fifth preset threshold, then the lung lesion is judged to have significantly increased in size. If the abnormal high-density lesion area in the pleural cavity of the current registered ICU bedside chest X-ray image is larger than the abnormal high-density lesion area in the most recent registered ICU bedside chest X-ray image by more than a fifth preset threshold, then the lung lesion is judged to have significantly increased in size. If the area of abnormal high-density lesions in the lungs in the registered ICU bedside chest X-ray image increases by more than a sixth preset threshold, the lung lesion is judged to have increased in size. If the area of abnormal high-density lesions in the lungs in the registered current ICU bedside chest X-ray image decreases by less than the sixth preset threshold compared to the area of abnormal high-density lesions in the lungs in the most recent registered ICU bedside chest X-ray image, the lung lesion is judged to have decreased in size. If the change in the area of abnormal high-density lesions in the pleural cavity in the registered current ICU bedside chest X-ray image compared to the area of abnormal high-density lesions in the lungs in the most recent registered ICU bedside chest X-ray image is less than or equal to the sixth preset threshold, the lung lesion is judged to have remained unchanged.
8. A method for automatic comparative diagnosis and intelligent reporting of ICU bedside chest X-ray films, characterized in that, include: Acquire the patient's current ICU bedside chest X-ray image, patient medical history information, previous images, and previous report information; wherein, the previous images are the patient's most recent ICU bedside chest X-ray image and all ICU bedside chest X-ray images prior to the most recent one; The current ICU bedside chest X-ray image and the previous images are input into multiple deep learning models for image segmentation, resulting in two types of segmentation data. These two types of segmentation data include segmentation data of chest structures and segmentation data of chest lesions. The segmentation data of chest structures includes the entire chest imaging region, lung field region, and mediastinal region of the current ICU bedside chest X-ray image and the previous images. The segmentation data of chest lesions includes the pleural cavity lesion region and the abnormal high-density lesion region of the lung in the current ICU bedside chest X-ray image and the most recent ICU bedside chest X-ray image. The pleural cavity lesion region includes the pneumothorax region and the pleural effusion region. Based on the segmentation data, the current ICU bedside chest X-ray image is registered with the previous images, and all registered images and the registered segmentation data are output. The registered images include: the registered current ICU bedside chest X-ray image, the registered most recent ICU bedside chest X-ray image, and all previously registered previous ICU bedside chest X-ray images. The registered segmentation data includes: the mediastinal region in the registered current ICU bedside chest X-ray image, the pleural cavity lesion region in the registered current ICU bedside chest X-ray image, the abnormal high-density lung lesion region in the registered current ICU bedside chest X-ray image, the mediastinal region in the registered most recent ICU bedside chest X-ray image, the pleural cavity lesion region in the registered most recent ICU bedside chest X-ray image, the abnormal high-density lung lesion region in the registered most recent ICU bedside chest X-ray image, and the mediastinal region in all previously registered previous ICU bedside chest X-ray images. Based on all the registered images, the registered segmentation data, and the previous report information, analyze the changes in related diseases on the current ICU bedside chest X-ray images; The system will automatically output the final diagnostic data based on the changes in the relevant diseases.
Citation Information
Patent Citations
Automated identification and classification of intravascular lesions
US20160157787A1
Patient monitoring system
US5802494A